Claim Charts book, Part II


This part of the book covers the different types of claim charts, and scenarios in which claim charts are employed, starting with initial or preliminary infringement contentions (PICs) — which are the subject of the lengthiest discussion, as also including pre-filing investigations which form the basis for PICs — onto non-initial infringement and invalidity contentions (ICs), and including such other types as domestic industry (DI) charts employed in International Trade Commission (ITC) importation actions; three-column charts incorporating claim construction; claim charts for design patents; expert charts; ‘elimitation charts’ sometimes used when litigating multiple patents with overlapping limitations; and several other types.

For some of the types covered here, e.g. statutory bar chart or enablement chart, the reader may think, “I’ve never heard of that as a type of claim chart,” and indeed, it is rare for contentions in some of these cases to be presented in chart form. However, it will be seen that the chart form, if not always necessary, is at least a useful option for presenting contentions such as lack of enablement.

Further, even while referring to claim “charts” or “tables,” there is no need to fetishize the two-column format, and most of the material herein is equally applicable to any text, whether employing a table or not, and even applicable to a graphical presentation, in which patent claims are split into limitations and compared on a limitation-by-limitation basis with something outside the four corners of the patent claims (in the case of enablement charts, the comparison is with the patent disclosures as a whole).

This part of the book also covers different theories or limitation types presented within charts, such as the doctrine of equivalents (DoE), obviousness, and indirect infringement. When working with claim charts, one must be aware — and this book covers — not only the elements (limitations) of the patent claim at issue, but also the legal elements at issue. For example, patent infringement under 35 USC 271 is defined as making, using, selling, offering for sale, or importing, the patented invention, and the claim chart should at least be the product of some thought as to which particular act is being alleged, and further, how this interacts with the type of patent claim, e.g. device, method, system, and interacts with the particular accused infringer, e.g. who is actually using the accused method? As further examples, DoE generally must be shown by employing a legal framework such as function/way/result (FWR) or known interchangeability; and when charting importation or exportation, the chart must include where the alleged infringement took place. While some legal elements, such as the knowledge or intent components of indirect infringement (inducement and contribution) may not require limitation-by-limitation treatment, and therefore can be handled outside the claim chart, most legal elements likely need to be reflected in the claim chart.

The major claim chart types and scenarios covered in this part of the book are:

  • Initial or preliminary infringement contentions (PICs), based on pre-filing investigation
  • Initial invalidity contentions
  • Infringement contentions
  • Invalidity contentions, especially anticipation (lack of novelty)
  • Method claims
  • Multiple-party issues
  • Doctrine of equivalents (DoE)
  • Means-plus-function (112(f))
  • Indirect infringement: charting inducement and contribution
  • Obviousness
  • “Alice & Bob” chart (subject-matter invalidity)
  • Non-infringement contentions (NICs), including declaratory judgment (DJ)
  • Validity contentions
  • Contentions employing explicit or alternate claim constructions
  • Enablement chart
  • Statutory bar chart (on-sale or public-use bar)
  • Working/practicing chart
  • Domestic industry (DI) chart in ITC importation actions
  • Charting importation (ITC) and exportation
  • Design patent chart
  • Expert claim charts
  • Limitations chart for multiple overlapping asserted patents
  • Representative instrumentality chart
  • Product-to-product comparison chart
  • Injunction/contempt-related chart
  • “Parts list” as demonstrative exhibit
  • Claim charts in petitions to PTO, PTAB for reexam, IPR, PGR, or CBM review
  • Claim chart employed in licensing negotiations
  • Claim construction chart for Markman hearing
  • Joint Claim Construction Chart
  • Deficiency chart, to highlight problems in the opponent’s charts
  • Attorney opinion letter, including Freedom to Operate (FTO)
  • Chart showing dependent-claim trees, patent families
  • Claim chart in jury verdict forms
  • Claim chart employed by PTO examiner

Examples are given for as many chart types as possible, from published opinions and/or from briefs available in cases with a known outcome. Example claim charts are difficult to acquire, as the contents of charts are often confidential, containing e.g. “attorney-eyes only” material from discovery. Even in published opinions regarding claim charts, the chart itself is often omitted (making it difficult to determine the holding of the case, or to compare case outlines; “ex facto jus oritur”). The examples presented here are often not ideal, but at least have the benefit of having been employed in actual litigation, as opposed to contrived training examples.

As indicated in the bulleted list above, this part of the book covers claim-construction charts, including Joint Claim Construction Charts. However, the coverage is brief, as these chart types are otherwise outside the scope of this book, both because they do not employ the “outside the four corners of the patent” comparison which is the focus of the book, and because there is already a wealth of material available on claim construction. One exception is made for an “inside the four corners of the patent” comparison in enablement charts, as these fit naturally with the other invalidity charting discussed here.

While many other types of charts might arise in patent litigation, e.g. charts used in damages calculations, these will not be covered here, as the common denominator here is not the mere use of a chart or tabular format, but rather the limitation-by-limitation parsing of patent claims (or, for design patents, the “every angle” approach) used as a basis for comparison with something outside the four corners of the patent.

Initial or preliminary infringement contentions (PICs), based on pre-filing investigation

The following unusually lengthy section on PICs may strike some readers as odd or out-of-date, as most district LPRs no longer distinguish between PICs and ICs. However, the basic point below is that pre-discovery generation of claim charts has special issues and problems, and this point remains whether the early pre-discovery claim charts are formally called PICs or not.

The initial infringement contentions in a patent infringement case are required under Local Patent Rules (LPRs) to be served early in the case, before discovery. This would also be a requirement under likely forthcoming legislation (see xxx above on HR 3309). Thus, these initial infringement contentions generally are necessarily based on publicly-accessible information gathered as part of the patent owner’s pre-filing investigation. In addition to the requirements imposed by the LPRs, this pre-filing investigation is governed by FRCP Rule 11 (“an inquiry reasonable under the circumstances”), and hence the contents of the initial infringement contentions are as well. See e.g. View Eng’g v. Robotic; Network Caching v. Novell; Eon-Net v. Flagstar. Courts have interpreted the LPR and Rule 11 requirements for initial infringement contentions as amounting to a “reverse engineering or its equivalent” standard for the level of detail required in initial infringement contentions; see e.g. Network Caching v. Novell; Intertrust v. Microsoft (for LPR 3-1, “a plaintiff must put forth information so specific that either reverse engineering or its equivalent is required”); GN Resound v. Callpod; but see STMicroelectronics v. Motorola (reverse engineering not synonymous with LPR 3-1 requirements).

In a few districts there now is, and in many districts there previously was, a formal distinction between so-called initial or preliminary infringement contentions (PICs) and the later infringement contentions (ICs) as amended and/or supplemented during discovery. Some district LPRs also formally refer to “final” infringement contentions. See e.g. ND Ill. LPR 3.1 discussed below at xxx on infringement contentions.

This formal distinction has largely been phased out from the LPRs. For example ND Cal. as early as 2008 eliminated PICs. An advisory subcommittee on LPRs noted the elimination of “the concept of ‘preliminary’ contentions in favor of a single round of contentions which can be modified only for good cause.” The reason given was that separate PICs were “not tying litigants sufficiently to their positions” (cited in James Ware and Brian Davy, The History, Content, Application and Influence of the Northern District of California’s Patent Local Rules, 25 Santa Clara High Tech. L.J. 965 (2012)).

Formally, there is now generally a single IC, deemed final unless there is good cause to amend, see e.g. E.D. Tex. LPR 3-6. Even when separate PICs existed, these too were to some extent deemed final; see e.g. Townshend v. Broadcom (“the default presumption is that a party’s PICs ‘shall be deemed to be that party’s final contentions’,” quoting then N.D. Cal. LPR 3-6).

The bottom line is that a patent owner’s very first claim chart need to be detailed and close to final; see e.g. Intertrust v. Microsoft (regarding deficiencies in P’s PICs, as identified by D, “Some of these deficiencies do appear, as InterTrust describes them, ‘nit picky.’ The purpose of Patent Local Rule 3-1, however, is in fact to be nit picky, to require a plaintiff to crystalize its theory of the case and patent claims. Accordingly, InterTrust must amend the deficiencies listed” by D).

The point is not only to early “crystallize” P’s theories of infringement, but to largely adhere to these early-crystallized theories; see e.g. Atmel v. Info. Storage (“The patent local rules were adopted by this district in order to give claim charts more ‘bite.’ The rules are designed to require parties to crystallize their theories of the case early in the litigation and to adhere to those theories once they have been disclosed. Rule 16-9(c) [providing only narrow bases for claim chart amendment] advances this purpose by making it difficult subsequently to revise claim charts through eleventh hour ‘discovery’ of facts. Unlike the liberal policy for amending pleadings, the philosophy behind amending claim charts is decidedly conservative, and designed to prevent the ‘shifting sands’ approach to claim construction”). [TODO: put “bite” into section on policies behind claim charts.]

The term PICs will be used here, even though generally not designated as such by the LPRs, to designate the patent-owner’s pre-discovery infringement contentions, based on a pre-filing investigation.

[TODO: show inadequate (Connectel v. Cisco) and adequate (AVG v. EA) PICs — explain why AVG chart was deemed sufficient, when it appears to mimic claim language; contrast Connectel v. Cisco, where plaintiff “made shotgun accusations of hundreds of products infringing hundreds of claims,” and AVG v. Elec. Arts, where plaintiff actually bought and played hundreds of accused video games, before filing suit; this difference not apparent from the face of the claim chart]

As noted above at xxx, the LPRs generally require, e.g., “A chart identifying specifically where each limitation of each asserted claim is found within each Accused Instrumentality…”, and this chart is required “Not later than 14 days after the Initial Case Management Conference…” (N.D. Cal. Patent L.R. 3-1(c)).

While not quite as early as the requirement of proposed federal legislation, in which the claim chart would be part of the initial complaint (see xxx re: HR 3309), the LPRs dictate that the patent owner develop its deemed-to-be-final infringement contentions before, or only shortly after, filing suit.

Given this timeframe, these charts unfortunately may often be ground out by an associate attorney, with little or no expert input, using superficial pattern matching of the claim language to vaguely similar wording at the accused infringer’s web site, or in other marketing materials. It appears somewhat rare for plaintiffs or their law firms to do complete “due diligence”. Hiring an expert early in the case, to reverse engineer the acquired product, is often viewed as premature. In any case, the expert will often be retained only after a decision to proceed with the case has already been made, thus limiting the extent to which the expert can impact the PICs (“the train has left the station”).

Yet the PICs almost by definition need to be close to what will end up being the final ICs, because as noted above amendment requires good cause, and good cause often requires not merely newly-discovered information, but information that could not have been discovered earlier even with sufficient “diligence.”

Diligence, in turn, means mining all publicly-available sources of information about accused products; the term used in some cases is “exhaustion”, e.g. NYU v. E.piphany (“The plaintiffs, of course, may first be required to demonstrate that they exhausted other ways of exploring potential infringement” before D will be required to disclose its source code; citing Connectel v. Cisco, where plaintiff “made shotgun accusations of hundreds of products infringing hundreds of claims,” and distinguishing AVG v. Elec. Arts, where plaintiff actually bought and played hundreds of accused video games, before filing suit).

As a small example, it is perhaps insufficiently diligent to treat information regarding an older product as unavailable, based on a single visit to the vendor’s web site, without also checking older versions of the web site at the “Wayback Machine” (; incidentally, while some litigants and experts avoid using the Wayback Machine, out of fear that its contents are hearsay and/or would require time-consuming authentication, its contents are widely though not universally accepted by the courts; see e.g. trademark case, Marten Transport v. PlatForm Advertising). [TODO: move this, and more on, to section on evidence used in claim charts.]

Far more important, the main source of information about an accused product is not marketing literature, nor even technical documentation for the product, but the product itself. Patent owners about to file a patent infringement suit against a product ought to know as much as possible about that product, which generally means purchasing it and taking it apart, i.e. reverse engineering. The category of “publicly-accessible information,” available pre-discovery, most importantly includes what can be learned by taking apart, or otherwise closely examining, products available on the open market.

This product inspection is not simply a generic “teardown” of the product, but rather has as its goal the acquisition of product information corresponding to limitations of the patent claim, such that a limitation-by-limitation claim chart can be assembled. Recall that the LPRs require “A chart identifying specifically where each limitation of each asserted claim is found within each Accused Instrumentality…” (emphasis added) — the goal of pre-filing and/or pre-discovery reverse engineering is to be able to point in the accused product to this where for each limitation.

Obviously there are questions how this applies to method claims, to the sale of services rather than tangible products, to products not sold on the open market, and so on. This is discussed in subsection below, which is followed by a subsection on what is meant by “reverse engineering or its equivalent,” and another on whether “reverse engineering or its equivalent” is truly the standard required for PICs and pre-filing investigation.

Acquiring the accused instrumentality

While most accused instrumentalities in patent litigation are presumably of interest precisely because they generate some revenue and are available on the open market, some products may be difficult to acquire. Courts are unsympathetic to difficulty or expense (see below), but will listen to assertions that the accused instrumentality is not reasonably accessible. In Intamin Ltd. v. Magnetar Technologies (CAFC, 2007), sanctions were inappropriate for failure to obtain and cut open the metal casing in a roller coaster breaking system. The court distinguished this from applicability of attorney sanctions when a sample would have been easy to obtain in Judin v. U.S. and Hewlett-Packard (CAFC, 1997) (no attempt made to inspect Postal Service bar-code scanners, beyond observation from a distance). If it’s more like a scanner than a roller coaster, plaintiff would be wise to get the product.

Method claims: Even for e.g. a web-based service, it is a bad idea to assert that “there is no ‘product’ to obtain” (Eon-Net v. Flagstar, WD Wash., 2006) (granting motion for Rule 11 sanctions; plaintiff’s attorney “didn’t know whether the server invoked a program, although he concluded that it must”). When accusing a process or service of infringing a method claim, plaintiff needs to ask who carries out the method. If it’s the end-user, then plaintiff is asserting indirect infringement, and can obtain and inspect the same product which it complains defendant is putting in the hands of end-users. If the defendant itself is carrying out the service or system, then perhaps only some “manifestation” is publicly accessible (e.g. what one sees when using Lexis isn’t necessarily going to tell you much about Lexis back-end processes; on the other hand, much can often be learned through dynamic reverse engineering; see below).

Older, outdated products: In Vasudevan v. IBM (ND Cal., 2011), Vasudevan admitted that it “did not acquire, install, and test the InfoSphere Warehouse product prior to filing the instant action,” and the court appears to have accepted (though merely for purposes of leave to amend, and explicitly setting aside Rule 11 questions) Vasudevan’s assertion that “it was impossible for it to have purchased earlier, outdated versions of the software.” The IBM software at issue included InfoSphere Warehouse and Alphablox. From a cursory inspection, different versions of Alphablox appear to have been available at IBM’s web site, in exchange for registering one’s name and email address with the site (plaintiff-to-be concerns over contacts with defendant web sites are noted below).

Future products: If plaintiff is going to accuse future products, it will need some basis such as a codename or beta. See Intertrust v. Microsoft (ND Cal., 2003) (“InterTrust asserts that past, present and as yet unmade future versions of Microsoft’s products do and will infringe its patents. Microsoft, however, cannot be expected to defend against claims of infringement for products that it has not produced yet”). See also product versions at XXX in Part III.

Expensive products: It is generally risky to plead the expense of buying and analyzing accused products. For example, in Shared Memory Graphics v. Apple (ND Cal., 2011), “SMG’s insistence that it need not identify the specific circuitry that purportedly satisfy these limitations because to do so would take months and cost hundreds of thousands, or even millions, of dollars is not well taken.” The court cites an earlier case (Bender v. Maxim, ND Cal., 2010) in which, too, the district court was unpersuaded by what it calls the “it is too expensive” argument: “While the Court recognizes that reverse engineering may well pose a financial hardship to plaintiff, it appears that he cannot maintain this lawsuit without undertaking reverse engineering or some equivalent that will enable him to better articulate his claims.” While Bender itself notes cases in which this was not required because it was “either impracticable or unnecessary to create a basis for adequate ICs,” those will be the exception not the rule. Getting and inspecting the product, or at least trying to, is the price of admission.

Negotiated product sales: Some products are arguably not quite on the completely-open market, because one must sign up or register, instead of or in addition to plunking down one’s credit card. This is generally true of web services, of course. Even the price of some products is only quoted following contact with the vendor. Plaintiff-to-be may fear even asking for a product, so as not to tip its hand (or even trigger reasonable apprehension for a declaratory judgment suit). Note however Judin cited earlier, in which the court points to plaintiff’s failure to ask the Post Office for a sample of the device, and note too that a web form requiring registration of name and email, for example, is not a sufficient form of having to ask to constitute public unavailability, because otherwise any non-cash transaction requiring some buyer information (credit card number, shipping address) would be non-public.

Sage Electrochromics v. View (ND Cal., 2013) discusses this issue (patent owner apparently needing to ask suspected infringer for a copy of its product) in some detail: “View conceded, View’s products are custom-made and not available for purchase in store or online. SAGE believed that a purchase in its own name from View, its competitor, was not realistic, which was borne out when View denied SAGE permission to purchase its products after it filed suit. Due to its parent company’s corporate ethics policies, SAGE was prohibited from purchasing View’s products under the name of another entity or individual. View produces samples of its products for trade shows, but the samples are subject to non-disclosure agreements and are not otherwise available. SAGE hired a private investigator to try to obtain discarded samples of View’s products from its waste disposal facility, but the facility was secured and the investigator was unable to do so.”

In other words, SAGE tried. View’s position as summarized by the court was that “where a party suspects infringement but is unable to obtain a sample of an alleged infringing product despite diligent efforts, the party should send a letter to the alleged infringer explaining the bases for its belief that there has been infringement before filing suit…. View argues that SAGE should have attempted to resolve its infringement allegations via a ‘Hoffmann-La Roche letter’ before filing suit, rather than filing its complaint without first obtaining samples of View’s products.” But according to the court, Hoffman-La Roche v. Invamed (CAFC, 2000) “did not hold that a party that suspects infringement but is unable to obtain a competitor’s product must attempt to enter into a non-disclosure agreement to exchange confidential information prior to filing suit,” and that court rejected the defendant’s position that because the plaintiffs ‘were unable to obtain and set forth in their complaint facts showing infringement, they should not have filed suit at all.'”

The View court “concludes that View’s products were not readily available to SAGE. Therefore, as SAGE’s preparation of its ICs [infringement contentions] was restricted by the fact that it did not have reasonable access to View’s products, SAGE appropriately relied on the information known to it through publicly available materials in formulating its ICs.”

Before taking the holding in View as a license to accuse without buying and inspecting, however, prospective plaintiffs should carefully re-read the fact pattern. Note too that what seems impossibly inaccessible to an attorney may be a relative piece of cake to a technical consultant or expert, who should be brought on earlier rather than later (see xxx).

[TODO: insert straw purchasers, using “private investigators” (“IP, PI”) for some purchases where the patent owner wishes to hide its interest from the suspected infringer.]

Too many accused products: Plaintiffs may complain that there are too many accused products to examine them all pre-filing. While in AVG v. Elec. Arts (ED Tex., 2005), plaintiff or its employees apparently felt little burden in purchasing and “testing” over 800 games, by contrast in Vigilos v. Sling Media (ND Cal., 2012, unreported), “While the Court is sensitive to plaintiff’s concerns about the expense of reverse engineering over 200 products, it is also sensitive to the desire of defendant not to be forced to produce proprietary schematics unnecessarily.” At the end of the day, plaintiff is master of the expense of proving its case. In Vasudevan, plaintiff made an “unduly burdensome” objection to defendant’s request that plaintiff perform 436 comparisons (4 products x 109 claim limitations). “This fails because VSi sued IBM and has control over the scope of this action.”

Standards as proxies for products: Plaintiff can in some cases look to a product’s stated reliance upon a standard, without necessarily inspecting the product’s conformance with that standard. Generally, implementation or use of the standard is relevant to individual claim limitations. When an entire claim is said to read on a standard, plaintiff may seek to compare the claim to the standard rather than to the product. In Townshend IP v. Broadcom (ND Cal., 2007), plaintiff apparently believed that any “V.90 compliant” product (V.90 is a dial-up internet standard) may infringe, “regardless of the inner workings of the modem or how it is designed.” Townshend’s PICs mapped its claims onto the V.90 dial-up internet standard, rather than onto the accused products, Broadcom’s “V.90 compliant” modems. The court required Townshend amend its PICs, in light of product source code which Broadcom had provided. Townshend could have had its way, reading the claims onto the standard rather than onto the product, but only if it had committed to not later amend its PICs in light of product information such as source code; it was unwilling to so strongly commit to its “mere compliance with the standard infringes” theory of the case.

Linex v. Belkin (ED Tex., 2008) notes there are times when an industry standard may be sufficiently particular to constitute the sole basis for an infringement case. Linex’s case was not however one of those times, as the 802.11n wireless draft standard in question did not specify implementation details sufficient to differentiate the disparate variety of products (wireless routers, networking cards, and laptops) which plaintiff accused. Plaintiff was given a month to amend its infringement contentions, this time pointing to particular provisions of the 802.11n draft standard, and explaining “how” defendant’s products infringe (as noted below, ED Tex. normally requires that plaintiff PICs point to “where,” but not that plaintiff explain “how”).

Eventually P may need to show that the accused product operates in accordance with the standard, if D denies this (see Fujitsu v. Netgear, CAFC, 2010) (accused infringer is free to either prove that the claims do not cover all implementations of the standard or to prove that it does not practice the standard). Whether or not required at the early PIC stage, plaintiff likely will want to think this through, before getting too deeply into the case based on a bare “our claim reads on the standard, and their web site says it complies with the standard” assumption. This assumption is likely to lead to “it must infringe somewhere” problems. See e.g. CSR v. Freescale (ND Cal., 2013): “Contending that the limitation must exist somewhere in the Accused Product [based on its use of the ATAPI protocol] is not the same as specifically identifying where each limitation is found within the Accused Product.”

Relying solely on marketing materials instead of getting the product: Are prospective plaintiffs required to rely on primary sources (the product itself, and whatever blueprints, diagrams, source code, etc. may be publicly available) rather than secondary sources (marketing literature, technical manuals, etc.)? Generally, yes: marketing literature and manuals do not typically identify what a product contains, much less where it contains it, as plaintiff must try to disclose as part of its PICs.

Marketing materials can be used as part of PICs, for example to satisfy a given simple claim limitation, such as use of a CPU or a network, but relying solely on such material for a claim in its entirety is, at best, risky. In Network Caching v. Novell (ND Cal., 2003), plaintiff’s PICs relied entirely upon defendant’s marketing materials. The court was not convinced that reverse engineering the accused products (sold by Novell, Akamai, Inktomi, and Cacheflow) would not have provided more information, and stayed all discovery until plaintiff provided more detailed PICs. The court found that FRCP 11 (requiring a reasonable pre-filing investigation) established a minimum level of detail required under Local Patent Rules, and that “reverse engineering or its equivalent is required.” At the same time, the court did not as requested dismiss the case, and gave plaintiff another chance, declining to award sanctions despite a finding that plaintiff did not conduct a reasonable pre-filing investigation. Despite harsh words for plaintiffs which fail to reasonably investigate accused instrumentalities, “death penalty” sanctions are rare.

Plaintiffs sometimes cite Q-Pharma v. Jergens (CAFC, 2004) for the proposition that plaintiffs need not acquire defendant’s product, relying instead solely on printed materials. See e.g. plaintiff’s citations to Q-Pharma in e.g. Centillion Data v. Convergys (SD Ind., 2006) (where plaintiff presented “largely advertising fluff”) and Vasudevan v. IBM (ND Cal., 2011). But Q-Pharma concerned a product named “Curel (R) Age Defying Therapeutic Moisturizing Lotion with Coenzyme Q10,” which product Q-Pharma did obtain, though not chemically analyze. The product was labelled with a list of active ingredients, including the eponymous “Coenzyme Q10,” which was the very subject of the patent.

In contrast, product marketing materials generally do not come with a list of ingredients, which plaintiffs instead must instead attempt to obtain from the product itself, using reverse engineering, and/or from any open source used by the product. Reverse engineering can take the place of labelling. When marketing material or other material like web screenshots are employed, plaintiff should not just dump masses of it into PICs; specific defendant statements must be linked to specific claim limitations.

Getting the case off on the wrong foot by not inspecting the accused product: Having shown how to annoy a magistrate or judge (don’t purchase the product you’re accusing, then say that it’s too expensive, or that you’ve accused too many products to afford to buy them, and then base your assertions entirely on marketing “fluff”), it must be noted that the punishment for this practice is generally not severe. Plaintiff is generally given 30 days to amend. Courts avoid what looks like Rule 11 satellite litigation, especially at the start of the case.

Courts will generally remind plaintiff of the need for buying and examining the accused product – just before giving plaintiff another month. In Vasudevan v. IBM (ND Cal, 2011), Vasudevan admitted that it “did not acquire, install, and test the InfoSphere Warehouse product prior to filing the instant action.” The plaintiff was just barely (?) granted leave to amend its claims which had relied solely on product manuals, when plaintiff subsequently learned from actually looking at the product that its conclusions from reading these manuals had been incorrect. IBM was unsuccessful in gaining discovery into the reasonableness of Vasudevan’s pre-filing investigation. The court also left “for another day” IBM’s assertion that plaintiff was obligated under Rule 11 to acquire and test the product before filing suit. That Vasudevan had subsequently examined the product was likely important in its living to fight another day.

One case with significant attorney sanctions involved a defendant’s counterclaim for infringement of its own patents. In View Eng’g v. Robotic Vision, CAFC, 2000), the law firm filed the counterclaim without seeing plaintiff’s products, basely on little more than an inventor’s belief in probable infringement (and the law firm’s stated annoyance at the other side’s obstructionism in discovery). In upholding the challenged attorney sanctions, the court noted the inventor’s belief was, in turn, based solely on advertising and statements made to customers, without any physical access to the accused machines. It does not seem to have helped that the inventor’s affidavit stated: “Further investigation was impossible since the View machine cost several hundred thousand dollars to purchase, and it did not make any economic sense to purchase a machine just for the purpose of making further investigation. In addition, since View is a competitor of [Robotic], it is doubtful that it would have sold [Robotic] a machine in any event.” In other words, they didn’t even try. To reiterate, sanctions were upheld. So the watchword is: at least diligently try to acquire and examine the accused instrumentality, and document all one’s diligent efforts to do so.

What is meant by “reverse engineering or its equivalent”?

The legal definition of reverse engineering is “starting with the known product and working backward to divine the process which aided in its development or manufacture.” Kewanee Oil v. Bicron (416 US 470 (1974)). However, rather than trying to reconstruct e.g. the original source code (for a software product), or duplicate the original product, the goal is generally to learn something about a product’s “as built” design, from lower-level details of the product. In non-litigation use, the purpose of software reverse engineering for example is generally interoperability (learning about program A in order to write program B which will work with program A), learning about flaws (especially security vulnerabilities), analyzing so-called “malware” (worms, trojans, spyware, and viruses), or understanding the dynamic behavior of a network-based system.

Reverse engineering typically appears in intellectual-property literature as a problem to be solved (clickwrap statements disapproving reverse engineering; DMCA anti-circumvention) or as a defense (to trade-secret misappropriation). Here, however, RE is an investigative tool employed in litigation. The reader’s likely “are we actually allowed to reverse engineer the other side’s product?” concerns are addressed in a later section. In patent litigation, reverse engineering “generally involves physical scrutiny and deconstruction of the actual product at issue to ascertain its operation” (Network Caching v. Novell, ND Cal., 2003). Note the purpose: to ascertain its operation. Pulling the product apart is simply a means. Further, ascertaining the actual product’s operation is not a holistic exercise. The goal is to find specific elements embodied in, or steps carried out by, the accused instrumentality, as the factual half of determining whether a reasonable basis exists for concluding that each limitation of a claim reads on the element or step. (Your expert or technical consultant, or a hired “teardown lab,” will need to be periodically reminded that this — not just pulling the product apart as in some car commercial — is the goal.)

[TODO: insert something about e.g., using their sample reports at, e.g. .]

The phrase “reverse engineering or its equivalent is required” appears in dozens of patent cases, albeit largely from ND Cal. For example, InterTrust v. Microsoft (ND Cal., 2003) (for 3-1, “a plaintiff must put forth information so specific that either reverse engineering or its equivalent is required”); GN Resound A/S v. Callpod (ND Cal., March 2013) (“Plaintiff’s contention does not provide the level of detail that the ‘reverse engineering or its equivalent’ standard requires). But see STMicroelectronics v. Motorola (ED Tex., 2004) (“The question of whether the disclosing party conducted reverse engineering or its equivalent is not synonymous with whether it has complied with R. 3-1, which requires only to set forth specific theories of infringement”).

Naturally, “required” is not as simple as it sounds; it is easiest to understand by first working through “or its equivalent.” It will be seen that “equivalent” generally means any method which, looking directly at the product itself, yields sufficient information on elements or steps in the accused product.

The court in Network Caching v. Novell explicitly rejected the notion that marketing materials can supply this equivalent: “NCT presents several engineering treatises and articles to support its view that examination of marketing materials alone can operate as the functional equivalent of traditional reverse engineering. [] Those treatises, however, recognize the limitation of relying solely on marketing materials, white papers and other product documentation.”

But the court also noted that it had rejected defendant’s view that plaintiff was required to reverse engineer to comply with Local Patent Rule 3-1. Instead, the court said, it had ruled that reverse engineering “or its equivalent” was required to provide the requisite level of preliminary-infringement information to defendant. The court observed that it had not earlier elaborated on what an “equivalent” entailed, but it had been animated by the concern that plaintiff’s contentions were “replete with vague discussions of the claim terms” and provided no information apart from the claim terms themselves (see PIC section below on “mimicking” claim terms). The court ends up concluding that, third time’s a charm, plaintiff had finally remedied the deficiencies in its 3rd revised PIC, as to Local Patent Rule 3-1. What had plaintiff done this time? It had continued to rely on marketing materials, white papers, and other publicly available documents. While providing what “may not be an exemplary model of disclosure,” the court stated that Rule 3-1 was satisfied. Apart from this particular case, certainly marketing materials could be used to satisfy a given simple claim limitation, such as use of a CPU or a network, even though likely not an entire claim with all its limitations.

Turning next to Rule 11, the court noted that “looking at these indirect sources of information fails to satisfy the pre-filing investigation requirements described both by View Eng’g [v. Robotic] and the court’s previous order” (emphasis added). “NCT, however, contends that it may compare its patent claims to an accused device without any actual examination of the accused device.” But NCT had done even less pre-filing investigation than was done in the Antonious case (see below), which was remanded to determine if attorney’s extrapolations from inspected golf clubs to non-inspected clubs had been reasonable. The Network Caching court found that plaintiff’s pre-filing investigation was not reasonable. However, it did not award sanctions at that time, as these would also require that the case be “baseless,” which was not be known at the time.

In other words, a plaintiff may be able to squeak by with the claim that marketing materials can provide the equivalent to reverse engineering, but it’s likely to leave a lingering bad taste.

In STMicroelectronics v. Motorola (ED Tex., 2004), the phrase employed is “reverse engineer the accused products or perform an equivalent, detailed analysis of the products.” In NYU v. E.piphany (SD NY, 2006), the court states that reverse engineering is not required “if the plaintiff has an equivalent way of showing the level of detail the rule [11] requires” (citing Renesas v. Nanya), this level of detail being, according to the court, an “element by element” comparison with at least one of each defendant’s products, for a given patent claim. Here, then, “reverse engineering or its equivalent” refers to an inspection of the product itself yielding information required for a preliminary infringement chart mapping each limitation of a claim to something specific in defendant’s product.

Among the dozens of cases referring to “reverse engineering or its equivalent,” the phrase is frequently coupled with “element by element comparison” as the purpose of this required methodology. E.g., Implicit Networks v. HP; Renesas v. Nanya; Bender v. Freescale Semiconductor; Rambus v. Nvidia; Pagemelding v. ESPN. In particular, the cases state that Rule 3-1 requires an element-by-element comparison for each asserted claim for at least one accused product, and “To make such a comparison, reverse engineering or its equivalent is required” (Renesas).

By stating the purpose of the methodology, the courts have thereby also provided guidance on the methodology itself: something that will let a party uncover and identify relevant elements (or steps) embodied in, or carried out by, a product or service available on the market, before filing a complaint, and hence before access is gained to any confidential designs, blueprints, or source code for the product or service. Given only what someone else has put on the market, patent holders are expected to work at identifying (though at this point not expected to prove up) each element of their invention. Rule 3-1 or its equivalent requires that, before discovery (and as a “do not pass go, do not collect $200” pre-requesite to discovery), plaintiff be able to identify the location “where” each element of his invention can be found in defendant’s pile. If it’s your invention, you ought to be able to spot it.

But not only spot or identify, but specify its location. Recall that Rules such as 3-1 typically require a “chart identifying specifically where” each limitation is found. “Where.” This requirement is analogous to a “pinpoint” citation. It might seem that if one can indicate some basis for believing that the product contains A, B, and C, it should be no more trouble to indicate where A, B, and C reside in the product. But identifying a product element is not the same as pinpointing its location. Were plaintiffs required only to identify an element without pinpointing location, plaintiffs could simply deduce the presence of the element (see CSR v. Freescale: “Contending that the limitation must exist somewhere in the Accused Product is not the same as specifically identifying where each limitation is found within the Accused Product”). The difference between “it’s in there” à la Prego, and pinpointing location, is surely even greater with steps carried out by accused processes.

It is hard to see how to achieve element-by-element PICs, short of reverse engineering. “Reverse engineering may well be the only method by which plaintiff will be able to state his infringement contentions with sufficient specificity, including identifying the location of each infringing element and grouping similar accused products” (Bender v. Maxim). Still, any technique or method yielding an element-by-element comparison between at least one claim and an accused product will likely be sufficient. If plaintiff presents sufficient PICs, the court will not much care how it got there (and any defendant complaint about not getting to that point via reverse engineering will sound like “they didn’t say Simon Says”). Conversely, if plaintiff does not present sufficient PICs, then the court will ask what it did, and what it could have done, using publicly-accessible information, and this question may be framed in terms of reverse engineering, e.g. Linex v. Belkin (ED Tex., 2008): “While reverse engineering ‘is not synonymous with whether [a plaintiff] has complied with Patent [Rule] 3-1,’ the completion of this task would evidence Plaintiff’s diligence in preparing the Infringement Contentions now at issue.”

Thus, the term “reverse engineering” here represents an examination of publicly-available primary sources (the product and/or its source code) rather than secondary sources (marketing materials, manuals, etc.), with the examination sufficiently detailed to locate individual elements or steps which can be compared with (i.e., at the same level of granularity as) the claim limitations. As noted earlier, this granularity is guided by, and should to some extent guide, one’s provisional claim construction (as seen in Judin and Eon-Net, an unreasonably broad construction of claim limitations may mislead plaintiff into underestimating what needs to be uncovered before filing suit). This examination is the mark of a reasonable pre-filing investigation under Rule 11, and is designed to yield the information necessary for claim charts under a Local Patent Rule such as 3-1.

Is “reverse engineering or its equivalent” truly required for PICs and pre-filing investigation?

Having determined that “reverse engineering or its equivalent” refers to an examination of an accused product, yielding specific elements or steps in an accused product, is this methodology truly required pre-discovery?

As we’ve seen, it’s easy enough to cite phrases from cases at least verbally maintaining this as a requirement. But, for example Shared Memory Graphics v. Apple (ND Cal., 2011) cites Network Caching as “holding that reverse engineering or its equivalent is required,” even though that case states, “While the court was reluctant to hold that in all cases such a comparison required reverse engineering of the competitor’s products, the court found that reverse engineering or its equivalent was required in the instant case” (ND Cal., 2002), and as seen above the plaintiff squeaked by with use of marketing materials. And see STMicroelectronics v. Motorola (ED Tex., 2004): “The question of whether the disclosing party conducted reverse engineering or its equivalent is not synonymous with whether it has complied with R. 3-1, which requires only to set forth specific theories of infringement.”

So is “reverse engineering or its equivalent” truly required or is it e.g. mere evidence of one’s diligence (as in Linex v. Belkin cited above)?

It helps to look closely at some frequently-cited cases: Intamin (roller coaster), Judin (post office scanner), Antonious (golf clubs), Q-Pharma (Jergens lotion), View Engeering (three-dimensional scanning equipment), and Bender (electrical circuit for amplifying signals). Some were reviewed above regarding mere product purchases; the issue now is reverse engineering of the product.

As noted earlier, in Intamin v. Magnetar (CAFC, 2007), Rule 11 sanctions were denied for failure to obtain and cut open the metal casing in a roller coaster breaking system, the court distinguishing this from Judin (see below). In Intamin, the magnets at issue were encased in metal tubes, and “a visual inspection would not disclose the orientation of the magnets within the tubes…. the technology presented the patentee with unreasonable obstacles to any effort to obtain a sample of Magnetar’s amusement ride brake system, let alone the difficulty of opening the casing.”

In Judin v. US itself (CAFC, 1997; Hewlett-Packard was a co-defendant), attorney sanctions were not only applicable, but the lower court had abused its discretion for failing to award them, when a sample Postal Service bar-code scanner could have been “easily obtained … for a nominal price,” but the plaintiff filed its case based only on what inventor and attorney “observed from a distance,” and inventor’s familiarity with the field. This was insufficient to create the reasonable belief, needed for filing, that the scanners contained two “critical” limitations (fiber optical source or its equivalent, and aspherical and converging lens). “By viewing the accused devices at a distance Judin was able to determine that they had a light source and a rounded tip through which the light passed, and that the light was focused in a pinpoint. That may have been sufficient to put Judin on inquiry about whether the Government was using a device that infringed his patent. But Rule 11 requires more. It requires that the inquiry be undertaken before the suit is filed, not after. Defendants have no choice when served with a complaint if they wish to avoid a default. They must undertake a defense, and that necessarily involves costs. Rule 11 prohibits imposing those costs upon a defendant absent a basis, well-grounded in fact, for bringing the suit.”

The lower court had described Judin’s rationale for not obtaining and reverse engineering – “the scanning devices were made to Government specifications and were not readily available for disassembly because they were in use at Government installations” – as “lame.” Judin and attorney did not attempt “to obtain the device, or a technical description of it, from HP or another vendor.” “The [lower] court noted that Judin could have purchased a device for a minuscule amount, compared to the cost of the litigation.”

In Antonious v. Spalding & Evenflo (CAFC, 2002), evidence that only one golf club had been cut open and inspected, with an inference made from the one to other golf clubs in the same product line based on a single sentence in marketing literature (“Top-Flite Intimidator fairway woods use the same titanium insert technology as the drivers”), was insufficient to determine whether to uphold attorney sanctions for accusing the entire product line.

In Q-Pharma v. Jergens (CAFC, 2004), while plaintiff did not chemically analyze “Curel (R) Age Defying Therapeutic Moisturizing Lotion with Coenzyme Q10,” the product characteristic relevant to the patent patent, “Coenzyme Q10,” was right in the name of the product, which also carried a list of ingredients. Sanctions were not upheld – even though it turned out, post filing, that the product did not contain a therapeutically effective amount of Coenzyme Q10.

In a kneejerk-counterclaim case, View Eng’g v. Robotic Vision, CAFC, 2000), attorney sanctions were upheld for the attorney’s sole reliance on the inventor’s say-so, which in turn was based solely on advertising and statements made to customers, without any physical access to the accused machines (for three-dimensional scanning of computer chips to insure proper lead alignment), whose expense and difficulty of purchasing from a competitor were cited as reasons for going no further.

Finally, in Bender v. Maxim (ND Cal., 2010), the court while denying a motion for dismissal and for sanctions, did find infringement contentions deficient and ordered 3rd revised contentions to be produced in less than one month, where plaintiff exhibited “unwillingness or inability to reverse engineering the accused products in order to determine who he believes they infringe” and where plaintiff “does not dispute that reverse engineering would enable him to discover further information about the allegedly infringing products. His primary contention, rather, is that he lacks sufficient funds to reverse engineer the products he wishes to accuse.”

The Bender court agreed that reverse engineering is not an absolute requirement (citing Intamin and Q-Pharma), but cases in which reverse engineering was not required “have tended to involve situations in which analyzing the accused product was either impracticable or unnecessary to create a basis for adequate ICs.” Here, since “plaintiff has had two chances to amend his infringement contentions and has not yet succeeded in bringing them up to par,” “Reverse engineering may well be the only method by which plaintiff will be able to state his infringement contentions with sufficient specificity, including identifying the location of each infringing element and grouping similar accused products.” In other words, “While the Court recognizes that reverse engineering may well pose a financial hardship to plaintiff, it appears that he cannot maintain this lawsuit without undertaking reverse engineering or some equivalent that will enable him to better articulate his claims.”

It can be seen that, in addition to the procedural posture of the case and precisely what is being requested at that stage (motion to strike, dismissal, Rule 11 sanctions), there are different levels of accessibility. This has already discussed above in terms of the difficulty of acquiring products. As to the difficulty of learning elements from products:

  • Intamin – no sanctions because element at issue was encased, visual inspection would not disclose, and technology was a roller coaster, presenting unreasonable obstacles to obtaining, “let alone the difficulty of opening”.
  • Judin – sanctions applicable when scanner would have been easy and inexpensive to obtain, and observation from a distance was insufficient to disclose two “critical” limitations.
  • Antonious – remand to determine whether sanctions applicable when extrapolating from one cut-open and inspected golf club, and one sentence in marketing literature, to a larger product line.
  • Q-Pharma – no sanctions when lotion product label itself listed relevant product characteristic, even though this turned out, post-filing, to not be present in sufficient quantity.
  • View Eng’g – sanctions when rely on inventor’s word who merely cited expense and difficulty of obtaining machine for scanning microprocessors.
  • Bender – no dismissal or sanctions, but harsh warning that case probably can’t continue case without reverse engineering, despite pleaded expense of reverse engineering circuits; exceptions to reverse-engineering requirement noted when impractical or unnecessary, but no showing of either made here; expense not seen as constituting impracticality.

A fair summary, then, is that “reverse engineering or its equivalent” is required when a critical limitation cannot be discerned from outward visual inspection of the product, and it is not roller-coaster-impractical to obtain and open up the product (the meaning of “open up” or “cut open,” vs. “visual inspection” in the context of software products and services, is an interesting question; see static vs. dynamic reverse engineering below). Expense of acquiring or reverse engineering carries no weight, such explanations being seen as “lame” and not as an example of impracticality. Impracticality (buying and opening up roller-coaster brakes?) and lack of necessity (the critical limitation is listed on the product packaging, albeit without specifying amount, which amount may be important) can be reasons to not reverse engineer at this stage.

Impracticality (including any obstacle created by license agreement language against reverse engineering); lack of necessity; selective reverse engineering of some accused products (“representative instrumentalities”); are taken up below.

Exception for reverse engineering when “impractical”: Taking the first part of the Bender court’s observation that cases not requiring reverse engineering “have tended to involve situations in which analyzing the accused product was either impracticable or unnecessary to create a basis for adequate ICs,” such impracticality may be a matter of expense, practical inaccessibility, and legal inaccessibility.

Disproportionate expense: It was seen above that courts lack sympathy for plaintiffs who plead the expense of buying a product that they’ve accused. This extends to the expense of reverse engineering a product. Courts are rightly outraged when plaintiffs forget that the burden is on them to provide some reason for believing that the defendant might infringe, not for the defendant to proactively show it doesn’t. In Bender, plaintiff’s “primary contention … is that he lacks sufficient funds to reverse engineer the products he wishes to accuse.” Reverse engineering can be expensive, indeed “may well pose a financial hardship to plaintiff,” but that may be the price of admission (“it appears that he cannot maintain this lawsuit without undertaking reverse engineering or some equivalent”). See also Shared Memory v. Apple (rejecting “it is too expensive” argument). As will be discussed further below at representative instrumentalities, however, plaintiff accusing multiple products may avoid the added expense of reverse engineering every accused product – if it can show a reasonable basis for believing that those reverse engineered are indeed representative of those which have not been. See Honeywell v. Acer.

Technical inaccessibility: It was noted earlier that courts don’t care how plaintiff achieves a sufficient PIC, so long as it does, but that given an inadequate PIC a court will ask what the plaintiff could have done. The court will look at whether plaintiff earnestly tried to learn everything possible from public information (see “exhaustion” above).

As noted earlier in the section titled “Who does the comparison? (Who knows more?),” P may have a difficult time determining the operation of D’s products or methods. Unsuccessful attempts will be excused for process/method claims asserted against an inaccessible process, where plaintiff did analyze the public product to determine the method by which it was made. In fact, this is codified in 35 USC 295 (“Presumption: Product made by patented process. In actions alleging infringement of a process patent based on the importation, sale, offered for sale, or use of a product which is made from a process patented in the United States, if the court finds (1) that a substantial likelihood exists that the product was made by the patented process, and (2) that the plaintiff has made a reasonable effort to determine the process actually used in the production of the product and was unable so to determine, the product shall be presumed to have been so made, and the burden of establishing that the product was not made by the process shall be on the party asserting that it was not so made”).

But even this remarkable-sounding burden shifting requires that plaintiff “made a reasonable effort to determine the process actually used in the production of the product and was unable so to determine.”

The steps plaintiff took in Hoffman (see above) provide one example of exhausting the publicly-accessible information. Dynamic reverse engineering provides indirect methods for determining steps carried out by a service, even when the service itself is inaccessible. Recall that what is required at this stage is a reasonable basis for asserting infringement and hence for requiring defendant to open its kimono. At any rate, impracticality of reverse engineering due to technical inaccessibility is not a ready-made excuse to be dropped in as boilerplate into one’s PICs. Attempt musts be made, and the attempts documented.

Impracticality of reverse engineering when source-specific information is required: A different argument is that reverse engineering is impractical because it will not yield the particular information necessary to plaintiff’s preliminary infringement contentions. Plaintiff can argue that the requirements for its PICs are so specific that only defendant’s own documents will reveal them. All things being equal, reverse engineering is less likely to provide the details necessary to show infringement of means-plus-function claims under 35 USC 112(6), because plaintiff must have some reason for believing not only that defendant carries out the claimed function (purpose, role, or output), but also uses the specific “means” (implementation) disclosed in the patent. WMS Gaming has been interpreted in software patent litigation as requiring identification of individual software routines to prove means-plus-function infringement, and this has in turn been seen as requiring source code. Local Patent Rules require that means-plus-function claims be specified in the PIC. Likewise, these rules require early assertion of the doctrine of equivalents, which is generally found to require non-conclusory parsing out of the distinct function, way, and result used by the accused instrumentality; note that “way” is, again, implementation. [TODO: xref to here from DoE and means-plus-function sections.]

This can produce curious litigation stances. Plaintiff, to avoid the work of reverse engineering before discovery (“why guess at what they’re doing, when they’re going to be required to tell us in discovery?”), may wish to stress the stringent PIC requirements for means-plus-function claims and DoE, and the impracticability of finding such specific, detailed, pinpoint information by reverse engineering. To force plaintiff to do reverse engineering before formal discovery commences – i.e., plaintiff should have to work hard at the facts, before defendant should have to – defendants will the need to argue that plaintiff’s requirements for meeting Rule 3-1 are not so fearsome as plaintiff paints, i.e., that plaintiff’s PICs, while requiring reverse engineering for sure, need show less than what plaintiff asserts it needs to show.

Impracticality from non-technical restrictions on reverse engineering: clickwrap and DMCA: One argument which one would think would be raised often, but which appears not to be, is “but Your Honor, we would have liked to reverse engineering defendant’s product, but we can’t, because it’s protected by a clickwrap license agreement which expressly forbids that” or “we tried to reverse engineer their product, but it’s encrypted, compressed, and obfuscated, and trying to unravel any of those would violate DMCA 1201(a) which prohibits efforts taken to ‘circumvent a technological measure that effectively controls access’ to the product, and our reverse engineering doesn’t come within the 1201(f) exception which is limited to reverse engineering ‘necessary to achieve interoperability'”.

This concern has been raised in the secondary literature; for example: “a ban on reverse engineering interferes with legitimate litigation-related investigations. Ironically, such a ban may make it difficult or impossible to detect patent infringement. Many software inventions are internal to the program, and their use cannot be detected without parsing the code. A patent owner who suspects a rival of infringing such a software patent may have no choice but to reverse engineer the rival’s software in order to gain the evidence it needs to file suit [citing View, Judin, and Refac, but noting Vista v. Trac-4 “general rule that Rule 11 requires an infringement plaintiff to examine the defendant’s product in all instances”]. … [This] may deter meritorious patent infringement suits from ever being filed [citing Pamela Samuelson on DMCA, 14 Berk. Tech. L.J. 519 (1999)].

One published opinion touches on this concern. In Centillion Data v. Convergys (SD Ind., 2006), Centillion made “too many inferences that should have been investigated. Centillion argues that it would not have changed its analysis if it had obtained Qwest’s allegedly infringing software because it already had the data it needed and because it could not have reverse engineered the software code pursuant to Qwest’s licensing agreements. But Centillion did not have the data it needed to ensure that Qwest’s software met every element of the claimed invention, either literally or under the doctrine of equivalents. Centillion only had product features from which it inferred that Qwest’s software should perform in a certain way. Moreover, the Court would not expect Centillion to violate a licensing agreement or other law to make a factual inquiry into potential infringement. In this case, obtaining the software and testing its limitations on manipulating and/or receiving data would have given Centillion a factual basis from which to conclude that Qwest’s software did in fact exhibit each of the ‘270 patent’s limitations.” (Emphasis added)

While the court found that plaintiff’s failure to obtain and use the software to ascertain how it functioned was objectively unreasonable, and that what plaintiff did use was “largely advertising fluff,” nonetheless the defendant had not suffered prejudice, and sanctions were inapplicable, because defendants waited too long to complain.

The Centillion court appears to be saying that it would not expect plaintiff to violate a license agreement or public law prohibiting reverse engineering, but that it did not need to, as it could have obtained the software and tested its limitations on manipulating/receiving data. Such testing is, presumably, distinct from reverse engineering. Indeed, one may need to look underneath the plain language of those ubiquitous license agreements to see the concern which animates them. Generally it is using reverse engineering to copy or clone the product. See for example IBM’s statements that, among reverse engineering techniques, only decompilation is “regarded as objectionable” (Brand & Katoh, Interfaces on Trial 2.0, at 138-9).

Apart from defendants appearing not to complain about license-agreement prohibitions or the DMCA in the context of patent litigation (everyone perhaps recognizing an implied judicial exception), with their invocation more likely to be used by insufficiently-diligence plaintiffs as an excuse for sub-par infringement contentions, there are the cases in which defendants proactively complained that plaintiffs didn’t tear apart their product. See e.g. STMicroelectronics v. Motorola (“STM argues that Patent Rule 3-1 preliminary infringement contentions require ‘the party asserting infringement [to] reverse engineer the accused products or perform an equivalent, detailed analysis of the products”). If the court finds PICs sufficient, this complaint will sound as if “reverse engineering” is being wielded as a talisman (“Your Honor, they failed to jump through the hoop labelled Reverse Engineering”).

This will be especially curious when defendant’s products carry the customary prohibitory language. See e.g. Intertrust v. Microsoft (“Microsoft contends that InterTrust has violated the Disclosure Rule because: (1) InterTrust describes Microsoft’s products by their function rather than their name … and (3) InterTrust fails to disclose how the accused Microsoft products allegedly practice the elements of InterTrust’s asserted patent claims … At the Patent Local Rule 3-1 Disclosure stage, a plaintiff must put forth information so specific that either reverse engineering or its equivalent is required”) coupled with defendant Microsoft’s standard end-user license agreement (EULA) (“You may not reverse engineer, decompile, or disassemble the Software, except and only to the extent that such activity is expressly permitted by applicable law notwithstanding this limitation”).

A Google search for the phrase “except and only to the extent that such activity is expressly permitted by applicable law notwithstanding” shows that this is common language, not only at Microsoft. Judicial requirements of reverse engineering before filing suit should (and likely implicitly are) viewed as permission by applicable law. When magistrates and judges order such reverse engineering, they are doing so with at least judicial notice of common license agreements and the DMCA.

The “or its equivalent” phrase can be employed here, as the necessary pre-filing techniques (especially dynamic reverse engineering discussed below, but also most types of static reverse engineering short of decompilation) may not fit the relevant definition of “reverse engineering.”

Exception for reverse engineering when “unnecessary”: Besides impracticality, the Bender court noted that lack of necessity provides the other reason not to reverse engineer an accused product. Reverse engineering may be unnecessary if the same information is available from a secondary source such as marketing materials, or if detailed technical information is available in the form of schematics or open source code, or if information on one product can reasonably be extrapolated from reverse engineering of another product in the same product line.

Information available from indirect sources: Plaintiff made this argument in Network Caching v. Novell (ND Cal., 2003), stating that reverse engineering the accused products would not have provided more information than the marketing material upon which it entirely relied. The court was unconvinced, staying discovery until plaintiff provided more detailed PICs. Note the schedule: first you learn more information from public sources, then you get discovery.

The point was more successful in Q-Pharma v. Jergens, where the court accepted the point that the product label already named the ingredient whose presence would have been revealed by chemical analysis. Recall that the eponymous “Coenzyme Q10” turned out to be present in a less-than-therapeutic amount, which fact presumably would have been revealed by inspecting the product itself. This is an excellent example of why a party’s own marketing materials cannot be relied upon for an accurate picture of that party’s products, at the level of detail needed in patent infringement, and thus why reverse engineering usually is necessary.

Representative instrumentalities: It has already been noted at xxx that claim charts may employ “representative instrumentalities” or “exemplars,” but that plaintiff must be able to explain why and how the products that were obtained and inspected are representative of those which were not. See e.g. Renesas v. Nanya: “A sufficient investigation for purposes of [FRCP 11 and Local Patent Rule 3-1] does not require a plaintiff to reverse engineer every accused product. Rather, a plaintiff need only provide sufficient evidence to convince the court that reverse engineering would not provide more detail regarding any potential infringement. The plaintiff’s investigation must be sufficient to permit a reasonable inference that all accused products infringe. Essentially, a plaintiff must be able to show why it believed before filing the claim that it had a reasonable chance of proving infringement” (emphasis added). Plaintiff had grouped the accused products into “A” and “B” version products, with inspection of two “A” products and one “B” product, and an explanation of why it could extrapolate to the others.

In Antonious v. Spalding & Evenflo (CAFC, 2002), the inventor had purchased one golf club from Spanding’s “Intiminator” product line, cut open the club head, inspected it, and determined the presence of a relevant structural characteristic. From this, combined with marketing literature, the inventor and attorneys inferred that the claim (as attorneys interpreted it) read on the entire product line. The question was thus whether reverse engineering the one product sufficed as a reasonable basis for accusing the line of products. The court noted that it sometimes can, when the investigation uncovered evidence sufficient to permit a reasonable inference that the non-examined products infringe.

Here, the inference was that the non-examined clubs would have an interior structure similar or identical to that of the driver that had been cut open. The inventor had “looked at” (but not cut open) several other Intimidator metal woods, and had reviewed all the available Spalding literature. As a reasonable basis for inference from one to the other, attorneys pointed to a single sentence: “Top-Flite Intimidator fairway woods use the same titanium insert technology as the drivers,” and noted that its literature search found “no results that were inconsistent with” that reference (see discussion of expert reports below for use of the “not inconsistent with” and similar phrases). From this, the district court had found that attorneys had no basis for asserting that the at-issue claim limitation was present in any of the accused golf clubs, and ordered sanctions for this as well as for the attorney’s claim construction.

The CAFC vacated the sanctions for claim construction (essentially, interpreting “extends into and connects into” as reading on tangential touching or abutment, was not so clearly incorrect as to be frivolous), and remanded the pre-filing investigation sanctions, stating that the facts as given were insufficient to determine whether the attorneys had evidence from which an objectively reasonable attorney would have inferred that each club in Spalding’s “Intimidator” product line had an interior structure that infringed the claim at issue.

The intersection here of legal stretching (a broad interpretation of the claim limitation) and thin factual basis (a single sentence about using “the same” technology) led plaintiff into a difficult position. Plaintiffs extrapolating from a reverse-engineered product to others will want to explain what it is that makes the one representative of the others. How defendants wish to respond is a different issue. It is perhaps not worth making a stink if the uninvestigated products actually are similar, in the relevant characteristics, to what plaintiff chose as its representative instrumentality – even if plaintiff didn’t have a good basis for leaping to this conclusion.

What reverse engineering methods and tools are available?

An excellent introduction is Robert Messler’s Reverse Engineering: Mechanisms, Structures, Systems & Materials (2013). See also Kevin Otto and Kristin Wood, Product Design: Techniques in Reverse Engineering and New Product Development (2000). These books cover reverse engineering of tangible items, and have only limited coverage of software reverse engineering. The following is a brief description of the two main types of reverse engineering — static and dynamic — using software as an example.

“Static” reverse engineering involves the classic image of reverse engineering, i.e., taking something apart. In the case of software, static reverse engineering treats the files that comprise a software product as texts to be examined, without running the product, and often without even installing it. These files are “read” in some manner, perhaps with some translation into a more readily-understandable language. Some of the methods available for static software reverse engineering include: string extraction, metadata extraction, disassembly, and (for some programming languages such as Java) decompilation.

In contrast, “dynamic” reverse engineering, which is especially useful when investigating method claims, involves running the product of interest, often under the control of some tool which can monitor or log (i.e., record to a file) the product’s behavior. In the case of software, dynamic reverse-engineering tools include debuggers (including those now built into most web browsers), network monitors (packet sniffers) such as the well-known “Wireshark” and “Fiddler” tools, file monitors, and API trace utilities. One danger of dynamic reverse engineering is that of over-generalizing from behavior seen in some number of tests, which may not fully reflect the product’s implementation as would be seen through static reverse engineering.

Of course, reverse engineering work will generally be done by a hired consultant or expert, or in-house by a patent owner, rather than by a law firm. There are also “teardown labs” which can help with hardware reverse engineering (less typically with software). In all cases, those hired to do reverse engineering must often be guided away from a tendency to get off in the weeds trying to achieve a holistic understanding of how the entire product works, as opposed to the limitation-by-limitation analysis needed to determine infringement, and to fill out a claim chart.

One should also always check the internet to see if someone else has already done a product teardown or otherwise reverse engineered at least some portion of the product of interest; it is surprisingly common to find such material.

Initial invalidity contentions

Not every district with LPRs has initial or preliminary invalidity contentions. A given district may simply have invalidity contentions, period. As with the relationship between PICs and infringement contentions, even those districts with explicitly-designated initial or preliminary contentions likely view them as an earlier version of the invalidity contentions, not (as in the past) as two separate filings [cite].

An accused infringer, if it is going to produce invalidity contentions, must do so on a very tight schedule. As noted earlier in the section on Local Patent Rules (LPRs), while the accused infringer generally has a long period (28 to 45 days following service of infringement contentions) in which to produce its initial invalidity contentions — in contrast to the seemingly much shorter period (10 days of the initial case management conference) that the patent owner has to produce its initial infringement contentions (PICs) — in reality the production of initial invalidity contentions is on a tighter deadline, as the defendant has little control over when the plaintiff will choose to file suit.

An accused infringer may quite deliberately chose not to produce initial invalidity contentions. It may decide to rest its case entirely on non-infringement, in which case it can wait until xxx days of receipt of the patent owner’s PICs, before filing non-infringement contentions. [TODO: timeline will generally give D a chance to see P’s PICs, before D has to decide on invalidity?]

However, this is risky. The defendant’s decision not to contest the patent’s validity will likely be based on the defendant’s assessment that the plaintiff’s infringement case is weak. That assessment may be premature. Having failed to file any initial invalidity contentions, the defendant may come to regret this strategic decision. If the party fails to produce initial invalidity contentions within the timeframe set forth in the LPRs, the party may be precluded from raising an invalidity challenge. See e.g. Joao Bock v. Online Resources (accused infringer strategically chose not to file any invalidity contentions at all before deadline; “strategic” in that based on original assessment of the merits of P’s infringement case); Emory University v. Nova Biogenetics; MASS Engineered v. Ergotron.

The same point also applies to invalidity contentions regarding some subset of the patents asserted in the case, and regarding some subset of possible invalidity theories. See e.g. ACCO Brands v. PC Guardian (failure to provide invalidity contentions as to a particular asserted patent; identifying invalidity theories in discovery responses did not suffice).

The theories of invalidity include anticipation (lack of novelty), the on-sale and public-use bars, obviousness, lack of enablement, written description, indefiniteness, 101 subject matter (especially Alice), inventorship defenses, and so on. Anticipation, prior-art references, and the on-sale and public-use bars are discussed below in the section on non-initial invalidity contentions, and obviousness and lack of enablement in respective sections. See also the section on “Alice and Bob” charts for 101 subject-matter defense. Other defenses are covered only peripherally, as generally not requiring a limitation-by-limitation analysis, i.e. a claim chart. [TODO: make sure there is enough on on-sale and public-use bars re: the patent owner/inventor’s own activities more than one year before patent filing date.]

The accused infringer with an invalidity case not only needs, on a tight schedule, to identify the particular patent claims it will change and the theories it will employ, but also must produce claim charts which meet the LPR requirements of a limitation-by-limitation analysis, e.g. S.D. Cal. LPR 3-3(c): “A chart identifying where specifically in each alleged item of prior art each element of each asserted claim is found, including for each element that such party contends is governed by 35 U.S.C. ç 112(6), the identity of the structure(s), act(s), or material(s) in each item of prior art that performs the claimed function.” [TODO: SDCAL has explicit “Preliminary Infringement Contentions”, but e.g. NDCAL does not.]

This early claim-chart requirement means that prior-art references cannot simply be listed, as a placeholder for later elaboration. See e.g. Anascape v. Microsoft (striking a portion of invalidity contentions which merely referenced a prior-art item, without mapping aspects of the reference to each element of the claim). Courts do not accept placeholders, boilerplate, or a “reservation of right to amend” generally in patent litigation contentions, including in initial invalidity contentions.

At the same, if a prior-art reference is known and not referenced, it will be extremely difficult to later use non-initial invalidity contentions which chart that prior-art reference (see cases below at xxx on invalidity contentions). Thus, the accused infringer’s best recourse is to make a good-faith effort at charting each known prior-art reference (up to any limit on prior-art references; see xxx) that is likely to become important, including under the court’s yet-unknown claim construction. [TODO: is this really good advice, or is it encouraging expensive overkill?]

The LPRs, in addition to the chart requirement above, also expect the accused infringer to make an early decision on obviousness, e.g. S.D. Cal. LPR 3-3(b): “Whether each item of prior art anticipates each asserted claim or renders it obvious. If obviousness is alleged, an explanation of why the prior art renders the asserted claim obvious, including an identification of any combinations of prior art showing obviousness.”

Note that, even though obviousness is noted separately from the chart requirement, the requirement of “why” and “any combinations” largely precludes an a boilerplate placeholder for later elaboration of an obviousness theory. See xxx below for cases in which a later attempt at obviousness contentions failed, for lack of earlier compliance with the rule. In particular, a mere list of prior-art references, without identifying any combinations among them, can’t disclose an obviousness theory of invalidity.

See xxx below on non-initial invalidity contentions.

Infringement contentions (post-discovery)

Producing and critiquing infringement contentions (ICs) is the core subject of this book, so little need be said about them here, except as they differ from initial or preliminary infringement contentions (PICs) discussed above. Even when a district’s rules do not distinguish between initial and non-initial or final ICs, there remains an important distinction between pre-discovery and post-discovery ICs. (This same distinction also applied to pre- vs. post-discovery invalidity contentions.)

N.D. Ill. is an example district where a distinction is made between initial and final ICs. N.D. Ill. LPR 3.1 requires service of “Final Infringement Contentions” within one-half year (21 weeks) after the due date for service of “Initial Infringement Contentions.” Cases from this district note e.g. that “the local patent rules require the plaintiff to file its final infringement contentions after the close of fact discovery, but before claim construction and expert discovery” (Sloan Valve v. Zurn).

Many district LPRs no longer refer to separate PICs, and state that ICs will be deemed final unless there is good cause to amend; e.g. E.D. Tex. LPR 3-6: “Each party’s ‘Infringement Contentions’ and ‘Invalidity Contentions’ shall be deemed to be that party’s final contentions, except as set forth below,” followed by exceptions in which leave to amend is not, or is, required (leave to amend not required if party believes in good faith that court’s claim construction requires amendment; otherwise, leave to amend required, with leave entered only upon a showing of good cause). See also the point made at PICs above (citing Townshend v. Broadcom) that even PICs may be regarded as final, absent leave to amend.

Since the initially submitted ICs — whether called PICs or not — are produced pre-discovery (indeed, may have been produced pre-filing, and should at any rate be the result of a pre-filing investigation), it is to be expected that discovery will have some effect, resulting in new information, theories, asserted claims, and/or accused products. Assuming good cause to amend, this new material will often be recorded in an amended or supplemental IC because, absent disclosure in an IC, the party will often be precluded from pursuing the theory, claim, or product. See xxx on results of inadequate claim charts. It would generally be easier to amend with newly-asserted claims than with newly-accused products.

Simply having better information from discovery to prove up an existing theory as to infringement of an already-asserted claim by an already-accused product would appear to not require a new IC in those districts which do not distinguish between PICs and ICs. However, particularly if the party in its IC only has conclusory statements (including mimicking claim language, over-use of “information and belief,” failure to explicate F/W/R for DoE, or SAM for means-plus-function), then it is dangerous to leave the IC as is. See cases at xxx where insufficient ICs resulted in, or at least helped lead to, unfavorable SJ, and see cases noted at the same place where omissions from the ICs precluded experts from testifying to a crucial issue, which in turn led to an unfavorable outcome on that issue. [TODO: as example of over-use of info & belief, e.g. DoE analysis where could show F & R, and only need info & belief for W (way/implementation), but instead just did blanket info & belief for DoE as a whole.]

Thus, the first-filed ICs created pre-discovery, whether called PICs or not, should often be amended or supplemented during or following discovery, to the extent allowable for good cause. A party may be compelled to amend its ICs (see xxx). Conversely, allowance to amend will depend upon such factors as the party’s diligence in seeking out information; whether the new information is truly “new” or instead would have been earlier accessible with greater diligence; brief elapsed time between finding the information and asking to amend the ICs; the other side’s reliance on existing ICs; whether discovery has closed; whether the court has ruled on claim construction; etc. (see xxx). Other factors aside, early amendment is much more likely to be granted than later amendment. See e.g. Cap Co. v. McAfee (because the court had not yet construed the claims, and the case was in an early stage, grant patentee leave to amend deficient ICs), contrast Allvoice v. Microsoft (CAFC; patentee waited over three months after claim construction order before seeking leave to amend ICs to identify a product component with a claim limitation; SJ not abuse of discretion).

The bottom line is that it is safest to get as much into the ICs as possible, as early as possible. At the same time, it doesn’t help much, if at all, to have mere boilerplate placeholders, as often found especially for the DoE; see e.g. Rambus v. Hynix (granting SJ of non-infringement under DoE, when patent-owner’s Final ICs notes, “to the extent that any limitation is found to be not literally present, Rambus asserts that such limitation is present under the doctrine of equivalents”; “The Patent Local Rules require a limitation-by-limitation analysis, not a boilerplate reservation. The doctrine of equivalents exists to prevent a ‘fraud on the patent.’ It is not designed to give a patentee a second shot at proving infringement…”). But also see xxx in which some courts have appeared to encourage the use boilerplate placeholders (e.g. Depomed v. Activis).

Assuming a distinction between pre-discovery ICs (whether designated PICs or not) on the one hand, and ICs during or after discovery on the other hand, some considerations for turning the first into the second:

  • If confidential information (CBI) such as source code has at this stage been made available in discovery, it can be referenced in the chart, to replace or supplement earlier PICs reliance on publicly-available (non-source) information.
  • Pinpointing: information from discovery, such as source code, can be used in the post-discovery ICs to “pinpoint,” or more precisely pinpoint, the location where each limitation is found in the accused/anticipating product; see xxx.
  • Any hedging or generality in the PICs should be replaced with a firm commitment to specific associations with claim limitations and product features/attributes: no more “See, e.g.,” or “for example but not necessarily limited to…” (these ought to be avoided even in the PICs).
  • “On information and belief” in the PICs should be replaced with better information, or the assertion should be removed (see xxx above on FRCP (11)(3)(3): the ACN points out that, should an information & belief assertion not pan out, the litigant is called upon “not thereafter to advocate such claims or defenses”).
  • Determine if new information from discovery results in a new theory of infringement, new product to accuse, or new claims to assert. See xxx and xxx on the obstacles to amending with “new” products that already existed, and asserting new theories which could have earlier been asserted.
  • Go through checklist at xxx of possible problems.

Note that the rules regarding infringement contentions also apply to defendant’s counterclaims; see e.g. View Eng’g v. Robotic.

Further discussion of e.g. DoE and means-plus-function is provided elsewhere.

Invalidity contentions, especially anticipation (lack of novelty)

Initial or preliminary invalidity contentions were discussed above. The relationship between initial and non-initial invalidity contentions is parallel to that between initial infringement contentions (PICs) and non-initial infringement contentions, also discussed above, though with some important differences.

Notably, whereas infringement contentions are a necessary part of patent litigation, invalidity contentions are not (apart from a declaratory judgment action; see xxx), as an accused infringer may choose not to contest (or may have no non-frivolous option of contesting) an asserted patent’s validity, and may confine itself to non-infringement contentions (see above). See initial invalidity contentions above on the choice of whether to contest validity in the first place.

Another notable difference between infringement and invalidity is that the accused infringer asserting invalidity faces an uphill battle, steeper than that faced by a patent owner alleging infringement, because while infringement need only be proved by a preponderance of the evidence, invalidity must be proved by clear and convincing evidence, to overcome the presumption of validity carried by patents which have already undergone an examination by the US PTO. The often-stated mirror image between infringement and invalidity (“that which infringes if after, anticipates if before”) does not account for this asymmetric burden of proof, which apart from the legal burden, is also represented in the regard with which many juries hold patents. [TODO: some question if the presumption varies for prior art which the PTO did not examine, especially if one can show that the patent owner knew of the prior art but failed to disclose it to the PTO (see xxx on inequitable conduct).]

Invalidity contentions will include theories of (based on 35 USC 102, 103, and 112) of anticipation (lack of novelty), on-sale bar, public-use bar, obviousness, lack of enablement, written description, and indefiniteness, as well as 101 subject-matter challenges under Alice, improper inventorship, and prior inventorship. [TODO: give one example for each.]

Here, the focus will be on anticipation (lack of novelty) under 35 USC 102, as obviousness under 103 and enablement under 112 are discussed elsewhere, and other theories of invalidity such as indefiniteness have less need for the per-limitation analysis of a claim chart.

For the most part, challenges to patent validity apply to individual claims, not to the patent as a whole.

A crucial point is that even if an independent claim is invalid, its dependent claims may be valid. This is in contrast to infringement, where a dependent claim is only infringed if the independent claim from which is depends is also infringed. For example, if independent claim 1 contains elements A, B, and C, then claim 1 is invalidated by prior art containing A+B+C. If dependent claim 2 adds element D, claim 2 would still be valid over that prior art, if the prior art does not contain D. In contrast, and accused product containing only A+B would not infringe claim 1, and could not possibly infringe claim 2 (which requires A+B+C+D) even if the accused product also contains D. The important implication for those preparing invalidity contention claim charts is to take dependent claims seriously. See also xxx on validity contentions (having knocked out prior art as to claim 1, don’t forget to consider dependent claim 2). [TODO: make sure this point is also in xxx on validity contentions.]

Rules regarding invalidity contentions, and anticipation and obviousness contentions in particular, were discussed earlier at xxx on LPRs. To recall, one of the requirements is “A chart identifying where specifically in each alleged item of prior art each limitation of each asserted claim is found…” (N.D. Cal. Patent L.R. 3-3(c)). In some cases accused infringers use what is labelled a “Prior Art Statement,” but this appears to have the same element-by-element requirements as an invalidity contention claim chart. [e.g. D.Minn.:] Also see invalidity charts presented to the PTO in requests for reexam (see xxx, including the claim-chart example given there from MPEP 2214).

The prior art available (see 35 USC 102) includes not only patents and prior-art publications (including, by implication, published patent applications), but also the class of prior art under 102(b) which is not a “printed publication,” including on-sale and public-use material (see e.g. N.D. Cal. Patent L.R. 3-3(b)) such as publicly-visible methods or devices. “Printed publication” is defined broadly to include electronic sources, software, database contents, and so on. Prior art must have been publicly available at the relevant time (see e.g. the “theses” cases, xxx). Except for on-sale and public-use by the inventor, the prior must have been capable of enabling the PHOSITA at the time to make and use the invention (see xxx). [TODO: reword, trying to use terminology neutral between pre-AIA and AIA 102.]

A good starting point for prior-art research is of course the prior art cited in the patent and in the file wrapper, but an invalidity argument which relies upon prior art that was also considered by the PTO examiner, and presumably taken into account in issuing the patent, is more difficult than an invalidity argument using prior art not considered by the PTO.

Below is part of an anticipation chart from Geffner v. Linear Bearings, 936 F.Supp. 1150 (EDNY, 1996); see also the “parts list” version, from the same opinion, at xxx. The opinion refers to the claim chart as showing that all limitations are present in the prior art, and uses the chart to make the point that, regardless of various distinguishing features between the patent claim and the prior art, “these distinguishing features are not stated in the language specified by the Geffner claims,” and therefore “have no consequence on the anticipation issue in this lawsuit”:


[TODO: possibly also use invalidity claim charts from GE v. Nintendo (xxx) or from Twin Disc v. US (10 Cl.Ct. 713 (1986)).]

Showing anticipation of a claim requires finding all limitations of that claim in a single reference. As noted elsewhere (see xxx), when comparing a patent claim to a prior-art patent, the comparison is not claim-to-claim, but rather between the patent claim at issue in the case, and the entire disclosure of the prior art patent, including the specification and drawings. A limitation may be present inherently, though this can be a difficult argument [TODO: see xxx].

Of course, a limitation may (and generally will) be represented in the prior art using different nomenclature from that used in the patent claim at issue. It is important to explicitly assert that what is called X in the prior-art reference is the same as what is called Y in the newer patent claim. (Conversely, a patent owner arguing validity may show that what the prior art calls X is still not the same as what its patent calls X.) This explicit assertion is different from merely “aping” or “mimicking” the claim language by baldly asserting that the prior art contains Y, but not linking this to the fact the ostensibly-corresponding element in the prior art appears to be something different, called X. [TODO: make this clearer with a better example; point to the Geffner/Manby claim chart as an example of the difference between tying and what courts have called “aping”.]

Speaking of nomenclature, the comparison must be from the prior art to an as-construed patent claim (see xxx on claim construction). For example, given a “means for…” claim limitation, the comparison to be meaningful should be the specific means disclosed in the patent specification, not to the raw language of claim. See means plus function below; the point here is merely that this applies to invalidity analysis as well as to infringement analysis. Often the invalidity contentions will need to be generated in the absence of a definitive Markman claim-construction ruling, and thus the invalidity contentions may need to deal with multiple alternate constructions. At the very least, the construction chosen in the invalidity contentions should be reasonable (e.g., xxx).

[TODO: the LPRs specifically note means-plus-function, e.g. N.D. Cal. LPR 3-3(c): “A chart … including for each limitation that such party contends is governed by 35 U.S.C. ç 112(6), the identity of the structure(s), act(s), or material(s) in each item of prior art that performs the claimed function.” In the Geffner/Manby example above, one of the limitations is “retainer means” — is this functional claiming, and does the right column identify a SAM?]

If all the limitations of a claim cannot be found in a single prior-art reference (including with inherency, see xxx), then the accused infringer should consider obviousness, with a combination of multiple prior-art references. See obviousness below. [TODO: probably need something here on inherency.]

Some points about invalidity contentions and prior-art references, derived from numerous cases in which invalidity contentions were found insufficient, and/or the accused infringer was denied leave to amend them:

  • Don’t merely reference a prior-art item, without mapping aspects of the reference to each element of the claim; e.g. Anascape v. Microsoft (striking this portion of the contentions).
  • Don’t use the expert report as a substitute for the invalidity contentions, either entirely or to “back-fill” something missing from the contentions; e.g. Better Bags v. Illinois Tool Works (striking obviousness contentions included in the export report, not in the party’s invalidity contentions).
  • Don’t expect it to be easy to later include prior-art references unless they are truly newly discovered and only newly discoverable: e.g. Avago v. Elan Microelectronics (granting motion to strike portion of expert report on invalidity, relying on new prior-art references not cited in party’s invalidity contentions; reject argument that because references were cited in other previously-identified prior art, it was permissible to rely); Allure Energy v. Nest Labs (denying leave to amend invalidity contentions to include allegedly new-discovered additional prior art, as expert had referenced the prior art, therefore not newly-discovered and without good cause to amend; conversely, the expert’s reference provides insufficient notice).
  • That a prior-art reference was newly discovered does not mean that it was only newly discoverable: e.g. e.g. Rackable v. Super Micro (“strategic decision not to commission certain searches for prior art until later in the case”)
  • There is a difference between newly-discovered prior art on the one hand, and known prior art with a newly-realized significance on the other hand; often the second is presented as if were the first, but the court may see through this: e.g. Veolia Water v. Siemens (failed to appreciate significance of already-known prior art); Breville v. Storebound (wants to amend not because new materials have come to light in discovery, but because it wishes to make a new argument, contrast case where new information was received in discovery, Vasudevan v. IBM); Cummins-Allison v. SBM (“A party cannot argue that because of its own inability and delay in understanding the divulged prior art, it possesses good cause to prepare new invalidity contentions. Accepting such an argument would encourage parties to engage in delay tactics under the guise of the inability to understand prior art and only seek expert advice late in the discovery process.” [TODO: case has list of factors — use in amending/supplementing section]
  • Don’t put forward unrelated reasons such as claim construction for newly disclosing prior art references: e.g. BreathableBaby v. Crown Crafts (the terms construed by the court appear unrelated to the newly-disclosed prior-art references; court adopted ordinary meaning), though the court’s claim construction can be a basis for amendment, if construction “changed the rules of the game” (see ePlus v. Lawson, where accused infringer failed to show that claim construction changed rules of the game, to justify new invalidity theories in expert report).
  • Don’t use delays or problems with the other side’s infringement contentions, as reason for delays or problems with invalidity contentions: e.g. MyMedicalRecords v. Walgreen (“Two wrongs don’t make a right”) [TODO: relate this to LPRs or FRCPs; what about ICs as “admission ticket” to discovery etc.?]
  • Having chosen one prior-art reference instead of another, due to limits on the number of prior-art references (imposed either by the court or by the party’s good sense and desire to focus its case), it will be difficult to undo that choice: e.g. Vehicle IP v. Werner Enterprises.
  • More broadly, strategic/tactical decisions, later regretted, do not create “good cause” for amendment: e.g. Wonderland v. Thorley; Joao Bock v. Online Resources (accused infringer strategically chose not to file any invalidity contentions at all before deadline; “strategic” in that originally based on assessment of the merits of P’s infringement case).
  • Be diligent in seeking out prior art, and timely in disclosing that prior art: e.g. Farstone v. Apple; Warner Chilcott v. Mylan; Par Pharmaceutical v. Takeda Pharmaceutical (knew of references, waited 4 months); Innovative Display v. Acer (lack of diligence in discovering new references, failed to show why important to case). [TODO: what are factors considered in determining whether diligent?; not only elapsed time, but reasons for elapsed time — put in section on amending/supplementing; TODO: what are factors considered in determining good cause to amend — put in section on amending/supplementing; partly depends not on time elapsed, but how much time left, e.g. before close of discovery, before SJ hearing, etc.; e.g. Catch a Wave v. Sirius XM]
  • Remember that, since prior art by definition was publicly-accessible at the relevant time, it should generally be possible to find it, and an argument about undiscoverability will often fail; this is especially true of prior art that was well known at the relevant time: e.g. GeoTag v. Frontier (deny leave to supplement contentions to include prior-art references, failed to show diligence where the references were “what might well be described as the largest and most popular internet systems of the 1980s and 1990s”, e.g. AA SABRE, Prodigy, AOL, CompuServe).
  • Of course, once-public materials may now only be available through discovery, and even for publicly-sold products, manuals and other documentation were not always on the open market: possibly e.g. DCG v. Checkpoint (product manuals produced by patentee), but see Brilliant Instruments v. Guidetech (deny leave to amend to assert device as prior art; had earlier knowledge of device but failed to investigate it; said that the device was trade secret, but patent openly refers to the device as prior art).
  • Especially for the on-sale bar, discovery may be required into prior art (e.g. manufacturer’s process), but this discovery can be denied if the requester didn’t earlier disclose whatever it already knew about the prior art (albeit without all details): e.g. Nippon Steel v. Posco (denying discovery into patentee’s prior art manufacturing process because the accused infringer in its invalidity contentions had not identified that process as being prior art it would rely on); this sounds like a Catch-22, but but presumably the invalidity contentions could have said something, as a hook/placeholder for the discovery request.
  • Once prior art is produced in discovery, don’t wait months to examine it: e.g. JS Products v. Kabo Tool. [TODO: see also source-code cases.]

The underlying basis for these decisions is that the parties and the court should be able to rely on the contentions to frame the case, albeit obviously with some adjustments and the occasional surprise; reliance rather than surprise should be the norm. A phrase from one case — the “parties had been operating under the current invalidity contentions for over a year” — is a good way of putting it: the parties operate under the infringement and invalidity contentions. [TODO: also put this in xxx on purposes/goals of claim charts.]

By the same token, invalidity contentions are generally deemed “final” invalidity contentions, unless there is a basis to amend or supplement under the rules: see LPRs on “final” contentions, e.g. E.D. Tex. LPR 3-6(a): “Each party’s ‘Infringement Contentions’ and ‘Invalidity Contentions’ shall be deemed to be that party’s final contentions, except as set forth below,” setting forth an allowed amendment, without leave required, following the court’s claim construction, with all other amendment requiring leave of the court for good cause.

See also separate discussions below of obviousness and enablement.

Method claims

A single patent often contains several independent claims of multiple types — such as an apparatus claim, a method claim, and a medium or system claim — all for the same single invention (it must be a single invention, as otherwise the patent would have been subject to a “divisional”).

Further, multiple claims of different types are often asserted against the same defendant in patent infringement litigation, and these claims may contain nearly identical wording, apart from the preamble, e.g. “An information processing apparatus, comprising…” vs. “A method of processing information, comprising…”. Even with nearly-identical wording, however, there are significant differences between apparatus and method claims, which call for different treatment in a claim chart.

In particular, note that a method claim is used, whereas an apparatus claim is made, sold, etc. In other words, method claims (and infringements of them) are dynamic, whereas the other claim types are somewhat static. [TODO: but an apparatus can also be used, as well as made or sold.]

Infringing acts asserted in a claim chart should follow this; it makes little sense, for example, to assert importation of a method claim (see xxx on ITC cases). Because copy and paste are frequently used to prepare claim charts, with attention perhaps lavished on claim 1 followed by pro forma grinding out of other claims, it is common for claim charts to contain nonsensical mismatches between claim type and infringement type, that were copied and pasted from claim 1. [TODO: really nonsensical to assert e.g. sale of a method?; what about sale of a service?]

The word “instrumentality” makes it easy to copy and paste infringement assertions from one row of a claim chart to another, without bothering to change “apparatus” to “method”, for example, and other generic words such as “embodies” or “practices” can also make do for more specific terms such as “makes” or “uses”, but such useful generic words can also help conceal the different types of analysis required. [TODO: show]

Some miscellaneous points about method claims:

Especially for software patents, a method claim may mostly be carried out by customers (“end users”) of the product, in which case the maker of the software is liable for indirect infringement (inducement). Note however that even when most use of the method claim will be by customers, there is likely some in-house use (tech support, debugging, testing, maintenance, etc.) and therefore direct infringement as well, albeit on a smaller scale. See xxx on indirect infringement.

If the accused instrumentality is capable of performing some method, but does not actually do so, e.g. “dead code” in software, this likely does not infringe a method claim, because the method is not being used, but it may still infringe a corresponding apparatus or medium claim, as the capability was made. See xxx on “latent code.”

A method claim is made up of steps. Usually the claim “comprises” one or more steps, in which another method infringes even if it also carries out additional steps. The order in which the steps are carried out is typically unimportant, except if the language of the claim imposes a particular order, by using words such as “then” or “followed by”. But see e.g. Mformation v. RIM. A single step in the patent claim may be satisfied by multiple steps in the accused method, and conversely multiple steps in the claim may be carried out by a single step in the accused method [cite].

For a method claim to be infringed, all steps must be carried out by a single defendant or by agents acting under the defendant’s direction and control [cite]. One interesting important exception, however, is that indirect infringement may include inducing multiple third parties to carry out different steps of the method; see Akamai v. Limelight.

Offshoring one or more steps can avoid infringement of a method patent (see xxx on patent territoriality), except if the overseas step is under the control and for the “actual beneficial use” of the US system; see NTP v. RIM.

A method claim may at first glance appear to require a step which is not being carried by the defendant, yet it is possible that the seeming requirement is not in fact a step in the method, but rather a mere precondition which must have been carried out at some point, but not necessarily by the defendant. Such preconditions are often signaled with the past-tense suffix -ed rather than with the active -ing, e.g. locked vs. locking, displayed vs. displaying. Also note preconditions referring to mere capabilities, e.g. “moveable” does not require that something be moved, merely that it be capable of movement; see xxx on “programmable.”

[TODO: also need a section for system claims, medium claims, e.g. medium requires some nexus?]

Multiple-party issues


Doctrine of equivalents (DoE)

The Local Patent Rules typically require that a patent owner indicate in its claim charts when it is asserting the doctrine of equivalents (DoE) and/or literal infringement. For example, E.D. Tex. LPR 3-1(d) states that information in the infringement contentions shall include:

“Whether each element of each asserted claim is claimed to be literally present or present under the doctrine of equivalents in the Accused Instrumentality”.

Note that the DoE assertion is on a specific per-element (limitation) basis, not global per-claim basis. See e.g. Finjan v. Sophos (grant motion to compel amended ICs, when DoE contentions used boilerplate language and sought to make function/way/result comparison to overall product rather than to individual claim limitations).

It is unfortunately common for claim charts, especially PICs, to contain “boilerplate” DoE assertions. For example, following a more-or-less detailed assertion of the literal presence of each claim limitation, with the goal of showing that D’s X satisfies the Y limitation of P’s claim, the chart will then state something along the lines of: “or, in the absence of literal infringement, limitation Y is present under the DoE, because X has the same or similar function as Y, uses substantially similar means as Y, and has the same or similar result to Y”.

In other words, the Function/Way/Result (F/W/R) test for the DoE is often recited as a mantra, rather than each of its three components being separately analyzed. In some cases, the party is fully aware of this, and, especially in its PICs, is merely hoping to create a boilerplate “placeholder” for a later serious assertion, should it become necessary, of the DoE. As seen in the discussion elsewhere of attempted amendments of claim charts to include DoE, this boilerplate placeholder is often unsuccessful. [TODO: give example here.]

At other times, the party even in its non-initial ICs is still stuck at a rote recitation of the F/W/R test, indicating that the author of the claim chart merely has a general sense of equivalence without being able to provide specifics. Note:

  • Function = purpose; role of this limitation in the claim as a whole
  • Way = implementation; means (of carrying out the function and generating the result)
  • Result = output

A failure to analyze how D’s X can be broken down into F/W/R and then mapped onto the F/W/R of P’s claim limitation may cast doubt on the same chart’s assertions of literal infringement, because it should be harder to establish literal infringement than equivalence, even though there is a defined framework for measuring equivalence and there is not such a framework for demonstrating literal equivalence; see xxx.

[TODO: this is too abstract; give a concrete example of what proper F/W/R analysis looks like (use the word “because”), after which it should be clear why inability to properly do F/W/R likely also means that the literal infringement analysis is superficial.]

While one can plead DoE in the alternative, by the time of non-initial ICs, a party should have committed to asserting either literal infringement or DoE. See e.g. Rambus v. Hynix, in which Rambus’s Final Infringement Contentions stated that to “the extent that any limitation is found to be not literally present, Rambus asserts that such limitation is present under the doctrine of equivalents” (“The Patent Local Rules require a limitation-by-limitation analysis, not a boilerplate reservation. The doctrine of equivalents exists to prevent ‘a fraud on the patent.’ It is not designed to give a patentee a second shot at proving infringement ‘[t]o the extent that any limitation is found to be not literally present.’ Rambus’s failure to comply with the Patent Local Rules provides ample, alternative justification for dismissing Rambus’s claims of infringement under the doctrine of equivalents”).

Some LPRs specifically require an F/W/R explanation for each DoE assertion. For example, W.D. Pa. LPR 3.2 requires for each element of each asserted claim, “if present under the doctrine of equivalents, the asserting party shall also explain each function, way, and result that it contends are equivalent, and why it contends that any differences are not substantial”. See also N.D. Ill. LPR 2.2(d), D.Utah LPR 2.3(d).

An additional method for demonstrating equivalence, from the Graver Tank and Warner-Jenkinson cases, looks at known interchangeability, and at evidence of copying vs. design-around. [TODO: say more about known interchangeability, in the context of facts re: welding from Graver Tank; lack of substantial differences.]

Note that equivalence is also a part of means-plus-function analysis under 35 USC 112(f); an allegedly infringing or anticipatory item matches the disclosed means for a functional limitation, when the item either literally matches the disclosed means or is equivalent. See means-plus-function below.

An aspect of DoE which receives a great deal of academic attention is prosecution history estoppel (PHE), which acts as a limit on the extra claim scope which can be created with the DoE. However, it is rare for PHE to appear in the context of claim charts. The basic principle behind PHE is that scope which the patentee gave up during patent prosecution, in order to receive the patent grant, should not be re-acquired when asserting infringement in litigation. This basic principle holds true even in a non-DoE context; accused infringers can use the file wrapper (prosecution history) to help keep patent claims within their proper scope.

Means-plus-function under 112(f)

Patent law generally cares more about structure than about purpose; this is one reason why preambles are generally not limiting: so long as the accused instrumentality does X, we typically don’t care if D is doing X to implement a floor wax, even if the preamble to P’s patent claim states that X’s purpose is to provide a dessert topping.

At the same time, it is sometimes convenient to speak in functional terms, and US patent law allows claim language which focuses on purpose/results/function rather than on way/implementation/means — but only so long as the functional claims language is backed-up by specific means disclosed in the specification. See 35 USC 112(f):

“An element in a claim for a combination may be expressed as a means or step for performing a specified function without the recital of structure, material, or acts in support thereof, and such claim shall be construed to cover the corresponding structure, material, or acts described in the specification and equivalents thereof.” [TODO: explain “for a combination”? show that both method and apparatus claims can employ means-plus-function.]

The important point for patent litigation is that, for such limitations, infringement will only be found when the specifically-disclosed means are present in the accused instrumentality, not merely when the claim’s function/purpose is present. This also applies to invalidity analysis. Thus, means-plus-function makes it more difficult to find both infringement and invalidity. [TODO: double-check]

Note “structure, material, or acts” (SAM) in 112(f). The need to identify SAM is somewhat different from the general requirement that, for each limitation, one show at least one “structure, process, algorithm, feature or function [SPAFF] of any accused product” (Connectel v. Cisco). [TODO: more on difference SAM v. SPAFF?]

Note “and equivalents thereof” in 112(f): means-plus-function is satisfied, not only when the identical structure, material, or acts (SAM) is found, but also when equivalents are found. Such equivalents are determined using the same tests as in the DoE (see above), particularly function/way/result (F/W/R), but an equivalent SAM is still regarded as literal infringement.

While 112(f) states “such claim shall be construed to cover…”, note that it starts off by referring to “an element in a claim,” and in fact it is limitations, rather than claims as a whole, which are or aren’t functional, and which therefore are or aren’t limited in scope by disclosure of “means” (implementation) in the patent specification. [TODO: show an example of a claim with multiple “means for…” limitations, and also with non-functional limitations.]

[TODO: above about means-plus-function limitations, not claims, might be wrong, because it is entire claim that is tested for presence of any structural support, i.e. it is absence of structural support, in claim (not just limitation), which makes for coverage under 112(f)? E.g. CCS Fitness v. Brunswick, “recites function without reciting sufficient structure for performing that function”.]

Functional language under 112(f) is usually created with the phrase “means for…”, but functional language can be anything which specifies what result is to be achieved, without specifying in the claim how it is to be achieved; see Williamson v. Citrix.

There are limits, of course, to what is seen as functional language: for example, “square root” specifies the result/output to be achieved, with no indication of how to go about doing so, but this function/part/element/step is so well known that a specific means/implementation need not appear in the specification, and likely any of the many possible implementations of square root, if found in the accused instrumentality, will be seen as meeting the claim limitation.

With the above prelude, what are implications for claim charts? Clearly, when comparing to a functional limitation such as one employing “means for…” language, the chart must compare the accused instrumentality or prior art, not merely with the claim language, but (as always) with the claim language as properly construed, which in this case means with the “corresponding structure, material, or acts described in the specification and equivalents thereof.”

This requirement is set forth in the Local Patent Rules, e.g. ED TX LPR 3-1(c), requires that infringement contentions include “for each element that such party contends is governed by 35 U.S.C. ç 112(6) [the pre-AIA version of 112(f)], the identity of the structure(s), act(s), or material(s) in the Accused Instrumentality that performs the claimed function”.

It is not always straightforward determining when an claim element is or is not governed by 112(f), and the LPR requirement that the IC author initially make this call (“… each element that such party contends is governed by…”) is interesting, since a claim element under 112(f) is narrower than one which isn’t. A party asserting infringement would prefer to not come under 112(f), as this limits them to the SAM disclosed in the patent specification. However, this does also help avoid some prior art. [TODO: is that really true?; example where means-plus-function made the difference in avoiding anticipation?]

Claim charts often breeze past the need to find a matching SAM for “means plus…” language. [TODO: cases; results; party can do this if it merely contends that elements are not governed by 112(f)?]

A common mistake is to identify a claim limitation as governed by 112(f), and to cite the SAM disclosed in the specification, but to then fail to find a corresponding SAM in the accused product or prior art, instead finding something that merely corresponds to the functional language of the claim. An example can perhaps be seen in the three-column claim chart below, from Bedrock v. Softlayer, in which the SAM has been spelled out as part of the court’s construction in the middle column, but the material in the right column appears to speak only to the unadorned functional claim language in the left column:


[TODO: show what a proper means-plus-function analysis would look like, weaving facts from accused product with the specifically-disclosed SAM.]

(As a further example of means-plus-function analyzed as part of claim construction, see the xxx chart in the section on the Joint Claim Construction Chart.)

As an alternative to placing the SAM disclosed in the specification in a third middle column of the claim chart, it can be placed in the left column, under the claim language, if properly set off for example with square brackets.

[TODO: include something on step-plus-function.]

Indirect infringement: charting inducement and contribution

Under 35 USC 271(b), “Whoever actively induces infringement of a patent shall be liable as an infringer.” Inducement, briefly, means encouraging or instructing another, knowing this will result in the other’s direct infringement.

Under 35 USC 271(c), whoever supplies a component constituting a “material part” of an invention, knowing the component to be “especially made” for the infringing use, and “not a staple article or commodity of commerce suitable for substantial noninfringing use,” shall be liable as an infringer. In other words, the supplied component must have certain characteristics in order to infringe.

Both inducement under 271(b) and “contribution” (as this component supply is called) under 271(c) have important knowledge and intent elements, but these will not be discussed here, except to the extent knowledge or intent would feature in a claim chart relating to indirect infringement, e.g. xxx. [TODO: contrast any per-limitation showing of knowledge or intent withcases where knowledge/intent are shown e.g. from advertising, Creagri v. Pinnaclife; Blue Spike v. Adobe?]

Importantly, there is no indirect infringement without one or more direct infringers. For example, the customers or users of a product may infringe a method claim, having been induced to do so by D’s sales and marketing of the device which, when used by the customer, carries out the method. Similarly, customers of a product may one missing part (e.g. a battery), and add it to a “component” (everything else, minus the batteries), and thereby infringe an apparatus claim directed at the entire ensemble (“component” plus battery). Note here that it is not the supply of the battery (a staple article of commerce), but rather of everything else, which is the component contributed by the indirect infringer.

The direct infringement could be the subject of a standard PIC or IC claim chart, with the elements of inducement or contribution shown in a non-chart portion of the contentions. However, there are several reasons why one might have a separate indirect infringement chart, or include facts regarding indirect infringement in the same chart as direct infringement:

  • While inducement is not on a limitation-by-limitation basis as such, both inducement and contribution are on a claim-by-claim basis (see LPRs below); given differences among multiple claims, where these differences depend on the limitations and not merely on a non-limiting preamble (including method vs. apparatus vs. system claim type), it may be safest to show that the alleged indirect infringer has induced as to each and every limitation, especially for method claims. [TODO: vague; give example.]
  • While contribution supplies a component, and therefore likely does not supply each and every limitation of a claim, demonstrating that the component is a “material part” of the invention is perhaps best done using a claim chart, to demonstrate which limitations the component does supply. Note that “material part of the invention” must be shown for at least one essential element or step for every asserted claim (see Arris v. BT). Conversely, parts which can do things outside the claim scope are unlikely “components” for 271(c) purposes (see Fujitsu v. Netgear).

The Local Patent Rules typically address indirect infringement by requiring (though not necessarily in claim chart form), “For each claim which is alleged to have been indirectly infringed, an identification of any direct infringement and a description of the acts of the alleged indirect infringer that contribute to or are inducing that direct infringement. Insofar as alleged direct infringement is based on joint acts of multiple parties, the role of each such party in the direct infringement must be described.” (N.D. Cal. LPR 3-1(d).)

While there must be direct infringement, P need not necessarily specifically identify the direct infringers, e.g. a list of all of D’s customers, especially up front. See e.g. DCG v. Checkpoint. This makes sense, as P will often be asserting inducement of a method patent precisely because there are too many direct infringers to handle separately. Curiously, it is possible to induce all steps of a method patent, even in the absence of a single direct infringer carrying out each and every step of the method; see Akamai v. Limelight. This is an exception to the rule that a single party must carry out (or be in control of carrying out, e.g. via agents) every step of a method patent (see above on joint/divided infringement, and on method claims).

However, there must be a definite showing of indirect infringement, e.g. by noting the names and model numbers of D’s devices, and then noting that these devices have users and/or customers. See e.g. Avago v. IPtronics. An indirect infringement case can easily fail for P’s insufficient showing of the seemingly-less-interesting direct infringement, which is easy to treat in too cursory a fashion as P is generally more interested in the indirect infringer as a “deep pocket” or direct infringer, and as it may seem counter-intuitive for P to be carefully telling D about acts of X, rather than P focusing on telling D on D’s own acts.

In many cases, a party has failed to assert indirect infringement until too late; as with DoE, one can’t “reserve a right” to amendment, as the LPRs require good cause (see above). However, it may work to broadly assert infringement in a way that suggests the later likelihood (and arguably provides some minimal notice to D) of a specific indirect infringement assertion; see e.g. Fenner v. Hewlett-Packard, differentiating Fenner v. Juniper). Similarly, in many cases a portion of the expert report relating to indirect infringement will be struck, as outside the scope of the party’s infringement contentions. [TODO: more on when court grants/denies amendment requests for, or motions to strike, possibly new indirect infringement contentions.]

There are several technical aspects of the “component” showing for contribution which may benefit from use of a chart. In particular, one of the tests for whether a component is “suitable for substantial noninfringing use” is whether an asserted noninfringing use merely derives from an add-on, packaging, or bundling. This easily arises in software cases. Essentially, the “suitable for substantial noninfringing use” requirement should come from what is sometimes called an “integrated” part of the accused product. See Lucent v. Gateway; i4i v. Microsoft; Fujitsu v. Netgear; Vita-Mix v. Basic. [TODO: okay, a chart, but why a limitation-by-limitation claim chart to counter assertion of substantial non-infringing use?]

[TODO: put something in section on multiple parties, on joint induced acts of multiple parties (Akamai v. Limelight), and not likely joinder issues, from P going after D for direct acts by 3P; though D will generally not want to pull in its customers; contrast cases where P goes directly after 3P customers, and their supplier intervenes.]

It is important to remember that if P can show that the infringing third party is acting directly under D’s control or direction, then the issue is direct infringement, rather than inducement or contribution. See xxx.

It is also important to remember that, even when customers are directly infringing a method claim, which indirect infringement by D, there is likely also some direct infringement by D who does in-house testing, tech support, maintenance, debugging, and so on. It is a rare company that does not “eat its own dog food.”

In importation actions before the ITC, indirect infringement can be used to address issues that arise when a vendor separately imports multiple parts (e.g. hardware and software) into the US, assembling them together only when importation is complete. See xxx.


An accused infringer, beside wanting to show non-infringement, will also want to show that the asserted patent is invalid. Ideally, the accused infringer will find a single piece of prior art on all fours with the asserted claims of the patent. See discussion of anticipation (lack of novelty) at invalidity contentions above. However, failing this it may find that all the limitations of the patent were disclosed, not in a single piece of prior art, but in multiple pieces of prior art. It will then try to assert that it would have been obvious for the person having ordinary skill in the art (PHOSITA) to combine this prior art at the time of invention. See 35 USC 103:

“A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains….”

Under Graham v. John Deere, obviousness is based on four factual determinations, each of which must be considered:

  1. the scope and content of the prior art
  2. differences between the prior art and the claims at issue
  3. the level of ordinary skill in the pertinent art
  4. objective indications of non-obviousness (“secondary considerations”), such as commercial success, long-felt but unsolved needs, and failure of others to make the claimed invention.

A showing of obviousness almost necessarily entails the use of multiple references, because if all the limitations of the claim were found in a single reference, anticipation under 35 USC 102 would instead be asserted. (As a rare exception, a single reference might be used to show obviousness, in combination with e.g. common PHOSITA knowledge.)

An obviousness claim chart uses multiple references, as in the example below from Pumatech v. Extended Systems:


Note the phraseology, “Claim 1 is invalid under 35 USC 103 on being obvious over [prior-art #1] in view of [prior-art #2].”

Since almost all inventions are comprised of previously-existing (and likely previously-patented) components, one cannot generate obviousness simply by combining the bits one wants from any combinations of prior art. A patented invention cannot simply be used as a template to go scavenger hunting through the prior art. Instead, the obviousness argument will have to be that the combination would have been obvious in the past, before the time of invention (pre-AIA) or before filing for the patent (AIA, see 103 above).

Further, the prior art cannot come from anywhere, but must come from the same field as the invention, or the “analogous arts,” or from the nature the problem to be solved.

On the other hand, the prior art need not be a patent or published patent application. Often, a gap between a prior art reference and the patent claim will be explained by pointing to general knowledge of the PHOSITA at the relevant time. This knowledge can be established by an expert declaration (so long as the expert remembers to speak to knowledge at the relevant time, not the present, and so long as the expert provides some basis for his opinion). It can even be a matter of common knowledge. [TODO: example, e.g. from PTO fill in gaps from common knowledge.]

The obviousness claim chart must establish that the PHOSITA would have been “motivated to combine” the prior art to arrive at the same combination as the patent claim. In the chart above, notice the statement: “One of ordinary skill in the art would be motivated to combine [prior-art #1] with [prior-art #2] or alternatively to combine [prior-art #1] with [prior-art #3] in order to efficiently represent and reconcile schedules containing recurring appointments.” However, the issue is not whether one today would be so motivated, but whether it existed at the relevant time. The phrase “teaching, suggestion, motivation” (TSM) indicates the need to find some pointer, existing at the time, that would have led the PHOSITA to the invention. Under KSR v. Teleflex, this need not be actual published or patented prior art (“in many cases a person of ordinary skill will be able to fit the teachings of multiple patents together like pieces of a puzzle”). Essentially, the test is now “flexible TSM.” [TODO: say something about “teaching away”?]

Note that 103 includes the clause “such that the claimed invention as a whole would have been obvious…” (emphasis added), indicating that there is a certain “holistic” analysis required in contrast to the usual emphasis in this book on disaggregated non-holistic analysis (see xxx). The difference is from showing that each of the elements or steps of the claimed invention would have been obvious. Finding a contemporaneous TSM to combine parts into a whole — the same whole represented by the patent claim at issue — largely takes the analysis from the parts to the whole. In addition, the claim chart drafter should consider (and if necessary explain in the chart) whether combining the previously-existed parts, per the contemporaneous TSM, would have resulted in an operable whole. [TODO: for example…]

The Local Patent Rules generally require that invalidity contentions include an obviousness explanation, or risk waiver of later assertions of obviousness. For example, N.D. Cal. LPR 3-3(b): “Whether each item of prior art anticipates each asserted claim or renders it obvious. If obviousness is alleged, an explanation of why the prior art renders the asserted claim obvious, including an identification of any combinations of prior art showing obviousness…”

There are many cases which include defendants denied their requests to amend their invalidity contentions to include an obviousness theory, for failure to include obviousness in their initial ICs. It makes sense that obviousness would initially be neglected in many initial charts, since obviousness is (despite being “the very heart of the patent system,” the “crowning glory of patent law,” and the like), a somewhat last-ditch effort in patent infringement litigation. A defendant generally wants to make its case with the much simpler showing of anticipation (lack of novelty), and it will generally in its initial invalidity contentions put forward several pieces of prior art that it views as knocking validity out of the park. It is only later, when it finds out that each of these prior-art references came short of anticipating of the patent claims, that it will reach for obviousness. (See however MedImmune v. BioPharma, in which curiously the original invalidity theory was obviousness, which was held as not waiving anticipation, in contrast to the usual situation in which only asserting anticipation may waive obviousness.)

At the same time, the case law indicates that where obviousness is newly asserted to cover already-disclosed prior art, an amendment may be granted; see e.g. Oracle v. Google (grant leave to add obviousness to ICs, based on four prior art references previously disclosed re: anticipation). It is when, to make a case for obviousness, the defendant needs to add previously-undisclosed prior-art references, that it more likely risks not being able to assert obviousness; see e.g. Verinata Health v. Sequenom. [TODO: case in which D thought it had anticipation via inherency, question whether this was sufficient notice of a possible obviousness theory; Semiconductor Energy La. v. Chi Mei Optoelectronics]

The bottom line is to establish the obviousness theory as early as possible, including of course the necessary prior-art references and asserted combinations, even if it seems superfluous to a “better” anticipation theory, because failure to assert essentially the whole obviousness theory — not just a placeholder (see e.g. DoE above), but largely the whole theory with necessary references and combinations — can result in waiver. See e.g. Better Bags v. Illinois Tool Works (obviousness struck from expert report, because not included in party’s ICs); Changzhou Kaidi Elec. v. Okin (preclude expert testimony as to new combinations not properly identified in party’s ICs).

It appears common to put forward prior-art references in a way that at least suggests a theory of obviousness, but without an identification of the combinations showing obviousness. See e.g. Saffran v. Johnson & Johnson (strike invalidity contentions, albeit with leave to amend, since contentions failed to identify the specific combinations of prior-art references); Medtronic v. W.L. Gore (strike invalidity contentions even though prior-art references were disclosed, for failure to identify combinations). [TODO: or identify some combinations, but not ones which later wish to use.]

Note however the ICs may not need to disclose all possible motivations to combine the identified prior art; see e.g. Fujifilm v. Motorola.

A recurrent issue in cases on obviousness and claim charts is how explicit the chart must be (and as a consequence how large and repetitive) when there is a large number of prior-art references and a large number of possible combinations. Prior-art references can be organized into groups, with combinations established between the groups. See Rambus v. Nvidia; Avago v. Elan Microelectronics; Keithly v. [TODO: show and/or example what commonality is required of each group.]

While 103 requires that the obviousness analysis include secondary considerations, these (with the possible exception of commercial success) would not require the limitation-by-limitation analysis of a claim chart, and at any rate would more likely be introduced by the patent owner (see below). Showing commercial success may require a claim chart, as of course it is commercial success of the patent invention, as found in one or more patent claims, which must be shown; see xxx on working/practicing charts.

The patent owner’s validity contentions (see above), in addition to showing lack of anticipation by single prior-art references put forward by the accused infringer in the invalidity contentions, will also respond to any obviousness contentions.

In addition to showing, of course, that the selected prior-art references do not in fact supply the pieces needed for the other side’s obviousness analysis, the validity contents can also attack:

  • Analogous arts, problem to be solved: Are the prior art references from appropriate fields, such that the PHOSITA would have looked there?
  • Flexible TSM: Has D really shown a teaching, suggestion, motivation from the relevant time, or has it merely waved its hands and chanted a “TSM” mantra?
  • As a whole: Has D really shown that the entire claim was obvious, or merely individual parts?
  • Operable whole: When the parts are put together, does it yield something that works for the purpose of the claim? [TODO: case?]

The patent owner will also want to introduce secondary considerations pointing away from obviousness. As indicated above, these generally do not require a claim chart, as they apply to the patent or claim as a whole, without requiring limitation-by-limitation analysis. However, as noted above, showing commercial success may require a working/practicing chart. A “nexus” between the commercial success and practice of the patented invention — e.g. commercial success is not attributable to a celebrity endorsement — should also be shown. Another secondary consideration, acquiescence by others in licensing, may be assisted by showing that others are practicing what they have licensed, as opposed for example to its inclusion, in passing, as part of a large portfolio.

[TODO: note 4-column obviousness chart format in appendix to Fox, U.S. Patent Opinions and Evaluations (2010)]

“Alice & Bob” chart (subject-matter invalidity)

Assertions of invalidity under 35 USC 101 (subject matter) were not common to patent litigation, until the time of cases concerning computer-implemented business methods, culminating in Alice v. CLS Bank. Alice has generated a new type of claim chart, the “Alice & Bob chart,” in which claim limitations in the left column are juxtaposed with acts performed by Alice, Bob, and Carol in the right column, to show that the patent claim is directed at a commonplace human activities which are not only well known but, more important, are abstract ideas that merely happen to be implemented on (i.e. have no necessary connection to) a computer. For example, from Walker Digital v. Google:


The Wikipedia entry for “Claim chart” has surprisingly useful coverage of these charts, including a second example by Prof. Richard Stern, based on Intellectual Ventures I LLC v. Manufacturers & Traders Trust:


Non-infringement contentions (NICs), including declaratory judgment (DJ)

The earlier LPRs section noted that many districts do not explicitly require non-infringement contentions (NICs), nor preliminary NICs (PNICs). Those districts whose LPRs do require NICs either require a formal admit/deny for each claim limitation spelled out in the PICs or ICs, or permit focusing only on those limitations being denied. Repeating the examples given earlier:

Admit or deny: “For each Asserted Claim against that party, a chart stating whether the party admits that that element is present in the Accused Device or contends that it is absent from the Accused Device. If the party contends that an element is absent from the Accused Device, it shall set forth in detail the basis for that contention.” (W.D. Wash. L.P.R. 121(a))

Assertion of absence: “A detailed description of the factual and legal grounds for contentions of noninfringement, if any, including a clear identification of each limitation of each asserted claim alleged not to be present in the Accused Instrumentality” (D. Nev. L.P.R. 16.1-8(a))

Those districts without an explicit NIC requirement, e.g. N.D. Cal., generally have requirements relating to NICs where they cover declaratory judgment (DJ) actions, in which the plaintiff is an accused infringer (albeit generally accused in an out-of-court context, such as licensing negotiations), seeking a court ruling that P does not infringe D’s patent (which in any event P asserts is invalid). In DJ cases, P is possibly in a position of formally showing non-infringement before a formal assertion of infringement. See e.g. N.D. Cal.:

“In all cases in which a party files a complaint or other pleading seeking a declaratory judgment that a patent is not infringed, is invalid, or is unenforceable, Local Patent Rules 3.1 [infringement contentions] and 3.2 [document production] shall not apply unless and until a claim for patent infringement is made by a party. If the defendant does not assert a claim for patent infringement in its answer to the complaint, no later than fourteen (14) days after the defendant serves its answer or fourteen (14) days after the Initial Rule 16 Conference, whichever is later, the party seeking a declaratory judgment of non-infringement, invalidity or unenforceability shall serve upon each opposing party its Disclosure of Non-Infringement, Invalidity and Unenforceability Contentions that conforms to Local Patent Rule 3.3 [invalidity contentions], and produce or make available for inspection and copying the documents described in Local Patent Rule 3.4.” (N.D. Cal. LPR 3-5(a))

In other words, there are 14 days to flush out an infringement chart from the DJ D, after which P shall serve its non-infringement chart (even if DJ D has not served an IC).

See also xxx below on charts used in licensing negotiations, on P’s actions (including presentation of a claim chart) sufficient to create a case or controversy for DJ.

Apart from the DJ context, in districts without formal NIC requirements, a patent-holder can use a contention interrogatory to elicit an accused infringer’s NICs. See e.g. Round Rock v. Dell (EDTX); Oasis Research v. Advrive (EDTX) (both cases rejecting argument that contention rogs were premature).

Where a NIC or PNIC is required, it cannot be merely conclusory, or else the accused infringer may be deemed to have admitted. See e.g. Agri-Labs v. TapLogic; Vincent v. Dux.

Court complaints about plaintiffs whose charts “mimic” or “parrot” the claim language are mirrored in complaints about defendants whose NICs do the same, e.g. “Product X does NOT” followed by the words of the limitation, with no further explanation.

Asking for more from an accused infringer, one who believes its product simply doesn’t do anything like the asserted patent claim, may appear to be asking to prove a negative, or explain an absence. However, at least one court was not persuaded, telling D to “meaningfully explain in detail why such element is not present literally or under the doctrine of equivalents” (Agri-Labs v. TapLogic). Presumably, language such as “… because Product X instead uses…” will sound more explanatory than “… because Product X entirely omits this step of its process…”.

Explaining the complete absence of a claim limitation is easier for negative limitations, e.g. “absence of noise” or “substantially without noise,” as D need only point to some substantial presence for the supposed-to-be-absent limitation (so long as it is present in a relevant part of the product).

For means-plus-function (and other functional) limitations under 112(f), D should explicitly point to the patent’s disclosed means for carrying out the functional limitation, rather than merely parroting the “means for…” language of the limitation, and should note how its product differs from the specifically-disclosed means.

To emphasize that NIC requirements are not the mirror image of IC requirements, it is not strictly necessary that the NIC chart each and every limitation, because whereas the patent-owner is faced with an AND expression in which it must show each and every limitation in order to show infringement, accused infringers only face and OR expression, in which the absence of a single limitation is sufficient to show non-infringement. Even in districts where the LPRs explicitly require an admit/deny format, accused infringers are presumably free to restrict their NICs to denied limitations, with all others being deemed admitted.

However, even if D need only show that a single limitation is missing, it generally will not want to stop there in its NICs, especially as amendment rules can be applied to NICs as to any other claim chart. See e.g. BreathableBaby v. Crown Crafts (strike amended NICs; good cause not automatic to amend NICs whenever judge’s claim-construction decision is adverse); Sloan Valve v. Zurn (refuse to consider NI theory that D first disclosed in SJ brief and had not included in NICs).

As a broader point, despite the name “non-infringement contentions,” these charts do not proactively assert non-infringement in some global sense. Instead, they respond to particular assertions of infringement. In a DJ case, DJ D may not yet have formally made these particular assertions, but because the court will take on an DJ action only when faced with an actual case or controversy under U.S. Const III(2)(1), DJ P likely has sufficient details of DJ D’s “case” to make out a NIC even in the absence of in-court IC. Note that under MedImmune v. Genentech, P need not have “reasonable apprehension of suit,” as previously required. [TODO: tie explicitly to whether P’s presentation of a claim chart is sufficient to give D DJ standing; see xxx below on charts used in licensing negotiations, including e.g. Sandisk v. STMicro.]

See LPRs: and .

Validity contentions

A patent owner who has asserted its patent against an accused infringer will likely face an assertion of invalidity, as well as non-infringement, by the accused infringer; see invalidity contentions above. The patent owner may rebut the invalidity contentions using a claim chart, or validity contentions. This claim chart compares the limitations of the asserted claims of the patent on the one hand, with the limitations ostensibly disclosed in the prior art in the invalidity contentions. The validity contentions can rebut assertions of anticipation (lack of novelty), obviousness, and statutory bars, as well as unenforceability. Here, rebuttal of anticipation is discussed; for obviousness and statutory bars, see relevant sections (unenforceability is not shown using a claim chart, and is not further discussed).

A patent granted by the US PTO comes with a presumption of validity, and the patent-owner need not re-prove validity at trial, apart from overcoming D’s clear and convincing evidence pointing to invalidity. Perhaps in part for this reason, few district LPRs require or otherwise refer to validity contentions. See e.g. Friskit v. RealNetworks: “There is nothing in the [N.D. Cal.] Local Rules or Fed.R.Civ.P. that require Friskit to provide a validity claim chart.” [TODO: context]

Several LPRs do require validity charts, in response to the invalidity chart, and provide helpful guidelines for the contents of the validity chart, which is largely a statement of absences. E.g. NJ LPR 3.4A: A “party defending the validity of the patent shall serve on all parties its ‘Responses to Invalidity Contentions’ which shall include the following: …(a) For each item of asserted prior art, the identification of each limitation of each asserted claim that the party believes is absent from the prior art…”; see also N. Ohio LPR 3.7.

An alternate is an admit/deny or merely deny format (in the latter case, anything not denied is deemed admitted), similar to that used in some non-infringement contentions (see below). E.g., N.D. Ill. 2.5: “The Initial Response to Invalidity Contentions shall contain a chart, responsive to the chart required by LPR 2.3(b)(3), that states as to each identified element in each asserted claim, to the extent then known, whether the party admits to the identity of elements in the prior art and, if not, the reason for such denial.”

[MORE: any cases with e.g. denied amendments for validity contentions.]

Contentions employing explicit or alternate claim constructions

All claim charts employ some claim construction, even if only implicitly by what is chosen in the right column to accompany the claim language in the left column. It is surprisingly rare to see claim charts explicitly showing the claim construction being applied. When a claim chart does show claim construction, it can either be as a square-bracketed or bulleted item in the left column [TODO: show], or made the subject of a separate column.

Because claim charts are typically generated before a Markman ruling, there is usually some uncertainty over which claim construction to use, and some charts provide multiple alternate claim constructions, generally from the two sides in the litigation. [TODO: show] This will typically be accompanied by text showing, not always convincingly, that the party’s ultimate position on infringement or invalidity still holds, whichever of the fiercely fought-over constructions is selected.

An excerpt from a sample three-column claim chart, from Bedrock v. Softlayer, is shown below; this chart should not be confused with a claim-construction chart as such (see below), as the chart here is intended to apply a claim construction, or a set of possible claim constructions, onto the accused instrumentality or prior art:


Note however that having included the court’s claim construction, this chart appears not to actually use the claim construction in the rightmost column. Instead, the walk-through of facts is in terms of the raw claim language, not the claim as construed by the court. In the example above, were the court’s construction being used, one would expect to see references in the rightmost column to e.g. “capable of containing two or more records”, a “pointer to the next record”, and so on. It is a very common error in claim charts, as in expert reports, even when diligently including a claim construction, to then forget to put it to use. A similar error occurs with means-plus-function limitations, where the claim chart will rarely show the means disclosed in the specification for the functional limitations, and will even more rarely explicitly show how the accused instrumentality maps onto the means (or its equivalent), rather than the bare function. See xxx.

See sections xxx and xxx for brief discussions of Markman claim-construction charts and the Joint Claim Construction Chart.

Enablement chart

To be valid, a patent must disclose sufficient information to enable a skilled person can make and use the invention. An accused infringer therefore can avoid the infringement accusation by showing that the patent is invalid for lack of “enablement.” In the words of 35 USC 112(a), “The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected…”

The Local Patent Rules (LPRs) don’t require an enablement or non-enablement chart. They do require that invalidity contentions (ICs) disclose the defendant’s enablement defense, but not necessarily in chart form. For example, N.D. Cal. LPR 3-3(d) requires, apart from the invalidity chart requirement in 3-3(c), disclosure of “Any grounds of invalidity based on … enablement or written description under 35 U.S.C. 112(1) of any of the asserted claims.” [TODO: written description (PTO may require claim chart with written description, to show new new matter); explain why no chart for definiteness (it’s a matter of the claim itself, not a comparison); and no best mode.]

D. Maryland LPR 805(1)(c)(iv) has more stringent requirements: “All grounds of invalidity other than anticipation or obviousness of any of the claims listed in the Claim Chart. This identification must be as specific as possible. For example, each party asserting an enablement defense must set forth with particularity what is lacking in the specification to enable one skilled in the art to make or use the invention, specifically citing information or materials obtained in discovery to the extent feasible.”

Whether employing a chart or not, determining enablement requires a comparison between the patent claims on the one hand, and the patent specification (i.e., the rest of the patent document, apart from claims) on the other hand. Such a comparison is an exception to the general rule otherwise followed in this book, of focusing on comparison of the patent claims with something outside the four corners of the patent. Claim construction charts are another such exception (see xxx).

It makes sense not to require a full-blown enablement chart, as the defendant will likely focus on the patent’s ostensible failure to enable just one or two limitations, and the patent owner in response will likewise confine itself to those one or two limitations.

However, there is a good reason to generate an enablement-related chart, at least for internal purposes, to ensure that each and every claim limitation has been enabled. The exercise of producing a chart, and thinking through each and every limitation, may raise more questions than would otherwise be raised without a chart. Note that failure to make a specific enablement argument in a timely manner may preclude later assertion of that argument, though generally courts appear open to enablement-related amendments: see e.g. Kilopass v. Sidense (denied), Lexar v. Pretec (granted), Chemfree v. Walter (granted), Verinata Health v. Ariosa (granted).

To enable, the patent specification need not state what the person ordinarily skilled in the art (PHOSITA) would have known, or could have obtained from elsewhere. Enablement may call upon the knowledge of the PHOSITA at the relevant time (and an expert may later testify to such knowledge at the time). A patent may enable, even if requring some experimentation by the PHOSITA, so long as it is not “undue experimentation.”

On the other hand, even clear disclosures in the specification may be insufficient if they are incommensurate with the scope of the claim. For example, a single example may not enable a claim to multiple possible embodiments, and a disclosed species may not enable a claimed genus. See e.g. MedImmune v. PDL Biopharma.

Note that that the enablement issue can be used, not only to invalidate, but also to limit the scope of asserted claims to what was disclosed in the patent. [TODO: more.]

Statutory-bar chart (on-sale or public-use bar)

To be valid, a patent of course must be novel and non-obvious over prior invention by others. See invalidity charts above. However, the patent also cannot have been preceded for too long by the patent owner’s or inventor’s own activities. This encourages application for a patent, rather than unpatented practicing of the invention; 35 USC 102(b): “A person shall be entitled to a patent unless … the invention was … in public use or on sale in this country, more than one year prior to the date of the application for patent in the United States” (this was rewritten in the AIA, but with largely the same effect, including the one-year “grace period”). This is known as the statutory bar. Note that sale includes offers for sale.

Charting a patent’s violation, or not, of the public-use and/or on-sale bar will look little different from charting any other non-patent prior art (including public use, sale, or offer for sale) in an invalidity chart. Here, there is extra care required for the date of the public use, sale, or offer for sale (more than one year before filing the patent application), extra care to connect the activity to the inventor or owner of the patent, and attention to several extra conditions, noted below.

In exchange for the one-year grace period, in which the inventor/owner is allowed one year of sales or use which would constitute prior art if done by another, set of invalidating acts is broadened: [TODO: double-check against AIA]

  • Even private/secret use by the inventor/owner is prior art, if for a commercial purpose (see Woodland v. Flowertree; Lough v. Brunswick; Lockwood v. AA) [TODO: check]
  • Sale or offer by the inventor/owner, of a product made by a secret patented process, is prior art (see Auld v. Chroma Graphics)
  • The use or sale by the owner need not be enabling to constitute prior art [cite]
  • Parties to the sale or offer need not know of the embodiment of the invention [cite; this is also true of sale by third party?]
  • Public use of the invention need not disclose the invention to the public [cite]

A claim chart may be important here because the parties clearly can dispute whether what was sold, offered, or used, was in fact an embodiment of what later became the patented invention. P is in a position of denying its own working/practicing, possibly making fine distinctions (which can then be used to confine the scope of P’s patent claims when exercised against D). [TODO: note this contradiction in section on working/practicing; P clearly wants to deny working/practicing more than 1 year before filing.]

The nature of the item sold or used can be tested on a limitation-by-limitation basis, comparing the sale/offer/use on the other hand, with the patent claim on the other. The legal test has evolved from a “substantially complete” test rejected in Pfaff v. Wells, to a “ready for patenting” test, i.e., based on what was sold or used, was the invention ready for patenting? [TODO: explain better; was sale/use based on drawings sufficiently detailed to enable (Robotic v. View Eng’g)?; see Space/Loral v. Lockwood.]

[TODO: maybe simplify this by noting that the sold/used item can be employed in anticipation or obviousness analysis, which is where item would be checked against the claim limitations.]

The LPRs address disclosure requirements relating to the on-sale and public-use bars, e.g. N.D. Cal. LPR 3-2(a) requires that the patent owner produce to the alleged infringer documentation regarding “sale of or offer to sell, or any public use of, the claimed invention prior to the date of application for the patent in suit. A party’s production of a document as required herein shall not constitute an admission that such document evidences or is prior art under 35 U.S.C. ç 102.” Note that the production is not limited to use or sale more than one year before filing.

See also e.g. N.D. Cal. LPR 3-3(b): “Prior art under 35 U.S.C. ç 102(b) shall be identified by specifying the item offered for sale or publicly used or known, the date the offer or use took place or the information became known, and the identity of the person or entity which made the use or which made and received the offer, or the person or entity which made the information known or to whom it was made known.”

A sale or use need only anticipate a single embodiment of the patent, to render the entire patent subject to the statutory bar. [TODO: is this right?; aren’t on-sale and public-use bars on a per-claim basis rather than affecting entire patent, just as invalidity is per-claim?]

Remember that if the sale, offer, or public use is by a third party, the effect is the same as with anticipation by non-published prior art (which must have been publicly accessible); see invalidity chart above.

[TODO: more on rebutting statutory bar: apart from “it was sold or offered,” etc., or that date is wrong, using claim chart to show that what was sold/used was not the patented invention, and e.g. why wasn’t ready for patenting.]

[TODO: see point in Part IV that P’s working/practicing chart is just one step away from becoming D’s statutory bar chart – if what P touts in the working/practicing chart can be extended back in time to before the 1-year grace period.]

Working/practicing chart

What might be called a “practicing chart” or “working chart” is suggested by some LPRs, e.g. ND CA 3-1(g): “If a party claiming patent infringement wishes to preserve the right to rely, for any purpose, on the assertion that its own apparatus, product, device, process, method, act, or other instrumentality practices the claimed invention, the party shall identify, separately for each asserted claim, each such apparatus, product, device, process, method, act, or other instrumentality that incorporates or reflects that particular claim.”

The LPR as written merely requires that the patent-holder identify on a per-claim basis each practicing instrumentality. While it is not enough to merely state that the product works or practices the patent as a whole, as opposed to a particular claim, at the same time a limitation-by-limitation exegesis is apparently not required. However, many patent holders will want to show, on a per-limitation basis, that it has a basis for stating that the identified instrumentality practices or works the specific patent claim.

What is this LPR referring to, and why would a patent-holder rely on practicing its own invention? The US, like most (all?) modern patent systems, does not have a working or practicing requirement (though this was the very origin of patent systems in 15th century Venice). This is why, for example, the US has the phenomenon of the “non-practicing entity” (NPE), or “troll,” which may include for example a university licensing department.

But there are nonetheless several circumstances in which a patent holder needs, or benefits by, pointing to its own practicing of its patent, and in this limited sense, the US does have a working/practicing requirement. The “for any purpose”, to which the LPR alludes, includes:

  • To show non-troll/non-NPE status, when requesting an injunction or other equitable relief, under eBay v. MercExchange; this can include products under development (see e.g. Mediatek v. Freescale).
  • To show commercial success of an embodiment of patent, as a secondary consideration for non-obviousness; P should show, on limitation-by-limitation basis, that the product embodies at least one claim of the asserted patent (as well as showing a “nexus” between the commercial success and the patented feature).
  • Perhaps, as to another non-obviousness consideration, acquiesence by others in licensing, it may be useful to show that licensees are in fact practicing the patent, i.e. they haven’t merely acquired it unthinkingly as part of a large portfolio; note that practicing/working by licensees may qualify as one’s own practicing/working (see ITC domestic industry section below).
  • To show technical ability to meet demand, for lost profits calculations; this could also be shown by past practicing of the patent, even if not current.
  • Perhaps most important, though not covered under the LPRs, to show “Domestic Industry” in an ITC importation action; see below.

In some circumstances, P need not show that it practices/works the asserted patent, but only that it practices/works something else, with which D is allegedly able to compete because of D’s infringement of P’s (otherwise non-worked) patent. See Rite-Hite v. Kelley.

When stating that its product A works, practices, or embodies its own patent, P must be careful that A does not predate the filing of its application for the patent by more than one year, or else it will have handed D an on-sale and/or public-use bar. [TODO: patent owner may be in position of denying that sale/use of its product was sale/use of something that practiced/worked/embodied the claimed invention; see on-sale and public-use bars above.]

Possibly, the accused infringer may point to the patent-holder’s own product, ostensibly covered by, embodying, or practicing the patent, as part of a “reverse DoE” type of argument, along the lines of: P says its product A embodies its patent; P says our product B infringes the patent; but B (e.g. a floor wax) and A (a dessert topping) have nothing in common (see xxx). [TODO: also note this in product-to-product comparisons.] More generally, D can use P’s mapping of its patent limitations onto P’s own product, with which P is presumably intimately familiar, as a gauge for P’s mapping of the same limitations onto D’s product with which it is presumably less familiar. See xxx on proactive use of the other side’s claim charts.

Another way of saying that P works/practices its patented invention is that P’s product “embodies” one or more claims of the patent. The concept of embodiment (not to be confused with enablement, nor with preferred embodiment) is somewhat counter-intuitive, and is covered towards the end of this book.

[TODO: see point in Part IV that P’s working/practicing chart may tread dangerously close to becoming D’s statutory bar chart, if D can push the dates for P’s touted limitations can be pushed back beyond the 1-year grace period.]

Domestic industry (DI) chart in ITC importation actions

As part of an importation action before the International Trade Commission (ITC; see below), the ITC expects a claim chart showing the patentee’s own practicing/working of the invention in the US; ITC Rules 210.12(a)(9)(ix):

“A showing that an industry in the United States, relating to the articles protected by the patent exists or is in the process of being established. The complainant shall make such showing by appropriate allegations, and when practicable, by a chart that applies an exemplary claim of each involved U.S. patent to a representative involved domestic article or to the process under which such article was produced…”

This “domestic industry” (DI) chart is a major type of working/practicing chart (see above). The ITC is not going to block importation of an item, if there is not some harm to domestic revenue or jobs. Domestic industry has two prongs, economic and technical. The economic prong requires e.g. investment in plant and equipment, employment, R&D, or licensing. The DI claim chart meets the technical prong, and shows that the domestic industry (from the economic prong) utilizes the infringed patent.

In the ITC rules, note “exemplary claim of each involved U.S. patent” — only a single claim from each asserted patent need be charted. Contrast e.g. PICs and ICs in which each and every asserted claim must be charted, or incorporated by reference. The ITC is sometimes presented with enormous charts created by overzealous associate attorneys concerned that all but one claim will be found not practiced. That said, if there is a legitimate concern that one claim might reasonably be found not to practice the invention, then inclusion of a substantially different claim is in order.

Note that the charted claim need not be one of the asserted claims; see e.g. 337-TA-432. [TODO: is DI ever relevant, to block importation of D’s product X1 which infringes P’s patent X2, when P’s DI does NOT practice X2, but instead practices patent Y, and when D’s X1 does compete with Y, similar to RiteHite analysis lost-profits?]

Also note in the ITC rules that the domestic industry “exists or is in the process of being established.” Working/practicing entirely in the past will not suffice (and further might trigger the statutory on-sale or public-use bars; see above). Curiously, the domestic working or practicing of the patent can be by a third-party licensee, even by an unwilling licensee, who lost an earlier infringement case [cite].

To block importation, it must of course also be shown that the imported item infringes the patent, and that the item is indeed imported, i.e. made outside the US and then sold or used inside the US; see below on claim charts dealing with territoriality.

As with working/practicing charts generally (see above), there is a useful interaction between P’s DI examples on the one hand, and infringement and invalidity on the other hand. For example, if P asserts that its product X is DI for its patent Y, and D asserts that an older product Z is prior art to patent Y, P cannot deny the relevance of features of Z that P itself has used to demonstrate DI for product X. [TODO: make X/Y/Z here consistent with X1/X2/Y above, or make simpler; also show how DI can help restrict infringement scope.]

Charting importation (ITC) and exportation

Patent law is territorial: it does not restrict acts happening outside its territorial jurisdiction. It is not infringement of a US patent when the invention is made and used in Germany, for example.

Further, it is not infringement when some limitations of a claim are made or used by one party in the US, if at least one limitation is instead made or used by a different party in Germany, unless with the “control and beneficial use” of the US party (see NTP v. RIM). Apart from one party as the agent of another party, there is no such thing as split or divided or joint infringement, much less across territorial borders. See xxx above on so-called split/divided/joint infringement, and contrast indirect infringement via inducement/contribution.

However, US patent law does address both import and export of items, including items made by a patented process. 35 USC 271(f) and (g). The word “items” should alert the reader to the issue of whether processes (corresponding to method claims, rather than apparatus/device claims) or signals/packets themselves can be imported or exported. See e.g. ClearCorrect v. ITC (no authority over electronic transmissions of digital data into US). [TODO: more on products produced by patented processes, including 35 USC 295 (presumption); discuss relation to e.g. software, e.g. is an API a method for producing a software product?]

Claim charts regarding import or export will largely match those in entirely-domestic scenarios, except that the chart may need to address the geographical location at which certain limitations occur. While geographical location can often be handled in some filing other than the chart, there are situations in which the location relates to one or more particular limitations, rather than to the claim as a whole, or in which all the limitations are not “put together” until after crossing a border. The latter scenario is addressed by 35 USC 271(f) regarding “components.” Such components correspond to a limitation or sublimitation of a claim, and can be shown in a claim chart. [TODO: show how this would be reflected in a claim chart; note any 271(f) case with chart; possibly Fenner v. Juniper, but not re: export?]

Import of a patented item can be taken before the International Trade Commission (ITC) — which is US federal agency not an international body — for an investigation and possible action under Section 337 of the Tariff Act; note that this differs from an action in federal court under 271(g), which also covers imports. 19 USC 1337(a)(1):

“… the following are unlawful, and when found by the Commission to exist shall be dealt with, in addition to any other provision of law, as provided in this section: … (B) The importation into the United States, the sale for importation, or the sale within the United States after importation by the owner, importer, or consignee, of articles that: (i) infringe a valid and enforceable United States patent or a valid and enforceable United States copyright registered under title 17; or (ii) are made, produced, processed, or mined under, or by means of, a process covered by the claims of a valid and enforceable United States patent.”

ITC domestic industry charts were discussed above; here, ITC infringement charts are discussed, to the extent different from other claim charts. As shown earlier at xxx, ITC Rules 210.12(a)(9)(viii) require “A showing that each person named as violating section 337 of the Tariff Act of 1930 is importing or selling the article covered by, or produced under the involved process covered by, the above specific claims of each involved U.S. patent. The complainant shall make such showing by appropriate allegations, and when practicable, by a chart that applies each asserted independent claim of each involved U.S. patent to a representative involved article of each person named as violating section 337 of the Tariff Act or to the process under which such article was produced.” [TODO: burden is on complainant to show why a claim chart is not “practicable”? any cases? what would make a chart not practicable?]

Note that while each asserted claim must be charted, on the other hand, it need only be applied to a “representative involved article” of each defendant. As the Section 337 Patent Investigation Management Guide observes, “This can still leave a large body of uncharted water to navigate”:

“Under existing practice, a complainant may satisfy its pleading obligation by providing a claim chart for just one model of one of the categories (e.g., an accused mobile phone) without having to disclose how the claim reads against any other models within that category or any other product category (e.g., camcorder) even though the two product categories may use entirely distinct technologies. As a result, complainant may then seek discovery over a far-ranging list of product categories without ever having had to have provided a specific allegation as to any product in those categories.”

As also noted earlier, the ITC Rules ask for drawings keyed to the claim chart: “(x) Drawings, photographs, or other visual representations of both the involved domestic article or process and the involved article of each person named as violating section 337 of the Tariff Act of 1930, or of the process utilized in producing the imported article, and, when a chart is furnished under paragraphs (a)(9)(viii) and (a)(9)(ix) of this section, the parts of such drawings, photographs, or other visual representations should be labeled so that they can be read in conjunction with such chart…”. Contrast this to one practice, seen at xxx, of producing almost entirely graphical claim charts, which attempt to use arrows from illustrations in the chart’s right column pointing to claim limitations in the left column. Here, while the illustrations are separate from the claim chart, the requirement that the illustration be labeled for reading in conjunction with the claim chart should, ideally, result in a very specific picture of where each claim limitation is located.

While it is possible to produce a claim chart showing infringement, in the same way one would in federal court, and separately assert the importation/sale for importation/sale after importation of the infringing item in a separate filing, a better practice is to include the specifically import-related facts directly in the claim chart. Note from xxx that the claim chart in any event needs to be specific about the infringing activity (making, using, selling, offering for sale), as this interacts with the claim type (apparatus, etc.), and especially if the chart is asserting a product made by a patented process, and details of the specific infringing activity can be combined with importation details. The chart would show two different geographical locations: one for origin of the item outside the US, and then sale or use inside the US (which may include “downstream” products or parties). [TODO: explain this better, and show an example.]

[TODO: component import issues, e.g. where all components for infringing are made abroad, then assembled inside US? or all components but one abroad, with last made in US?; especially with combined hardware/software products, where software is put on the hardware in the US, e.g. via OTA update. See Suprema v. ITC (using indirect infringement at ITC).]

Responses: ITC Rule 210.13(b)(1) states, “If it is asserted in defense that the article imported or sold by respondents is not covered by, or produced under a process covered by, the claims of each involved U.S. patent, a showing of such noncoverage for each involved claim in each U.S. patent in question shall be made, which showing may be made by appropriate allegations and, when practicable, by a chart that applies the involved claims of each U.S. patent in question to a representative involved imported article of the respondent or to the process under which such article was produced…” [TODO: do more respondents tend to view the chart as not practicable, especially given rapid timeline of ITC cases?; Section 337 Patent Investigation Management Guide states, “While Commission Rule 210.13(b)(1) ‘encourages’ respondents to include with their responses non-infringement claim charts, in practice many respondents do not do so.” Was this re: a pre-amendment version of the rule?]

This “noncoverage” chart is essentially non-infringement contentions, discussed at xxx above. As with the complaint, the chart need only show a “representative” article or process. Presumably, the chart could also show that the accused article is not imported, e.g. it is made in the US, or one or more limitations are met within the US. [TODO: more on how ITC treats cross-border production, and how this would be reflected in a limitations-based chart.]

Design patent chart

A design patent covers the ornamental design of an item; 35 USC 171(a): “Whoever invents any new, original and ornamental design for an article of manufacture may obtain a patent therefor, subject to the conditions and requirements of this title.”

A design patent does not have claims, and instead features one or more drawings.

Further, the test for infringement of a design patent is holistic, covering the design as a whole (its visual impression), not individual design elements. See Oddzon v. Just Toys, citing Gorham v. White; see also Amini Innovation v. Anthony.

Contrast the point made elsewhere in this book that infringement and invalidity are not holistic exercises, being based instead on parsing the asserted claims into limitations, with comparisons performed on a separate limitation-by-limitation basis, this indeed being a major reason for using a chart, with separate rows for each limitation, as a way to analyze and disclose patent infringement and invalidity.

Interestingly, the holistic manner in which design patent infringement is to be assessed is also a reason to use a chart. See Crocs v. ITC (598 F.3d 1294 (CAFC, 2010)), in which the ITC “placed undue emphasis on particular details of its written description of the patented design. Those details became a mistaken checklist for infringement. Without a view to the design as a whole, the Commission used minor differences between the patented design and the accused products to prevent a finding of infringement. In other words, the concentration on small differences in isolation distracted from the overall impression of the claimed ornamental features. [] The proper comparison requires a side-by-side view of the drawings of the ‘789 patent design and the accused products.” (Emphasis added.)

In other words, a chart with the patent drawings on one side, and the accused product on the other side. The CAFC decision in Crocs includes such charts, for example: [TODO: this is not an infringement chart; Crocs v. ITC re: prior art and domestic industry]


The proper comparison is between the patent drawings and the accused product, not between commercial embodiment and accused product. This is the same point made earlier in the “Comparing What to What?” section. However, if there is no significant distinction between the patent and a commercial embodiment, one may compare this embodiment with the accused product; see e.g. Lee v. Dayton-Hudson.

The LPRs generally do not address design patent infringement. However, see the exception in NDNY LPR 3-1(c)(ii): “For each design patent or variety of plant patent claim that is alleged, a chart displaying each view of the design or variety of plant patent drawings and a view of the accused design or variety of plant from every available angle for all embodiments.” [TODO: what mean, “all embodiments”?]

While “every available angle” may appear analogous to the design-patent-infringement requirement earlier referred to as “each x3” (each limitation of each claim, for each instrumentality), it likely refers to performing the comparison along every angle that was shown in the patent drawings. [TODO: what is the basis for this, phrase “each angle depicted in the design patent” from xxx?; note that Crocs patent D517,789 contains more drawings, depicting more angles, than shown above.]

The comparison itself employs the “ordinary observer” test, see Gorham v. White, with this observer giving such attention as a purchaser usually gives; following Egyptian Goddess v. Swisa, there is no longer an additional “point of novelty” test. The test is for sufficient resemblance such as to deceive, sufficient to induce a purchase supposing D’s design to be P’s. For example, were the patented and accused designs randomly mixed, could the ordinary observer restore them to their proper arrangement without careful, prolonged effort? (See Crocs v. ITC.)

As further examples, see the four design patents in Apple v. Samsung, and the following portion of a claim chart from Nike v. Wal-Mart (see


[TODO: design patents in validity contentions, e.g. NJ LPR: “design patents, where the party shall supply an explanation why the prior art does not anticipate the claim…”; see design-patent carve-outs in NJ LPRs.]

Expert claim charts

An expert’s claim chart (incorporated in the expert report) will not necessarily appear different in format from that of the party hiring the expert. However, there are important differences.

While the party’s charts are covered under the LPRs and e.g. FRCP Rule 11, the expert’s charts are not. Instead, the expert’s charts (as incorporated in the expert report) are covered under FRCP 26(a)(2)(B), as well as under FRE 701-705 and Daubert. (A party’s claim charts are unlikely to come under FRE 701-705 and Daubert, unless the chart is being used as evidence rather than as a pleading or rog response; see xxx above).

An expert’s claim chart is generally expected to provide an explanation as to why or how a particular claim limitation reads on a selected feature of an instrumentality or prior art; while many cases indicate that this is also expected of party claim charts (see e.g. xxx, and especially with regard to DoE and obviousness), an expert is generally obligated to do so under FRCP 26(a)(2)(B)(i) requiring that the expert report not only disclose the expert’s opinions, but also “the basis and reasons for them.” (This should not be seen as somehow trumped by FRE 705, allowing an expert to state an opinion “without first testifying to the underlying facts or data” as the latter refers to questioning at trial, and does not alter the expert-report obligations.)

Importantly, the expert’s claim “chart,” i.e. the expert’s mapping of claim limitations to an instrumentality or prior art, may be in textual rather than tabular format, especially if the expert has been hired by D, and is not disputing all limitations, such that the expert can focus in text on a few specific limitations.

In addition to the expert report incorporating a claim chart or its textual equivalent, an expert may also supply other claim-chart-related material, including:

  • Declarations or affidavits noting deficiencies in the other side’s chart; such a declaration would then be cited in a brief accompanying a motion to strike, an SJ motion, etc.
  • Support for the party’s representative instrumentality assertions (see “representative instrumentality chart” below)
  • Material to be cited in the party’s claim chart, establishing linkages (D’s product feature A matches P’s claim limitation X), for which the party would otherwise want written evidence; however, the expert must avoid the tendency towards mere expert ipse dixit
  • Support for the party’s claim construction, based on the expert’s opinion of what the PHOSITA would have understood a term to mean at the relevant time
  • Support for any inherency assertions (see xxx)
  • Results of testing or experiments, to be cited in the right column of the claim chart
  • [TODO: expert declaration supporting/opposing SJ; conclusory expert statements have little weight re: SJ; see e.g. Telemac v. Topp; use claim chart format to avoid conclusoriness, which is otherwise common (expert bases opinions “on my experience and training,” i.e. ipse dixit, without the necessary parsing of terminology into component parts, such parsing being seen as “that’s what the lawyers do”).]

Importantly, it does not work to employ the expert to “back-fill” or “backdoor” theories of infringement or invalidity which the party failed to present in a timely manner in its claim charts. See the examples cited earlier in the “results” section, e.g. Asus v. Round Rock: “P cannot include a placeholder in its contentions that generally asserts indirect infringement, and then later clarify the infringement theory through an expert report.” In many cases, the court has struck portions of the expert report when these put forward theories (especially DoE or indirect infringement) not earlier disclosed, or sufficiently disclosed, in the party’s claim charts.

It is also a problem if the expert merely adopts the party’s claim chart wholesale, while presenting it essentially as “all my own work.” [TODO: possibly see IPPV v. Echostar?] A related problem is posed by the expert whose report merely attaches (or “incorporates by reference”) the party’s claim charts, while putting forward theories and/or evidence substantially different from that in the endorsed claim chart.

There is a surprising tendency for experts to use the wrong dates or wrong time-frame when assessing claim construction obviousness, opining as to what the expert today regards as the meaning of a claim term, or today regards as “obvious” or not, without properly seating the opinion in the past. [TODO: see Patent Case Mgmt Judicial Guide on this as a common problem.]

Also see a separate section towards the end of this book on experts and claim charts. [TODO: though how does that section differ from this one?]

Representative instrumentality chart

Where a patentee accuses multiple “instrumentalities” (products, versions, models, etc.) of infringement, the patentee must prove that each such accused instrumentality infringes. It must fulfill the “each x3” requirement of the LPRs, discussed above (each limitation x each asserted claim in each asserted patent x each accused instrumentality). It will not do to present proof for one product and then vaguely but hopefully wave one’s hands in the direct of other, unproven, products. See Bender v. Maxim, granting motion to compel P to supplement charts, and finding three “representative” claim charts for over 100 accused products failed to show “that the accused products share the same or substantially the same infringing circuitry as any other product,” failed to adequately identify the accused products, and instead improperly used broad and vague names to identify accused products.

If there are multiple accused products, and the litigant does not wish to chart each one individually (and the other side is not insisting that it chart each one individually; see below), and the accused products have sufficient commonalities such that a “representative” product can be chosen to stand in for the others, the litigant can simply chart this “representative instrumentality” rather than also grind out multiple repetitive charts for each of its kin. In other words, the patentee may meet its proof obligations by showing that a representative example infringes and that each accused instrumentality conforms, in the relevant way, to that exemplar.

However, “simply” is not entirely accurate, as the litigant must also show why the product chosen to be charted is indeed representative of those not charted. See e.g. Implicit Networks v. Hewlett-Packard; Vigilos v. Sling Media; Cap v. McAfee. The party’s say-so is insufficient (Silicon Labs v. Cresta, noting that this is a recurring dispute in patent litigation). The burden is on the chart maker to show representativeness, not on the opponent to show non-representativeness. This burden may be carried by an expert testifying that his detailed testimony regarding one model also applies to other models, without discussing those other models in detail (Tivo v. EchoStar, CAFC; see also experts & representative products in Shared Memory v. Apple; Renesas v. Nanya).

The party must, in other words, present a reasonable explanation. One such explanation is simply that all the accused products belong to the same product family, or are simply different versions of the same product, and if true, this is likely a sufficient explanation for why picking one exemplar should suffice.

In many cases, the explanation will not be so simple. For example, P accuses different products from different families across D’s entire product line (but all from the same D; also see the multi-defendant scenario below). What these products have in common, in P’s eyes, is precisely their infringement of P’s patents, i.e. that all the products embody limitations X, Y, and Z. But they may so embody X, Y, and Z in substantially different ways, however, in which case they likely require separate charts.

A representative instrumentality should be representative of some known set of accused instrumentalities, in contrast to an open-ended accusation. See e.g. Finjan v. Proofpoint (striking patentee’s use of “including but not limited to” identification of accused products).

Further, it is a mistake to pick product 1 as representative of products 2 and 3, when the commonality between 1, 2, and 3 relates not to the accused limitations X, Y, and Z, but rather to some largely-unrelated feature A, B, or C. See e.g. Avago v. IPtronics, where the use of a representative instrumentality was successful, because the alleged difference between the representative product and the uncharted product did not relate to any claim limitation. This was the point of the “conforms, in the relevant way” above.

One way in which representativeness can be shown is with a second chart — and this is what this section refers to as a specifically representative-instrumentality chart — showing a product-to-product comparison between the charted product and the otherwise-uncharted products. The comparison, per the point made above from Avago v. IPtronics, should relate to the claim limitations at issue, and not merely recite some unrelated commonalities among the products. [TODO: show example]

This really only works when there is substantially commonality among the multiple products. While small differences might be noted at the appropriate place in a single chart [TODO: show], this should be contrasted with the attempt to shove multiple, substantially-differing products, in a single chart (see xxx on the problem of “Frankencharts,” and the danger that the chart maker will mistakenly or otherwise pick some limitations from one product, and other limitations from another product, when the two products are not used in conjunction with each other).

If some name is used for each category of accused product, it is best if this name matches that chosen by the vendor itself, and certainly best not to mimic the claim language in designating product categories. Also take care that each accused product fits into some category (which is then made the subject of a claim chart). See Bender v. Maxim, where P categorized some products (e.g., “LNAs”, “IF IQ Demod”, “RF VCO Buffers”, etc.), but did not categorize others; “Since no information is given that would suggest the proper category for each of the remaining products, the infringement contentions fail to comply with Patent L.R. 3-1.”

The representative-instrumentality chart should almost definitely be avoided when the multiple products belong to different vendors. See e.g. Linex v. Belkin, in which plaintiff created a single claim chart for each asserted claim and served that same claim chart on every defendant (including D-Link, Cisco-Linksys, Netgear, Dell, Gateway, and Toshiba), notwithstanding the fact that Ds each manufacture different products; P fails to take into account that the way in which each product is alleged to infringe will not necessarily be the same, even though all products are compliant with the 802.11n standard. “Not only are there numerous different kinds of accused products — including wireless routers, wireless networking cards, and laptops, but the 802.11n standard does not outline specific implementation details for compliant devices.”

As just suggested, care is required when the feature common to the accused products — even when from a single vendor, must less from multiple vendors, as in Linex v. Belkin — is a standard. Use of a standard is a reasonable starting point, but only that. See also SmartPhone v. HTC (“it is improper to simply chart to standards, e.g., the 802.11 standard, to illustrate how representative products satisfy claim limitations”).

Care is also required when asserting multiple patents against multiple products, and trying to do everything in a single chart. While multiple products can be handled with a representative-instrumentality chart, and multiple patents with a limitations chart (see below), it may be too easy to miss something important if these are combined [TODO: vague; example]. See the section of this book on “multiplicity” problems.

Finally, note that “representative instrumentality” issues frequently arise later in the case, and/or in the context of claim-chart amendments or supplementation. Typically, P will have accused D’s products 1 and 2, and later wants its claim chart to also cover D’s products 3 and 4. While product 4 is perhaps brand new, and P could not reasonably have included it in an earlier chart, product 3 could have been included but wasn’t. P will argue that the earlier claim chart is simply “representative” of other, uncharted products. This rarely works, as the representative instrumentalities must be explicit, and explicit as to precisely which other instrumentalities they are standing in for. See e.g. Digital Reg v. Adobe (P failed to show that earlier claim chart was representative of the newly accused products, as the new products had different software platforms from the originally-accused products); see also CANVS v. US; SmartPhone v. HTC; Computer Accel. v. Microsoft. This problem is similar to that of claim charts whose accused features are labelled as merely exemplary (“See, e.g….”), discussed at xxx.

[TODO: make clearer that this same tool could be used for multiple pieces of overlapping prior art in an invalidity chart (and might help with problem of many different combinations in some obviousness charts?). If individual references in the prior art are given names or numbers, this can appear similar to the limitations chart, see below.]

Injunction/contempt-related charts

If a party, already found to infringe a patent, and/or already subject to an injunction (perhaps a TRO), is now allegedly violating the injunction (and therefore in contempt of court) by continuing to practice the same technology as was enjoined, yet under a different name or as part of a different product or version, this may be shown using a chart that compares the current product/version with the enjoined product/version, asserting under e.g. Tivo v. EchoStar a lack of “colorable differences.”

In other words, the resulting chart is not comparing the allegedly still-infringing technology to the original asserted patent, but rather to the older (already-enjoined) product, and therefore this is a rare instance in which a product-to-product comparison is a valid exercise in patent litigation. See product-to-product comparison chart, below.

Also injunction-related, but employing a standard claim chart, a request for a TRO or preliminary injunction (PI), likely ought — in order to show likelihood of success on the merits — be accompanied by a claim chart which rises above the bare minimum acceptable for PICs, merely showing why the litigant believes it has a plausible case (see Iqbal, above) that it believes it can prove. The patent holder, after all, is not merely using the chart to force D to open its kimono in discovery, but is using the chart to ask the court to halt D’s product. See PICs above on what can be done to take pre-discovery charts above the bare minimum.

A request for equitable relief (TRO or PI) likely also requires a working/practicing chart (see above), to show that the litigant practices what it preaches, so to speak, i.e. is not a non-practicing entity (NPE) that could be made whole entirely through monetary damages.

Product-to-product comparison chart

As was noted earlier in the “Comparing What to What?” section, claim charts use the claims of a patent as the basis for comparison with something outside the patent, and this is to be distinguished from a frequent misconception of solo inventors litigating pro se, or hoping to play a large role in their litigation, who wish to use their own product — upon which their patent is “based” or at least bears some loose relationship, about which the inventor may be unclear — as a basis for comparison with the product of an allegedly infringing competitor.

However, this are several occasions for product-to-product comparisons. As shown above at “representative instrumentality chart,” such a product-to-product comparison can help show that one allegedly infringing product chosen for charting is representative of, and that its chart can stand in for, other accused products. As another example (see “injunction/contempt-related charts” above), following an injunction, showing that the infringer’s “new” product violates the injunction may be shown better by comparing the new product with that which was enjoined, rather than going back to first principles to re-show infringement. Finally, the accused infringer may wish to point to the patent-holder’s own product, ostensibly covered by the patent, as part of a “reverse doctrine of equivalents” type of argument (see xxx).

The resulting chart may be more akin to a copyright-infringement chart than to a patent-infringement chart. [TODO: show small example?]

“Parts list” as demonstrative exhibit

A good example of a patent claim “parts list” may be found in Geffner v. Linear Rotary Bearings, 936 F.Supp. 1150 (1996):


The left column was based on claim 1 of the ‘128 patent at issue, which read in part, “In a rectilinear and rotary motion bearing having relatively spaced first and second race members, one of said members having relatively spaced bearing lands, rollable bearing elements in said bearing for movement between said first and second race members and movable into and out of engagement with said bearing lands, retainer means between and movable relative to said first and second race members and having a plurality of closed loop paths for retaining said bearing elements between said first and second race members…”

It can be seen that, besides producing a standard claim chart broken into limitations (as also appears in the Geffner v. Linear opinion), the patent claim has also been broken into small parts, with a quick check that each such part has a correspondent in the cited piece of prior art (here, Manby).

Such short-hand summaries of each limitation or sub-limitation may in the right column instead simply contain a checkmark, perhaps to be filled out as a demonstrative exhibit at trial, to help a walk-through with a deponent, or as an internal case checklist.

[TODO: sublimitations: give an example of the difference between a limitation and a part, e.g. in what the parties will likely view as a single limitation (because semicolon delimited) “a download engine communicatively coupled to the digital engine and configured to indicate, in response to the request for the piece of selectable content, that a first version of the piece of selectable content is available for downloading…,” there arguably are these separate parts: download engine, digital engine, coupling between the two engines, configuration of the download engine, a request, and so on; also compare to “terms” in claim construction, which the LPRs sometimes describe as “terms, phrases, or clauses” — which are more linguistic than a sublimitation, which is xxx.]

Apart from the benefit of a quick, concise overview of the litigant’s assertion, the focus on separate rows for parts, rather than for limitations (here, delimited by commas though usually delimited by semicolons), forces issues such as whether e.g. the prior art really contains two separate race members (the first corresponding to Manby 11 and 30, the second corresponding to a shaft in Manby 10), which might have been glossed over were a single row devoted to an entire limitation.

At the same time, there is a danger (also present in limitations charts; see below) that in focusing on parts, rather than on entire limitations, and therefore omitting small connecting words and phrases (e.g. “one of said members having”, “in said”, “plurality of”, “for”, etc.), that non-part requirements of the claim will be glossed over. Here, for example, the parts list, while easy to digest, omits any indication of whether the ostensible closed loop path[s] (the circuits formed by trackways 27 and end portions 28) are “for retaining…” (which, however, may not be a limiting feature of the claim).

Another, related, danger is what is sometimes referred to as the “laundry list”, “bag of parts,” or, indeed, “parts list” problem that also occurs with standard claim charts: having found some feature or aspect of the accused instrumentality, or of the prior art, corresponding to each and every specified part in the claim, do all those features or aspects from the accused instrumentality or prior art fit together to form a coherent whole? See xxx on the laundry list problem, noting the danger for example of “Frankencharts,” in which limitations are supplied from product or feature A and some from product or feature B, when A and B are separate or incompatible. (Of course, the danger is not a concern for disaggregated “kit” patent claims, see e.g. Application of Venezia.)

One solution in the context of claim charts is to add a final row to the parts list: “Do all parts fit together?” or “Do all parts form a single whole?” (And, perhaps, do the parts fit together in the same way as the parts extracted from the patent claim?)

Limitations chart for multiple overlapping asserted patents

It was seen above that representative-instrumentality charts can help reduce the claim-chart workload when there is a large number of accused instrumentalities (or, perhaps related pieces of prior art). A related chart type, the limitations chart, helps when there there is a large number of asserted patents and claims, and there is significant repetition of particular limitations.

Of course, some litigants actively desire a huge stack of repetitive claim charts, see xxx, and perhaps even desire the wasteful make-work involved in generating them, and some litigants more sincerely fear using anything other than the standard format. However, the limitations chart can at the very least be a useful internal tool for organizing the material that will go into claim charts (and, for better or worse, “merge” scripts can be created for Microsoft Word to use the two parts of a limitations chart to grind out a much larger volume of claims charts; one could also envision an “un-merge” script to read in a large mass of claim charts and spit out the much smaller limitations chart that, arguably, ought to have instead been produced).

The limitations chart generally has two sections: in the first, all asserted claims, from all asserted patents, are listed in the left column, with the right column showing corresponding limitation numbers or names in the right column, representing the limitations which comprise each listed claim. It is expected (and indeed, is the entire point of the limitations chart) that the same limitation numbers or names will appear repeatedly in the right column. When the exercise is completed, it may turn out that over 100 asserted claims boil down into, say, 20 unique limitations. Limitation numbers may be as simple as “LIM1”, “LIM2”, and so on, or more readably (but perhaps also misleadingly, given the greater complexity of most limitations) “ROLLER_BEARING”, “CLOSED_LOOP_PATH”, and so on.

In the second section, the left column contains the extracted limitations, and the right column shows the corresponding assertions and information relating to the accused instrumentality or prior art.

[TODO: show contrived example: one table in which each asserted claim in each asserted patent -> LIM1, LIM2, LIM3, etc., or perhaps SUB1, SUB2, SUB3, such that multiple claims & patents will share LIM1 or SUB1, etc.; and then separate large table juxtaposing LIM1 to accused product, LIM2 to accused product, etc.]

Consolidating a large number of claims into a smaller number of limitations often requires ignoring “minor” wording differences between almost-but-not-quite identical limitations appearing in different claims or different related patents. Given the principle that different claims necessarily have different meanings (the basis for claim differentiation), it is entirely possible that two almost-identical limitations are sufficiently distinct that, at the very least, a footnote is required in the limitations chart. [TODO: give example]

Naturally, the limitations chart can be combined with the parts list (see above), to generate a parts chart, in which our hypothetical 100 asserted claims boil down into, perhaps, 30 unique parts (and here, names such as “ROLLER_BEARING” or “CLOSED_LOOP_PATH” would likely be more accurate than if such names were assigned to entire limitations). This has all the benefits, and all the dangers, of the parts list discussed above.

Claim charts in petitions to PTO, PTAB for reexam, IPR, PGR, or CBM review

The PTO’s Manual of Patent Examining Practice (MPEP) 2214, Content of Request for Ex Parte Reexamination Filed under 35 U.S.C. 302, presents a straightforward example claim chart to satisfy the requirements for a detailed explanation under 37 CFR 1.510(b):


(Also see a disaggregated version of claim 1 from this chart, at the end of this section.)

MPEP 2660 (Optional Inter Parties Reexamination / First Office Action): “Where the request for reexamination includes material such as a claim chart to explain a proposed rejection in order to establish the existence of a substantial new question of patentability (for a reexamination filed under that standard), or a reasonable likelihood that the requester will prevail (for a reexamination filed under that standard), the examiner may bodily incorporate the claim chart (or other matter) within the Office action. The examiner must, however, carefully review the claim chart (or other material) to ensure that any items incorporated in a statement of the rejection clearly and completely address the patentability of the claims….”

Petitioners to PTAB have on several occasions been told that claim charts are not to include argument. For example, VMWare v. Electronics & Telecommunications Research Institute (IPR2014-00901): “Claim charts should only be used to provide an element-by-element showing as to how the prior art teaches the limitations of a claim (e.g., citations to a prior art reference, quotations from a prior art reference). Claim charts may not include arguments, claim construction, statements of the law, or detailed explanations as to why a claim limitation is taught or rendered obvious by the prior art.” For instance in the sentence “In addition, Hathorn Fig. 5 discloses shadowing data across multiple disks to create a remote dual copy, which is a RAID architecture commonly known as RAID 1. (Ex. 1005 at 8:64-9:51; 12:54-60; Ex. 1003, ¶ 143.),” the italicized phrase “which is a RAID architecture commonly known as RAID 1” constitutes such argument, rendering the IPR petition defective. This appears to be a recurrent problem: Marvell Semiconductor, Inc. v. Intellectual Ventures I (IPR2014-00547), Google v. Visual Real Estate (IPR2014-01338), B/E Aerospace, v. MAG Aerospace Industries (IPR2014-01510). [TODO: but what’s the issue?; the decisions talk about fonts, single-spacing, and page limits which hardly seems to merit the decisions.]

At the same time, a petition can be rejected for failure to explain the contents of claim charts. See GN ReSound v. Oticon (IPR2015-00103), citing 37 C.F.R. ç 42.22(a)(2) (the “petition . . . must include . . . [a] full statement of the reasons for the relief requested, including a detailed explanation of the significance of the evidence”), and noting that petitioner’s claim chart failed to link up prior-art references with specific claim components. For example, in row [6f] the example below, it is not indicated which of the references in the right column corresponds to the recited “shield element” in the left column (this is a very frequent problem with claim charts generally; see xxx on the lack of specificity that comes from parsing claims only on semicolons):


The petitioner attempted to be more specific in its Request for Rehearing by color-coding the links between individual components (sublimitations) on the left side and prior-art references on the right, but by then it was too late (“the color-coded mapping of claim elements to the prior art disclosure were not presented in the Petition”):


While not limited in relevance to PTAB practice, a possible lesson here is that the sort of careful linkage or tying shown above — going beyond the normal dump of an entire semicolon-delimited phrase into the left column, and a dump of a mass of corresponding evidence in the right column — would be an excellent practice, as matter of course, perhaps with the linkages made textually, or with separate rows, rather than in four colors. See xxx on tying, linking, and/or weaving the right and left sides of a claim chart. [TODO: would a textual linkage, e.g. “an antenna formed as a metallic cylinder is a shield element,” constitute apparently impermissible “argument”?]

As an example, returning to the first example in this section from MPEP 2214, the same material could be presented by disaggregating compound limitations into separate rows, as appears in the Wikipedia article for claim charts:


[TODO: Possibly combine this section with other PTO section below; here, charts submitted to PTO/PTAB by applicants; other section contains charts provided by PTO examiner]

[TODO: add e.g. Covered Business Method review (CBM) claim chart with non-patent prior art (review of 8,209,389); example 2]

Claim chart employed in licensing negotiations

Claim charts are not only employed in litigation. For example, claim charts should be prepared before litigation, as part of a reasonable pre-filing investigation (and then presented, perhaps in abbreviated form, as PICs, see above, with the complete pre-filing charts kept in the attorney’s back pocket as evidence, should it ever be needed, for a Rule 11 hearing into the reasonableness of a pre-filing investigation).

As another example, claim charts may be used in a context which attempts to avoid litigation entirely — licensing “negotiations” or demands — though importantly the use of claim charts in this context can lead to litigation, on somewhat on the other side’s terms, in that presentation to the other side of claim charts allegedly showing their infringement of a patent may be sufficient to, or at least be excellent evidence of, the existence of a sufficient controversy to justify a declaratory judgement (DJ) suit by the other side. As discussed at xxx above, DJ reverses the usual roles of plaintiff (P) and defendant (D). Instead of patent-owner P suing D for infringement, in DJ accused infringer P sues patent-owner D, seeking a DJ ruling of non-infringement (and often invalidity); because the judiciary will not issue “advisory opinions,” and requires an actual case or controversy (US Const. III(2)(1)), DJ P has to show a sufficient controversy to give it DJ standing. A patent-owner’s presentation of a claim chart alleging infringement can help establish such DJ standing. The accused infringer has the advantage of picking the time and place of the DJ suit.

An excellent example appears in Sandisk v. STMicroelectronics, 480 F.3d 1372 (CAFC, 2007; citations omitted):

“On August 27, 2004, the licensing meeting was held. Jorgenson [ST’s VP of IP and licensing], two ST licensing attorneys, and three technical experts retained by ST[Microelectronics] to perform the infringement analyses of SanDisk’s products, attended on behalf of ST. … a four- to five-hour presentation by ST’s technical experts, during which they identified and discussed the specific claims of each patent and alleged that they were infringed by SanDisk. According to Thompson, the presentation by ST’s technical experts included ‘mapp[ing] the elements of each of the allegedly infringed claims to the aspects of the accused SanDisk products alleged to practice the elements.’ Thompson declares that ‘the experts liberally referred to SanDisk’s (alleged) infringement of [ST’s] products.’ … At the end of the meeting, Jorgenson handed Thompson a packet of materials containing, for each of ST’s fourteen patents under discussion, a copy of the patent, reverse engineering reports for certain of SanDisk’s products, and diagrams showing how elements of ST’s patent claims cover SanDisk’s products. According to SanDisk, Jorgenson indicated (in words to this effect): ‘I know that this is material that would allow SanDisk to DJ [ST] on. We have had some internal discussions on whether I should be giving you a copy of these materials in light of that fact. But I have decided that I will go ahead and give you these materials.”

Indeed, the four- to five-hour presentation by experts of their reverse-engineering results, presented in claim-chart form, was part of what ultimately led the CAFC to overrule the district court and decide that there was a sufficiently actual controversy to “allow SanDisk to DJ [ST] on,” just as ST’s representative said. The CAFC applied the Supreme Court’s decision in MedImmune v. Genentech, rejecting the older “reasonable apprehension of suit” test for DJ standing; costs were awarded to SanDisk.

DJ-triggering actions can occur with demand letters, as well as in licensing negotiations. There is a bit of an art form for crafting demand letters which do not trigger DJ standing; e.g. “We look forward to working with you to ensure that our patent rights are protected” is supposed to not make the recipient apprehensive (though it’s hard to imagine anything more ominous sounding, short of “This letter with accompanying claim chart shall serve as a formal notice of your infringement of our patent”).

As an example of a claim chart employed in licensing “discussions,” the following is from an inclusion in a Feb. 2008 letter sent by Gibson Guitar to Activision, regarding the latter’s Guitar Hero system. This claim chart was included by Activision in its March 2008 declaratory judgment court filing: [ ]


[TODO: say more about this chart; what was outcome of Activision v. Gibson DJ case; note somewhat different claim charts employed in settlement negotiations, including e.g. design-around?]

Claim construction chart for Markman hearing

To some patent litigators, the phrase “claim chart” may mean a claim-construction chart prepared for a Markman hearing, more than it means the infringement and invalidity charts which are the focus of the book. Markman-related charts are quite different from infringement/invalidity-related charts, and will only be briefly discussed here, especially as a wealth of information on patent claim construction is widely available, compared to the dearth of information noted earlier on the determination of e.g. literal infringement (which is perhaps felt to be too simple to require guidance, but which it is the intent of this book to show is a complex subject). A book on patent litigation without claim construction may seem like Hamlet without the Prince of Denmark, as cases are often won or lost on claim construction. However, it is the factual digging on infringement and invalidity that gets to the case to the point where claim construction is dispositive.

According to the Markman decision, claim construction is a legal matter for the court, not a factual matter for a jury. By extension, claim-construction charts are performing a legal rather than factual analysis. The claim charts which are the focus of the book compare the claims (as construed) to some external reality, outside the four corners of the patent (e.g. an infringing product or prior art). In contrast, claim-construction charts largely extract the meaning of claim terms from the patent itself.

Facts relating to infringement or invalidity are not intended to enter into claim construction, see SRI v. Matsushita (“construed without reference to the accused device”), even though each party’s claim construction postures will, of course, be determined by the facts it has regarding infringement and invalidity (see the point made earlier about claim charts “teeing up” the case for a Markman hearing).

However, the distinction just drawn between the two types of claim chart is somewhat misleading, in part because claim construction can employ extrinsic evidence as well as intrinsic evidence. While extrinsic evidence, such as definitions from technical dictionaries and testimony from expert witnesses (regarding e.g. the PHOSITA’s understanding of a claim term at the relevant time), is disfavored in comparison to intrinsic evidence, see Phillips v. AWH, it may be used, including in a claim-construction chart. Further, “intrinsic evidence” includes not only the patent itself, but also the history of the patent’s prosecution at the PTO (file wrapper), and prior art cited within the patent (prior art not considered by the PTO examiner is extrinsic evidence; see e.g. Tate Access Floors v. Interface).

Conversely, infringement/invalidity claim charts may explicitly employ one or more claim constructions (see xxx), and all claim charts at least imply some claim construction, by virtue of which facts they choose to align with each claim limitation.

One helpful treatise, Creel’s Patent Claim Construction and Markman Hearings (PLI, 2013) provides a “working claim chart” for claim construction, to be used internally as a worksheet in preparing the party’s Markman presentation; the following is adopted from the blank chart in the PLI treatise and its explanatory text (pages 6-6 to 6-13):


The PLI treatise also includes a blank format of the chart that would be submitted to the court:


One might question the benefits of breaking out “extrinsic evidence” as a separate column. If largely empty, it perhaps emphasizes the construction’s reliance on more-favored intrinsic evidence; alternatively, it or perhaps can help show that one is not contradicting intrinsic evidence.

[TODO: note claim construction at ITC, with “Staff’s Construction”]

Joint Claim Construction Chart

As with individual parties’ separate Markman charts above, the Joint Claim Construction (CC) Chart is included here for completeness as these are otherwise not the type of claim chart that is the focus of this book.

In a Joint CC Chart, the opposing parties collaborate to help define and narrow the focus of an upcoming Markman claim-construction hearing. The Patent Case Management Judicial Guide, 3rd ed., includes Appendix 5.2 Sample Joint Claim Construction Charts; for example:


The LPRs typically cover the joint production of this chart, associated with a claim construction “meet and confer.” The earlier section on LPRs quoted as an example SDCAL LPR 4.2, whose Joint CC Chart requirements are here provided in bulleted-list form:

  • A column listing complete language of disputed claims with the disputed terms in bold type;
  • Separate columns for each party’s proposed construction of each disputed term;
  • Each party’s proposed construction of each disputed claim term, phrase, or clause, must identify all references from the specification or prosecution history that support that construction;
  • Identify any extrinsic evidence known to the party on which it intends to rely either to support its proposed construction of the claim or to oppose any party’s proposed construction of the claim, including, but not limited to, as permitted by law, dictionary definitions, citations to learned treatises and prior art, and testimony of percipient and expert witnesses;
  • For every claim with a disputed term, each party shall identify with specificity the impact of the proposed constructions on the merits of the case.

The chart as shown above does not confirm to these requirements. The Patent Case Management Judicial Guide has an additional example, see below, with some citations to intrinsic evidence, and a “Significance” column:


(It is worth noting that means-plus-function here is handled as a matter of disputed claim construction, rather than straightforward incorporation of disclosed means from the specification, see xxx above.)

In collaborating and haggling over the Joint CC Chart, the parties will likely push up against limits, defined in the LPRs, on the number of claim terms in dispute. See e.g. NDCAL LPR 4-3(c), requiring “an identification of the terms whose construction will be most significant to the resolution of the case up to a maximum of ten (10) terms. The parties must also identify any term among the ten (10) whose construction will be case or claim dispositive. If the parties cannot agree on the ten (10) most significant terms, the parties must identify the ones which they do agree are most significant and then they may evenly divide the remainder with each party identifying what it believes are the remaining most significant terms. However, the total terms identified by all parties as most significant cannot exceed ten (10).”

The N.Ill. LPR 4.1 has the following comment: “In some cases, the parties may dispute the construction of more than ten terms. But because construction of outcome-determinative or otherwise significant claim terms may lead to settlement or entry of summary judgment, in the majority of cases the need to construe other claim terms of lesser importance may be obviated. The limitation to ten claim terms to be presented for construction is intended to require the parties to focus upon outcome-determinative or otherwise significant disputes.”

(Note that a claim “term” is not necessarily the same as, and is generally smaller than, a claim limitation. The LPRs refer to claim “terms, phrases, or clauses.” Compare these with sub-limitations, discussed elsewhere at xxx.)

See and .

Deficiency chart, to highlight problems in the opponent’s charts

Litigants frequently point to alleged deficiencies in each other’s claim charts; and experts are often called upon in their reports or in declarations to note deficiencies in the other side’s claim charts. Sometimes it is useful to do so, or makes a stronger impression, by using a chart of the deficiencies. A good example appears in pretrial management findings and conclusions from H-W v. Apple (2012 WL 3650597; see


Comparing this with the claim chart itself (see 2012 WL 1745721), one can see the basis for the allegations; though an attractive four-color presentation, the chart makes certain assumptions which, though likely true (surely, generic computer memory can store a contextual-search program, and here the memory merely needs to be operable for such storage, not even doing the storage), could easily have been made explicit by adding some text to the use of arrows:


Proceeding with the deficiency chart, and comparing it with the underlying claim chart the problems become clearer:



Here, unlike the allowance one might make for generic memory “operable for storing,” for example, the limitation very specifically requires “circuitry for submitting…” and “circuitry for receiving…” and merely pointing to a four-color photograph of circuitry shows nothing about what the circuitry is in fact for.

[TODO: say more about this graphical claim chart format, which attempt to use arrows from illustrations in the chart’s right column pointing to claim limitations in the left column. This has the benefit of taking seriously the requirement that claim charts point to WHERE each limitation is supposed to be located. However, it is clearly a forced fit when e.g. using a circuit board to show software locations. Also note ITC Rules request for drawings keyed to claim chart (“parts of such drawings, photographs, or other visual representations should be labeled so that they can be read in conjunction with such chart”).]

As a further example, see a deficiency chart in Bedrock v. Softlayer (, e.g.:


[TODO: walk through one example from Bedrock v. Softlayer deficiency chart; also note Cisco’s 4-color markup of Connectel claim chart in 2005 WL 6152756.]

Attorney opinion letter, including Freedom to Operate (FTO)

Just as a patent owner will do a pre-filing investigation of another’s infringement of its patent (see PICs above), similarly that other party or its attorney may do its own non-litigation investigation of its own potential infringement or non-infringement of one of more patents, whose validity it will typically also investigate. This investigation is typically memorialized in an attorney opinion letter. To be meaningful, the opinion of non-infringement should be based on a limitation-by-limitation analysis, i.e., a claim chart or equivalent.

Before the decision In re Seagate (CAFC, 2007), having a component non-infringement and/or invalidity opinion letter from counsel was important in avoiding a finding of willful infringement, and hence treble damages, as the legal standard was an affirmative duty of due care to avoid infringement. There was an industry devoted to generating such opinion letters. A party’s willfulness further depended in part on the reasonableness of the party’s reliance on the attorney opinion (which, of course, later turned out to be “wrong,” hence the dispute over willful infringement). Reasonable reliance in turn could depend on the manifest thoroughness of the attorney’s analysis of non-infringement (see e.g. Golden Blount v. Peterson, CAFC, 2006) (“oral opinions obtained by Peterson were rendered without counsel having examined either the patent’s prosecution history or the accused device, and were thus incompetent”).

The AIA codifies the reduce role of opinion letters in 35 USC 298: “The failure of an infringer to obtain the advice of counsel with respect to any allegedly infringed patent, or the failure of the infringer to present such advice to the court or jury, may not be used to prove that the accused infringer willfully infringed the patent or that the infringer intended to induce infringement of the patent.”

Note that the AIA refers not only to willfulness, but also to inducement (see indirect infringement at xxx). [TODO: the same evidence may also help establish a substantial question re: infringement or invalidity to avoid preliminary injunction (PI)?]

At the same time, parties of course will still desire actual advice (as opposed to what was perhaps in some cases the mere writing of a “CYA” letter for the files), which — within time and budget constraints — will require a limitation-by-limitation analysis, a reasonable claim construction, careful review of the patent’s file wrapper, and so on.

Further, an opinion letter can still be used, proactively, to rebut willfulness and inducement. See Finistar v. DirectTV (CAFC, 2008), holding that Seagate imposes no affirmative obligation to obtain opinion of counsel, but opinion can negate an accusation of objective recklessness, if reliance on the opinion was reasonable, e.g. did counsel have all relevant information necessary to properly analyze non-infringement; see also Commil v. Cisco (CAFC, 2013): Good faith belief that patent invalid can be defense to induced infringement.

Apart from attorney opinion, a party’s own investigation can show due diligence to negate willfulness. See Cohesive Tech v. Waters (D.Mass., 2007), where D retained scientists to test for infringement.

Post-Seagate, cases continue to hold that whether an infringer conducted an infringement investigation continues to be relevant, though not presumptive, of willfulness. The difference is if the defendant was on notice of possible infringement. E.g., Veracode v. Appthority, 137 F.Supp.3d 17: “The Seagate court accordingly shifted the focus away from actual knowledge of the patent to knowledge — either actual or constructive — of a risk of infringement of a patent…. Although ‘Seagate removed the presumption of willful infringement flowing from an infringer’s failure to exercise due care to avoid infringement,’ it is still relevant whether an infringer conducted a freedom to operate search, obtained advice of counsel, or conducted some other form of investigation upon receiving notice of the likelihood of a relevant patent,” citing Spectralytics v. Cordis (CAFC, 2011), in turn applying enhanced-damages factors from Read Corp. v. Portec (CAFC, 1992). [TODO: does this also explain Broadcom v. Qualcomm (CAFC, 2008): failure to obtain noninfringement opinion was probative of inducement? is this inconsistent with AIA, 35 USC 298?]

While normally such materials would be protected under attorney/client privilege and the work-product doctrine, if D chooses to rely on an opinion-letter defense, naturally the opinion letter and ancillary materials become discoverable [cite].

In addition to targeted comparisons of a company’s product with a specific patents, companies also engage in broader Freedom to Operate (FTO) searches. The same considerations as above apply. [TODO: more, e.g. scope of search methodology?; sometimes search & analysis done using short-hand descriptions of limitations or parts, i.e. using a parts list (see xxx) rather than a claim chart?]

[TODO: see large book, David Fox, U.S. Patent Opinions and Evaluations (2010); has sample invalidity charts with claim construction for “An anvil for cartoon use…”, including a 4-column obviousness chart]

Chart showing dependent-claim trees, patent families

For completeness, note that other types of “claim chart” — but lacking the limitation-by-limitation and comparison-with-something-outside-the-claims features emphasized in this book —  are available as tools for those working in patent prosecution and patent litigation. Such tools include patent claims trees and patent family maps.

A useful “Patent Claims Tree” add-on is available for the Chrome browser, taking patents displayed at Google Patents, and generating a concise picture of the claims, including claim type, a word count (providing a rough sense of the claim’s breadth or narrowness, as all else being equal a shorter claim is broader than a lengthy claim), and a tree of dependent claims ( For example:


Patent-family or patent-portfolio tools can be found with a Google search for “patent family mapping”. [TODO: e.g.; MaxVal ; see Patent Information User Group (PIUG) list of tools:,+Mapping,+and+Visualization+Tools]

While not entirely on-topic, it is worth noting the area of patent valuation or rating, including patents in this area, some of which take limitations into consideration (as opposed to those which focus e.g. on citations to the patent), e.g. 7,676,375 (System and method for valuing patents); 9,177,349 (Method and system for rating patents and other intangible assets).

Claim chart in jury verdict forms

Jury verdict forms typically concern claim numbers (“Has P proven it is more likely than not that D’s product X has directly infringed claim 1 of the ‘123 patent?”) and dollar amounts (“If you found at least 1 asserted claim of the ‘123 patent is infringed by D’s product X, what reasonable royalty damages do you award to P?”). It appears rare for a verdict form to focus the jury’s decision on a limitation-by-limitation parsing of claims, as would be the case were the jury verdict form to contain a claim chart or parts list (see above).

However, in one published case the losing party complained that the claim preamble was omitted from the jury verdict form, suggesting that the form in other respects proceeded on a limitation-by-limitation (the omission was found harmless). Lucent v. Newbridge (168 F.Supp.2d 181), citing EMI v. Cypress (104 F.Supp.2d 370, 388), which explicitly states “For each asserted claim, the verdict form contained a claim chart instructing the jury to mark ‘YES’ or ‘NO’ for each claim element present or absent in the accused devices.” [TODO: court in EMI found that the jury instructions and verdict form improperly directed the jury to disregard elements of the claims when rendering a verdict on anticipation and obviousness; preamble was also harmlessly omitted, not limiting.]

Claim chart employed by PTO examiner

Though not typically presented using a tabular format, PTO examiners proceed on a limitation-by-limitation basis in Office Actions regarding anticipation and obviousness, and the PTO Manual of Patent Examining Procedure (MPEP) notes that in some circumstances the examiner must perform what is effectively a limitation-by-limitation comparison:

“For each rejection under 35 U.S.C. 102 or 103  where there are questions as to how limitations in the claims correspond to features in the art even after the examiner complies with the requirements of paragraphs (I)(3) and (4) above, the examiner must compare at least one of the rejected claims feature-by-feature with the art relied upon in the rejection. The comparison shall align the language of the claim side-by-side with a reference to the specific page or column, line number, drawing reference number, and quotation from the reference, as appropriate.” (MPEP 2677 (Examiner’s Answer) / I Requirements for Examiner’s Answer / (I) Grounds of rejection / (5).)

(Cited paragraphs I(3) and (4) require, for 102 and 103 respectively, that the examiner must explain the rejection, “pointing out where all of the specific limitations recited in the rejected claims are found in the prior art relied upon in the rejection”.)

One case (TV Interactive Data v. Sony, 2012 US Dist LEXIS 171526 (2012)) references a PTO examiner’s use of a claim chart in a re-exam: “Sony contends that in the subsequent Office Action letter, the examiner provided a detailed claim chart demonstrating that every element of the `307 claims was present in the CDTV Manual. According to Sony, TVI responded by adding the ‘without rebooting the host device’ limitation, and the examiner at that point allowed the claims.”

The MPEP otherwise largely confines mention of claim charts to those provided by applicants in requests for (pre-AIA) interferences, e.g. MPEP 41.202 (Suggesting an interference): “provide a claim chart comparing at least one claim of each party corresponding to the count and show why the claims interfere” and “If a claim has been added or amended to provoke an interference, provide a claim chart showing the written description for each claim in the applicant’s specification”. [TODO: no new matter; note this as analogy at enablement?]

Claim charts provided by applicants also appear in the context of requests discussed above at “Claim charts in petitions to PTO, PTAB”. [TODO: combine two sections into one?]

Conclusion to Part II

The point of the seeming laundry list above of claim chart types, and scenarios in which claim charts appear, is not to fetishize a chart or tabular format, but rather to emphasize the use of a disaggregated limitation-by-limitation analysis, and presentation.

[MORE, briefly sum up main types (PIC, IC, etc.) and special issues within major types (DoE, means+function, obviousness, etc.]

Part III



Print Friendly, PDF & Email