Missing the target with functional claim language

As a companion to our recent article on eligibility, the United States Patent and Trademark Office (USPTO) also announced revised 2019 guidelines for applications with ‘functional limitations’ in their claims. These new guidelines are particularly relevant for computer-implemented inventions, including artificial intelligence (AI) and machine learning (ML).

What’s the problem with ‘functional limitations’?

Patents involving computer software typically define the invention with intentionally broad and vague language. A particular goal or objective is explained, but often without giving explicit details of how that is to be achieved, these ‘implementation details’ instead being left for the notional ‘person skilled in the art’. The aim is to get a broad monopoly over the invention, regardless of how a competitor decides to code that particular feature. In the past few years, the US Court of Appeals for the Federal Circuit have heard several cases with these ‘functional’ claims, but held that the patents involved were invalid for ‘indefiniteness’. Therefore, it is important to be aware of this balance between broad claim scope, which is good for catching infringers, as against the risk of the patent being struck down as ‘too broad’.

What is in the new guidance?

‘Functional limitations’ which define the invention in purely functional terms are given special treatment under section 112(f) of the US patent law, and must also meet the requirement of ‘definiteness’ under section 112(b). These two sections are being applied together in Part I of the new guidance to put a limit on purely functional claiming.

Also, patent applications have to describe an invention with enough detail and precision that it can be carried out by the notional person skilled in the art. Part II of the new guidance covers this requirement for written description and ‘enablement’ under 35 USC 112(a).

“Means” and ways

The words in a patent claim are normally given their ‘plain, ordinary meaning’, but under section 112(f) a functional limitation is restricted only to the examples given by the patentee in the description, i.e. “the structure, material or act described in the specification as performing the entire claimed function and equivalents to the disclosed structure, material or act”.

Under US law there is a long-standing presumption that the words “means for” will trigger the special interpretation of section 112(f) – e.g. “means for displaying”. However, as emphasised in the new guidance, the same rule can be invoked by any other generic placeholder plus purely functional language. Hence, words such as ‘mechanism for’, ‘module for’, ‘device for’, ‘unit for’, ‘component for’, ‘element for’, ‘member for’, ‘apparatus for’, ‘machine for’ or ‘system for’ a particular function can all trigger 112(f). According to the guidance:

“Note that there is no fixed list of generic placeholders that always result in 35 U.S.C. § 112(f) interpretation, and likewise there is no fixed list of words that always avoid 35 U.S.C. § 112(f) interpretation. Every case will turn on its own unique set of facts.”

By contrast, 112(f) is avoided when a generic placeholder term is modified by ‘sufficient structure, material or acts for performing the function’. For example, the termdetent mechanism’ has a specific meaning for a mechanical engineer to avoid 112(f), whereas ‘colourant selection mechanism’ or ‘distributed learning control module’ in a software claim had no such widely understood specific meaning. It is helpful to look up the words in dictionaries and show they have the intended recognised meaning in practice as a noun denoting structure, as with ‘CPU’, ‘memory’, ‘network card’ and so on, and to avoid the temptation to coin new technical-sounding compound terms or phrases.

The hidden danger of 112(b)

The new guidance highlights that, according to the case law of the Federal Circuit, the patent application must also disclose an algorithm for performing the specified computer function, in order to meet ‘definiteness’ under 112(b). In other words, the description needs to include “a finite sequence of steps for solving a logical or mathematical problem or performing a task”, which will create a special-purpose computer to achieve the specified function. This algorithm can be expressed “in any understandable terms including as a mathematical formula, in prose, or as a flow chart, or in any other manner that provides sufficient structure”. It is not enough that the skilled person could create such an algorithm based on the functional description in the application, which would satisfy enablement under 112(a), but instead the algorithm must be explicitly present in the application. Failure to disclose an operative algorithm in sufficient detail leads to invalidity for ‘indefiniteness’ under 112(b) when interpreting the claim according to the special treatment of functional limitations under 112(f).

Summary

From our UK and European perspective, those of us preparing new patent applications will look to avoid invoking 112(f) if at all possible, by making careful use of meaningful claim language, and to prepare the detailed description with plenty of practical details (e.g. mentioning specific APIs or DLLs relevant to implementation of the invention) to also withstand a challenge of indefiniteness under 112(b), over and above the regular requirement of enablement under 112(a). At Appleyard Lees we work closely with many US colleagues to give clients the best possible advice in this difficult field.

Link:

2019 Guidance on Functional Limitations for Computer-Implemented Inventions

Author

Related posts