Recent developments in case law surrounding the application of 35 U.S.C. § 112(f) to software patents have further defined requirements for overcoming indefiniteness attacks against those patents and provided guidance for prosecutors and litigators on assessing the strength of any means plus function elements found in software claims. After a brief discussion of the history of §112(f), I will examine Ergo Licensing LLC, v. Carefusion 303 Inc., 673 F.3d 1361 (CAFC 2012), which clouds the present state of judicial examining of means plus function elements in software patents and provides a cautionary tale about disclosing too little information to support software means plus function elements.
- A Brief History
- How Courts Today View Software Means Plus Function Elements
- Software Means Plus Function in Light of Ergo Licensing
A Brief History
Originally, patents were drafted and their scope was defined by what was actually built by the inventor. In the early nineteenth century, American patents began using claims to broaden the scope of the patent and also provide guidance as far as what the invention included and did not include. Many early cases involving the interpretation of claims were, as we understand them today, essentially doctrine of equivalents analyses performed by the judges. As this trend progressed into the middle nineteenth century, inventors began abstracting claim language to encompass functional claiming, meaning claim language that claimed a function (a means for fastening) as opposed to a structure (a button). 
As America moved into the 20th century, functional claiming increased in popularity, until the 1946 Supreme Court decision in Halliburton severely curtailed the application of these functional claim elements. Though many practitioners believed Halliburton ended functional claiming, the utility of functional claiming was not exhausted. In the Patent Act of 1952, Congress decided to essentially undo Halliburton and explicitly allow functional claiming, though with greater limitations than functional claims enjoyed previously. Namely, the scope of a means plus function claim element is narrowed to those embodiments disclosed in the specification and equivalents thereof.
In relation to software patents, a few points should be noted. First, software patents are unique from many other types of utility patents in that most functionality can be completely separated from the structures that support and produce them. Almost all software patents require the same structural elements (a processor, memory, etc.); if functional language is couched in structural terminology, then that function becomes protected from any implementation that uses the same universal structures (i.e. since “a computer programmed to receive email” can be the same computer as “a computer programmed to produce sums,” all that distinguishes them is the functionality). This has led to widespread use of “’capability claims,’ which assert ownership of any device that is capable of implementing that idea whether or not the device actually does so.” Although infringement is still limited to computers that actually have the code to infringe a patent, this “capability claiming,” has widened the scope of functional claiming in software patents.
How Courts Today View Software Means Plus Function Elements
Patents are presumed to be written for persons skilled in the field of the invention. For that reason, the proper construction of a disputed claim term requires that the term be interpreted as it would by a skilled artisan at the time of the invention. In an attempt to rein in this functional claiming, Courts today impose additional limitations on means plus function elements in software patents. The Federal Circuit has held, “that the structure disclosed in the specification be more than a general purpose computer or microprocessor.” Courts now require that any means plus function software element be limited by the algorithms disclosed in the specification. These algorithms can be “in any understandable terms including as a mathematical formula, in prose, or as a flow chart, or in any other matter that provides sufficient structure.” These disclosure requirements do not require computer code, but they must be sufficient to enable one of ordinary skill in the art to understand the boundaries of the claim. It is also important to note that when the claim element claims more than one type of functionality, the algorithm used to support that element must describe each of the functionalities, otherwise the court will treat the element like it has no support from the specification. Of particular relevance to recently decided cases, courts need not rely on expert testimony regarding the disclosure if no algorithm is present.
Software Means Plus Function in Light of Ergo Licensing
Based on the above, it is clear that an algorithm must be present in the specification for a software means plus function claim element to be properly supported for the purposes of §112(b). However, the recent case of Ergo Licensing LLC, v. Carefusion 303 Inc., 673 F.3d 1361 (CAFC 2012), illustrates a lack of clarity regarding disclosure requirements and presents a specter for prospective plaintiffs of judges using discretion to decide whether disclosure is adequate without having to frame the issue in terms of whether a person of ordinary skill in the art would know if the disclosure was adequate.
Ergo Licensing sued Carefusion 303 for infringement of its US Patent No. 5,507,412 relating to an infusion system designed to meter and deliver fluids from multiple sources to a patient’s body. These fluids were metered by an adjustment means coupled to a controller, which controls the flow of different fluids through the adjustment means. The controller is claimed in claim 1 as “control means” and in claim 18 as “programmable control means.” Prior to a claim construction hearing, both parties stipulated that these terms were means plus function terms. Despite Ergo pointing to multiple disclosures in the specification, the District Court found inadequate disclosure in the specification to support the inclusion of these means plus function elements and Ergo appealed.
Ergo argued that the “control means” element was supported in the specification as a control device and that one of ordinary skill in the art would understand what was being referenced. Using Ergo’s own expert against them, the court relied on the expert’s testimony that in fact the control device could be one of a microprocessor, discrete circuits connected to stepper motors, and analog circuits, and hence, was not a defined structure. Not finding an adequate algorithm, the court held that under §112(b), the disclosure of the control device was not specific enough. In accordance with previous cases, since the control device would require specific programming, a recitation of a generic computer would not be satisfactory. “The specification merely provides functional language and does not contain any step-by-step process for controlling the adjusted means. As a result, we hold that the district court correctly determined that the ‘control means’ terms are indefinite for failure to disclose corresponding structure.”
The dissent in Ergo compellingly argues many points against the majority’s opinion. They believe that the columns of text dedicated to the invention adequately describe the invention, that the control means are not the point of novelty for a system arranging known components in a novel way for new results, and that the control means are adequately disclosed in the prior art. Correctly, the dissent admonishes his colleagues for not deferring to the expertise of the patent examiner, who presumably thought disclosure was adequate, stating, “This court’s foray into patent draftsmanship, finding standard presentations now to be fatally deficient, adds grievous unreliability to duly granted patents.” After demonstrating that the structures disclosed could be adequate, the dissent dissects the algorithm argument using In re Katz to show that a person of ordinary skill in the art would understand how to construct the control means based the information disclosed. The dissent then concludes:
The court’s new position simply taints thousands of heretofore innocent patents, adding a further infusion of unreliability to the patent grant. This is of particular concern because the panel majority is not relying on newly cited references or other possible examination oversights, but is invalidating a patent on purely formalistic grounds – grounds on which the expertise of patent examination is normally superior to that of judges.
Based on Ergo, it seems that the questions of indefiniteness as it relates to software patent means plus function elements now has a heightened level of scrutiny: it is not adequate to disclose enough that a person of ordinary skill in the art would understand what is claimed; instead, there must be enough information to convince the judge of adequate disclosure, regardless of what a person of ordinary skill in the art understands.
From a practice standpoint, this ruling essentially raises the bar on how much information is required to adequately support software means plus function elements. For prospective plaintiffs, this could mean that means plus function elements that are weakly supported, either by not disclosing a specific structure or algorithm or both, may be in danger of invalidation despite a person of ordinary skill in the art being able to practice the invention nonetheless. For those planning on using means plus function elements in software-related patents, err on the side of disclosure, including a step-by-step algorithm for describing the functionality of any software means plus function element, and where possible, including any applicable code.
 Software Patents and the Return of Financial Claiming, Mark A. Lemley, 2012, SSRN-id2117302
 Halliburton Oil Well Cementing Co. v. Walker, 329 U.S. 1 (1946).
 See 35 U.S.C. §112 (f)
 Lemley, at 16
 Lemley, at 19
 See Phillips v. AWH Corp., 415 F.3D, 1303, 1314 (Fed. Cir. 2005)(en banc)
 See Innova/Pure Water, Inc. v. Safari Water Filtration Sys., Inc., 381 F.3D 1111, 1116 (Fed. Cir. 2004)
 See Aristocrat Techs. Austl. Pty Ltd. V. Int’l Game Tech, 521 F.3d 1328, 1333 (Fed. Cir. 2008)
 See Net MoneyIN, Inc. v. VeriSign, Inc., 545 F.3d 1359, 1367 (Fed. Cir. 2008)
 Finisar Corp. v. DirecTV Grp., Inc., 523 F.3d 1323, 1340 (Fed. Cir. 2008)
 See AllVoice Computing PLC v. Nuance Commc’ns, 504 F.3d 1236, 1245 (Fed. Cir. 2007)
 See Noah Sys. V. Intuit Inc., 675 F.3d 1302, 1314 (Fed. Cir. 2012)
 See Default Proof Credit Card Sys., Inc. v. Home Depot U.S.A., Inc., 412 F.3d 1291, 1302 (Fed. Cir. 2005)
 See In re Katz Interactive Call Processing Patent Litigation, 639 F.3d 1303, 1306 (Fed. Cir. 2011)
 See Ergo, at 1365
 Id., at 1372