Functional Claiming of Computer-Implemented Inventions in View of Recent Decisions

Part 1 of this articled — Avoiding Invocation of Functional Claim Language — focused on the general principles of law surrounding functional claim language in the context of computer-implemented inventions. Part 2, below, focuses on how the CAFC has applied some of these principles by analyzing recent decisions. The cases provide the standards used to analyze whether § 112(f)[1] should be invoked and also reiterate the amount of disclosure necessary to survive an indefiniteness analysis under 112(b) following the invocation.

Williamson v. Citrix Online, LLC (Fed. Cir. 2015) (en banc)

In the November 2014 opinion of Williamson v. Citrix, the majority held that the district court erred in concluding that “distributed learning control module” was a means-plus-function claim term subject to § 112(f). In Williamson v. Citrix (en banc)[2] the CAFC substituted the previous opinion with the en banc opinion overruling precedent that characterizes as “strong” the presumption that a limitation lacking the word “means” is not subject to § 112(f).[3]

As mentioned in part 1, the standard for assessing whether a claim term invokes § 112(f) recited in the en banc opinion “is whether the words of the claim are understood by persons of ordinary skill in the art to have a sufficiently definite meaning as the name for structure.”[4] Applying the standard in the en banc opinion, the court found that the term “distributed learning control module” was subject to § 112(f) and indefinite under § 112(b).

Claim 8 of U.S. Pat. No. 6,155,840 sets forth, inter alia:

a distributed learning control module for receiving communications transmitted between the presenter and the audience member computer systems and for relaying the communications to an intended receiving computer system and for coordinating the operation of the streaming data module.

The analysis of whether the term invokes § 112(f) begins with pointing out that the full term of the claim must be considered and not merely the introductory phrase, “distributed learning control module.” The opinion states that, as claimed, the full term is a format consistent with traditional means-plus-function claim limitations.[5] The court reasoned that the term “means” is replaced with the term “module” and the claim recites functions performed by the “distributed learning control module.”[6]

In the context of § 112, “module” is a well-known nonce word that can operate as a substitute for “means.”[7] The word “module” in claim 8 does not provide any indication of structure because it sets forth the same black box recitation of structure for providing the same specified function as if the term “means” had been used.[8]

I expected the en banc decision to mention that dictionary definitions may be used to determine whether a word may indicate structure. Interestingly, in contrast to the en banc opinion, the original November 2014 opinion reaches the conclusion that § 112(f) did not apply by relying heavily on the following dictionary definitions, which included structure in the definition of “module” to find structure in the claim term module.

  • The IBM Corporation, IBM Dictionary of Computing 439 (1st ed. 1994) – a packaged functional hardware unit designed for use with other components and a part of a program that usually performs a particular function of related functions.
  • Alan Freedman, The Computer Glossary 268 (8th ed. 1998) – a self-contained hardware or software component that interfaces with a larger system.
  • John Daintith & Edmund Wright, Dictionary of Computing 315 (4th ed. 1996) – programming or specification construct that defines a software component and a component of a hardware system that can be subdivided.

However, dictionary definitions were not considered whatsoever in the en banc opinion to determine whether the words of the claim are understood by persons of ordinary skill in the art to have a sufficiently definite meaning as the name for structure. Rather, the en banc opinion reasoned that the term “module” was a generic “black box” term that performed the recited computer-implemented functions and the prefix “distributed learning control” do not depart structure into the term “module.”[9] The court notes that “module” is simply a generic description for software or hardware that performs a specified function.[10]

Although the “distributed learning control module” is described in a certain level of detail in the written description, the written description failed to impart any structural significance to the term.[11] The specification of the ‘840 patent supports that the distributed learning control module operates as a functional unit of the distributed learning server and coordinates the operation of the streaming data module through input from the presenter computer system. The specification also makes clear that the distributed learning control module includes software that runs on a portion of the distributed learning server.

Consider the portions of the specification related to Figs. 2 and 3, which the court relied on in their original conclusion that § 112(f) does not apply. Fig. 2 of the ‘840 patent is reprinted below.

Figure 2

Fig. 2 is a high-level block diagram illustrating a typical computer system 200 for implementing the distributed learning server (DLS) 102, on which the claimed distributed learning control module is implemented. Fig. 3 (reprinted below) shows a block diagram illustrating the functional units of the DLS 102, including a distributed learning control module (DLCM) 310, a classroom environment module 312, and a streaming data module 314. The specification states that the DLCM 310 controls the communications among the various computer systems 106, 108 in the distributed learning system 100 and manages the other modules in the DLS 102.


However, the written description fails to impart any structure significance to the term learning control module. The court did not find anything that lead them to construe the expression as the name of a sufficiently definite structure to avoid the invocation of § 112(f).[12]

Looking back to the language of the claim, one fault pointed out by the court was that the claim does not describe how the “distributed learning control module” interacts with other components in the distributed learning [control] server in a way that might inform the structural character of the limitation in question or otherwise impart structure to the “distributed learning control module.”[13] Note that in claim 8, the distributed learning server comprises the “distributed learning control module.”

Further, the en banc opinion reiterates the principle that mere testimony declaring that one having ordinary skill in the art, reading the specification, would know how to program a computer to perform the recited functions, does not replace structure where none is otherwise disclosed.[14]

Having found that the term invokes § 112(f), the decision continues with analyzing whether the specification discloses sufficient structure that corresponds to the claimed function. The district court identified three functions in the claim: 1) receiving communications transmitted between the presenter and the audience member computer systems; 2) relaying the communications to an intended receiving computer system; and 3) coordinating the operation of the streaming data module.

The opinion focused on whether adequate structure corresponding to the “coordinating” function is disclosed in the specification. After determining that a special purpose computer is required to perform the function, the court searched for an algorithm for performing the function, but did not find one. The court rejected Williamson’s argument that the distributed learning control module controls communications among the various computer systems and that the “coordinating” function provides a presenter with streaming media selection functionality. The disclosures relied upon by Williamson were thought of by the court as merely functions of the distributed learning control module and opined that the specification does not set forth an algorithm for performing the claimed functions.

Indeed, none of Figs. 1-5 or the written description disclose an algorithm for coordinating the operation of the streaming data module. As a result, the claim is indefinite under § 112(b) and therefore invalid.

The en banc opinion is important case law to understand. It indicates the priority the Federal Circuit is giving to functional language. It also echoes the importance of adequately disclosing the claimed functions in the specification and figures. The opinion especially reiterates the importance of adequately disclosing algorithms of the claimed functions in computer-implemented inventions.

Robert Bosch, LLC v. Snap-On Inc (Fed. Cir. 2014)

In October 2014, the Federal Circuit decided Robert Bosch, LLC v. Snap-On Inc[15] upholding indefiniteness, affirming that both claim terms “program recognition device” and “program loading device” are means-plus-function terms and the specification does not disclose the requisite corresponding structures. Therefore, the claims were held invalid as indefinite.

Claim 1 of U.S. Pat. No. 6,782,313 sets forth, inter alia:

a program recognition and program loading device,
wherein a program version contained in a connected control unit is queried and recognized by means of the program recognition device, and, if the program available in the motor vehicle and recognized via the diagnostic/test plug is not stored there in a latest and most current version, a respective most current version is loaded by the program loading device into a program storage device of the pertinent control unit of the motor vehicle…

Turning to the analysis of whether the terms invoke § 112(f), the court asked “if the claim language, read in light of the specification, recites sufficiently definite structure to avoid” § 112(f).[16] The court elaborated the issue, adding “the question is whether the claim language names particular structures or, instead, refers only to a general category of whatever may perform specified functions.”[17]

In particular, the court stated that the word “device” has been found to be non-structural in other precedent.[18] In addition, the other words of the phrase do nothing more than identify functions for the “device” to perform.[19] Bosch argued that the terms “program recognition device” and “program loading device” both name physical electronic structures with physical connections to other components discussed in the ‘313 patent. However, the court noted that the patent merely explains the device’s function and does not provide structural guidance whatsoever of the program recognition device itself. The ‘313 patent is silent on what such a “program loading device” consists of.[20]

The decision also mentions Inventio, in which the claimed “computing unit” and “modernizing device” were held to connote sufficiently definite structure, thereby avoiding an invocation of § 112(f). In Inventio, the CAFC stated that the claims indicate that the “modernizing device” functions as an electrical circuit that receives signals, processes signals, and outputs signals to other components in the patented system.[21] The court, in Inventio, also considered the specification, which depicts the modernizing device and its internal components, namely, the processor, signal generator, converter, memory, and signal receiver elements.[22] Fig. 3 of the patent-in-suit of Inventio is printed below, which shows the components of the modernizing device.

Fig 1

The written description of the patent in Inventio also shows how the elements are connected together and to the elevator control and computing unit components of the elevator system.[23] Accordingly, in Inventio, the claims recited sufficient structure to perform the recited function, which was also supported by the depiction of the components and connections thereof in the specification.

On the other hand, the court distinguished Bosch’s claims from Inventio since the claims involved functional terms and there is no language in the specification or dependent claims that defines the terms to refer to structures. [24]

The court found the specification silent about any interaction between the “program recognition device” and other components of the system, including the external diagnostic tester.[25] As stated by the CAFC, “the specification does not teach how the “program recognition device” receives and processes signals, as the words “signal” and “process” are not even in the specification.”[26] The same rationale was used with respect to the term “program loading device.”

Finding the presumption against invocation of § 112(f) was overcome, the court turned to whether the claim terms were indefinite, as required. However, the court easily found the claims indefinite since there was no guidance about the structure of the program recognition device or program loading device. As a result, all claims were held invalid.

Bosch reiterates the principles that a general recitation of a device does not set forth enough structure in the claim to avoid the invocation of § 112(f), where the device is modified by functional language.

The PTAB Informative Decisions

In April and March of 2013, the PTAB published some informative decisions, which provide additional, yet subordinate, guidance. As mentioned above, each of the four decisions involve functional limitations relating to processors and in each case, the PTAB entered, sua sponte, the combination of rejections under § 112(b) and (f).

Ex Parte Smith

In Ex Parte Smith, claim 1 provided to the PTAB set forth:

A computer system comprising:

a memory; and

a processor in communication with the memory,

the processor programmed to:

receive, from a user, a first review of an asset;

store the first review of the asset in association with a user identifier in a memory device;

receive, from the user, a second review of the asset;

store the second review in association with the first review and the user identifier in a memory device; and

generate an opinion timeline for the asset for the user associated with the user identifier.

To determine whether the limitations invoke § 112(f), the PTAB resorted to dictionaries to consider how one having ordinary skill in the art would understand the word “processor.”   The PTAB concluded that, after a review of the dictionary definition of processor, one having ordinary skill in the art would not recognize processor as having sufficient structure for “generat[ing] an opinion timeline…” and performing the other limitations. Rather, the term processor simply means a general-purpose computer. Indeed, the specification provided only “non-limiting” examples of a CPU and a graphics processing unit.

The PTAB also considered whether the functions performed by the processor in claim 1 are functions typically found in a commercially available off-the-shelf processor. As stated in the decision, if a general purpose processor would be capable of performing the claimed functions, then a skilled artisan might understand the claim term “processor” to provide sufficient structure for performing those functions.[27] However, the PTAB concluded that “at least generating the opinion timeline by the processor in claim 1 is not a typical function found in a general purpose processor and would require additional programming of the processor to implement, even where the processor is an off-the-shelf [graphics processing unit] GPU.”

The decision also mentions Inventio,[28]in which the claimed “computing unit” and “modernizing device” were held to connote sufficiently definite structure. The PTAB distinguished Inventio, however, since there is no structure connected to the processor in claim 1. Although the claim states that the processor is in communication with a memory, a memory does not have sufficient structure for performing the limitations.

With respect to the rejection under § 112(b), resort to the specification only disclosed a “general description which does not disclose how to generate an opinion timeline for the asset for the user associated with the user identifier, as recited in claim 1.” See step 420 of Fig. 4 below. Although, the specification includes a flowchart illustrating a five-step process for creating an opinion timeline, the PTAB stated that the specification merely provides non-limiting examples for each step.

figure 4

Further, the PTAB found that the specification “does not specifically indicate that either processor 106 or processor 616 may generate an opinion timeline.” The specification is devoid of explaining how the first and second reviews of a user could be used to generate an opinion timeline and “there is no algorithm to transform a general purpose processor into a special purpose processor for performing the function of generating an opinion timeline.”[29]

Ex Parte Erol

In Ex Parte Erol, at the time of appeal to the PTAB, claim 18 set forth:

A system comprising:

a reader adapted to read a machine readable identifier; and

a processor adapted to:

determine, from the machine readable identifier, a first object descriptor associated with an object, the first object descriptor specifying one or more features of content within the object;

determine a set of one or more objects from digital media content;

generate an object descriptor for each object in the set of objects, each object descriptor specifying one or more features of content within an object from the set;

identify at least one object descriptor from the object descriptors determined for the set of objects that matches the first object descriptor determined from the machine readable identifier; and

perform an action in response to identifying the at least one object descriptor that matches the first object descriptor.

As mentioned above, the PTAB found that the claims invoked § 112(f). The analysis used is nearly identical to the analysis of Ex Parte Smith. In analyzing the rejection under § 112(b), the PTAB cites Default Proof Credit Card Sys., Inc. v. Home Depot[30] for stating that while the corresponding structure need not include all things necessary to enable the claimed invention to work, it must include all structure that actually performs the recited function.[31]

Para. [0040] of the publication corresponding to the application states that:

Various different types of actions may be initiated or performed. Examples of actions include annotating the media content, performing an action using the media content, updating a database, sending a message, invoking a URL, or other like actions. Metadata information, if any, associated with the matching object descriptor in information 116 decoded from machine readable identifier 114 may be used as part of the action.

Citing para. [0040], the PTAB found that with respect to “perform an action in response to identifying the at least one object descriptor that matches the first object descriptor” in claim 1, the only portion of the specification that relates to the limitation are merely exemplary actions that may be performed. These exemplary actions were viewed as “an impermissible attempt to claim every way to “perform an action” under the sun, and do not constitute a sequence of steps of a particular algorithm required to meet the definiteness requirements of [§ 112(b)].”

Accordingly, the claims were rejected as indefinite. Interestingly, however, following the PTAB’s decision in Ex Parte Erol, the applicant amended the claims to recite:

A system comprising:

a reader adapted to read a machine readable identifier;

one or more processors; and

one or more modules stored on a memory and executable by the one or more processors, the one or more modules configured to:

The Examiner did not maintain the rejection of claim 18 entered by the PTAB. Of the four informative opinions, the application in Ex Parte Erol was the only application that continued prosecution following the decision of the PTAB.


Drafting claim language with enough structure to avoid the invocation must be considered to keep the patent out of troubled waters since a claim that invokes § 112(f) necessarily faces the indefiniteness hurdles of §112(b). The Williamson v. Citrix en banc opinion lowers the bar of presumption against the invocation of § 112(f). Using words that have a meaning to those having ordinary skill in the art to connote structure is essential in a claim that uses functional language to avoid the invocation of § 112(f). In addition, it is best practice to explain the structure of a term used in the claims so that the written description may be relied upon for support that the term connotes structure. In particular, practitioners should avoid using “nonce” terms alone (without modifiers) that could be thought of as replacing the word “means” in a claim and instead define the structure that performs the function.

Now that the presumption is lowered, Judge Newman, in dissent, remarked that “the examiner must guess whether the term is intended as a means-plus-function term, now that the court holds that the signal “means for” need not be used. Section 112(f) has morphed from a clear legal instruction into a litigator’s delight.

It seems the CAFC and PTAB find it easy to invalidate the claim as indefinite when a claim is found to invoke § 112(f). In any event, it is good practice to thoroughly explain the process of a claimed function in the form of algorithms, for example, and the structures performing the function. One rule of thumb is if a general, off-the-shelf computer, cannot implement the function without special programming (it probably cannot), then the algorithm (process) must be explained. In addition, the specification may inform the claims to avoid the invocation and the understanding of a person having ordinary skill in the art is considered.

Although challenges based on prior art during the lifetime of a patent are difficult for a practitioner to anticipate, practitioners do have control over challenges based on § 112 from the start of drafting the application. Therefore it is essential to keep the above principles in mind and understand how they are applied to provide quality patent prosecution.


[1] 35 U.S.C. § 112 was amended and subsections were renamed by the America Invents Act, Pub. L. No. 112-29 (“AIA”), which took effect on September 16, 2012. For simplicity,
pre-AIA 35 U.S.C. § 112, second paragraph will be referred to as § 112(b), and pre-AIA 35 U.S.C. § 112, sixth paragraph will be referred to as § 112(f)

[2] Williamson v. Citrix Online, LLC, No. 13-1130 (Fed. Cir. June 16, 2015)

[3] Williamson v. Citrix, No. 13-1130, slip op. at 15-16 (Fed. Cir. 2015)

[4] Id. at 16

[5] Id. at 17

[6] Id.

[7] Id.

[8] Id. at 18

[9] Id.

[10] Id. at 17

[11] Id. at 18

[12] Id. at 18-19

[13] Id. at 19

[14] Id. citing Function Media, L.L.C. v. Google, Inc., 708 F.3d 1310, 1319 (Fed. Cir. 2013)

[15] Robert Bosch, LLC v. Snap-On Inc., 769 F.3d 1094 (Fed. Cir. 2014)

[16] Inventio at 1357

[17] See Laitram Corp. v. Rexnord, Inc., 939 F.2d 1533, 1536 (Fed. Cir. 1991) (“[t]he recited structure tells only what the means-for-joining does, not what it is structurally”)

[18] See Mass. Inst. of Tech. v. Abacus Software, 462 F.3d 1344, 1354 (Fed. Cir. 2006); Personalized Media Commc’ns, LLC v. Int’l Trade Comm’n, 161 F.3d 696, 705 (Fed. Cir. 1998).

[19] Bosch at 1099

[20] Id.

[21] Inventio at 1358

[22] Fig. 3 of U.S. Patent 6,892,861 at issue

[23] Inventio at 1358

[24] Bosch at 1101.

[25] Id. at 1100.

[26] Id. at 1101.

[27] See In re Katz Interactive Call Processing Patent Litig., 639 F.3d 14 1303, 1316 (Fed. Cir. 2011) (functions such as “processing,” “receiving,” and “storing” that can be achieved by any general purpose computer without special programming do not require disclosure of more structure than the general purpose processor that performs those functions). This is also discussed above.

[28] Relying on Inventio at 1359-60

[29] Aristocrat at 1333

[30] Default Proof Credit Card Sys., Inc. v. Home Depot U.S.A., Inc., 412 F.3d 1291, 1302 (Fed. Cir. 2005)

[31] Id. at 1298 citing Cardiac Pacemakers, Inc. v. St. Jude Med., Inc., 296 F.3d 1106, 1119 (Fed. Cir. 2002)

The Author

Nicholas R. Mattingly

Nicholas R. Mattingly is an associate at Mattingly & Malur, PC. He focuses on patent prosecution and patent portfolio development for startups and large companies, both foreign and domestic. He is experienced in handling patent matters in various technologies including networking, storage systems, mobile devices, biometric identification, cloud storage, digital rights management, digital and analog circuits, electron microscopy, solar cell technologies, particle therapy systems, wireless communications, construction machines, and other various mechanical technologies.

Before going to law school Mr. Mattingly worked as a registered patent agent for Mattingly & Malur. During law school he was a staff member on the University of Baltimore Intellectual Property Law Journal and served as a judicial intern with the Honorable Clayton Greene, Jr. on the Court of Appeals of Maryland.

Warning & Disclaimer: The pages, articles and comments on do not constitute legal advice, nor do they create any attorney-client relationship. The articles published express the personal opinion and views of the author and should not be attributed to the author’s employer, clients or the sponsors of Read more.

Discuss this

There are currently 1 Comment comments.

  1. Edward Heller June 20, 2015 9:08 am

    Nick, thank you for your article. But let me make a comment. The Federal Circuit and the patent office are both pretending that cannot claim structure functionally. There is no statute and there is no case anywhere that I know of that is actually held this. Halliburton as clarified by Faulkner v. Gibbs held that a claim functional at “the exact point of novelty” is indefinite. That leaves open functionally claiming structure which is not novel, or which is not at the exact point of novelty. Indeed, Faulkner involved functionally claimed elements were the novelty was in the combination, rather than a particular element. The vice that the Supreme Court is seeking to prohibit is not functionally claiming structure but functionally claiming the invention. If a combination claim is equivalent to a single means claim because all the other elements are old and there is no new functional relationship between the one element that is new and the other elements which are old, then the claim is effectively a single means claim.

    We should also note that §112(f) is permissive. It is by its own terms, optional. It is invoked by the applicant. If a claim that is functional at the point of novelty does not use the statute however, the claim may be invalid as functionally claiming the invention if the applicant does not invoke §112(f).

    So what I am going to do henceforth is add a statement to the specification that states that the applicant does not intend, and specifically intends not to have his claims construed under §112(f) unless the claim at least uses “means for” or “step for.” This would end any debate, any legitimate debate that is, whether the applicant is intending to invoke the “optional” benefit of §112(f). A court would then have the find some case support for the proposition that one cannot claim structure functionally in order to invalidate a claim that is not functional at the point of novelty. That case support does not exist.

    This, of course, does not eliminate the requirement that the claim be supported under 112(a). But when one is claiming known structure, or software, that is fully enabled by a description of what it does, then one should have no problem under 112(a).