About the Authors

  • The Authors and Contributors of "Patent Docs" are patent attorneys and agents, many of whom hold doctorates in a diverse array of disciplines.
2018 Juristant Badge - MBHB_165
Juristat #4 Overall Rank

E-mail Newsletter

  • Enter your e-mail address below to receive the "Patent Docs" e-mail newsletter.

Contact the Docs


  • "Patent Docs" does not contain any legal advice whatsoever. This weblog is for informational purposes only, and its publication does not create an attorney-client relationship. In addition, nothing on "Patent Docs" constitutes a solicitation for business. This weblog is intended primarily for other attorneys. Moreover, "Patent Docs" is the personal weblog of the Authors; it is not edited by the Authors' employers or clients and, as such, no part of this weblog may be so attributed. All posts on "Patent Docs" should be double-checked for their accuracy and current applicability.
Juristat #8 Overall Rank


« Conference & CLE Calendar | Main | Commil USA, LLC v. Cisco Systems, Inc. (2015) »

May 25, 2015


The sentence "After all, software instructions are indeed just another form of data, and without a medium to be stored on, lack physical structure." goes too far, and like the court itself, misses what software is.

But setting that aside for the moment, let's look at another impact here: copyright.

If this particular definition were to hold, then a gaping hole (the size of the Grand Canyon) has just been blown through copyright protection. All one has to do is copy the "signal" and one is free and clear. Since copyright also extends only to that which is captured in a tangible media, anything intangible is not - and cannot - be protected.

Does this extend to copying the signal of movies and music?

Pandora called - she wants her box back.

US readers should bear in mind that In re Nuijten and Digitech do not apply in Europe, and that the EPO will in appropriate cases grant claims to signals. However, because of the severe tests for added subject matter and intermediate generalization under a.123(2) EPC it is essential to include appropriate statements of invention in the specification if such claims are to be pursued here.

That being said if we analyze claim 60, apart from a broad statement of the goal of providing a universal speech-recognition interface without an indication of any new and surprising or ingenious technical features that make this possible, there are specified input means, output means and audio playback means. These features are present in the now fairly venerable iMAC on which this comment is being typed, were present in the Epson HX20 portable computer which was the first that I owned in the early 1980's and were also present in the Sinclair Spectrum that I bought for my children at that time. It is a little surprising that this claim was considered litigating at all, still less as far as the Federal Circuit.

If software is not data, then what, pray tell, is it?

@ Mike

From my recollection software is an ordered assembly of instructions for be supplied to a computer processor. The processor operates on data in accordance with the instructions. As concepts these are quite distinct, as any software engineer will tell you.

Software instructions are data - they can be read, written, modified and deleted like any other form of data. They may have a functional difference from other forms of data, in that they can be used to tell a processor what to do, but they do not have a structural difference.

One of the first programming assignments I had at university was to parse through an executable file and look for certain types of information. The software code was the data to another program.


Careful there with your "definition" of data, as along that logical abstraction, everything is "data" (you fall into the Tegmark camp).

"misses what software is"

if the claim recited "what software is", there would not have been a problem. the claim recites "what software does" (function laden language) and does not fit into the four statutory categories. my grandmother can come up with dozens of items to magically do stuff she wants done, but that's not patentable. squeeze the magic out of the claim and you'll make progress.

"If software is not data, then what, pray tell, is it?"

at what point in time of the existence of "software" do you want a definition? it's not patentable when it exists initially in the mind of the programmer. it's not patentable when written down on paper. it's not patentable when stored in memory. it's not patentable as an ordered assembly of instructions. it's patentable when compiled because it is then a MACHINE--an electric circuit where a higher voltage at one node causes electrons to travel toward a node at a lower voltage. does one claim it that way? not if the prior art does not require such narrow language. perhaps my generation of electrical engineers is obsolete--where we had to write programs and compile logic circuits. a processor is not a black box to my classmates. i understand that college coursework is so specialized these days that graduates majoring in software need not understand hardware details, and vice versa. claiming hardware with high level functional language will not get you a patent, a disclosed and claimed novel and non-obvious algorithm will.

The point (well, my point anyway) is that the CAFC made a reasonable determination that the claim covered software as data, bereft of a tangible medium, and is therefore unpatentable. This is in line with precedent and makes technical sense, as opposed to what the Supremes did with Alice.

With respect to how Claims 60-68 might have been written, note the other claims in the `273 patent that are directed to various methods and "An editor work station" and were held indefinite and not infringed in this suit. Claims 73 and 74, directed to a "computer usable medium," were not asserted against Microsoft, though they and Claims 60 - 68 were successfully asserted against Dragon in AllVoice Computing PLC v. Nuance Communications, Inc., No. 06-1440 (Fed. Cir. Oct. 12, 2007). In that earlier case, the Federal Circuit reversed district court rulings that these claims were indefinite or were invalid for failure to disclose the best mode of practicing them.

Also there are two related AllVoice patents, 5,857,099 and 6,961,700, with further sets of claims including "computer usable medium" and method claims. So there are a variety of claims of various types, some of which have been held invalid and some not.

I confess that I do not understand this technology and am not inclined to study the claims in these patents to assess their differences and similarities to Claims 60 - 68 of the '273 patent. It would be interesting to know just how the infringement cases were laid out in both AllVoice lawsuits. It would be interesting to know just how the district court held most of the asserted apparatus and method claims to be indefinite and why Allvoice did not appeal that decision.

Returning to the AllVoice Computing PLC v. Nuance Communications, Inc. decision, I see that the panel found corresponding "structure" in the specification to support all three "means" expressions of Claim 60. But, in pursuing Microsoft, Allvoice conceded that these "means" limitations were limited to software instructions without any hardware limitations. Why was AllVoice forced to make this concession?

Would someone address these two cases?

What I find interesting is the lack of focus on the use of means-plus-function claim language. As you know, MPF claims inherently limit the claim element to structural embodiments (disclosed in the specification, and equivalents, etc.).

Thus, if the holding was that the MPF claim element didn't have corresponding structure in the specification, it would seem to be more appropriate to invalidate as indefinite under 35USC112.

What an interesting time for patent law!

That Claim 60 was a disaster from the git-go.

1. The preamble should have explicitly referenced statutory matter -- i.e. "an interface device," which is what is described in the spec.

2. The novel/non-obvious characteristic should not be in the preamble, it should be cited as either an element or as a limitation of an element. I would have used MPF here:

means for enabling operative coupling of a speech-recognition engine to at least any one of a plurality of different computer-related applications;

This MPF is adequately supported by flo charts in the spec so as to convert the algorithm into "structure" ala the recent Eon case.

It was a disaster in the making. Didn't affect the outcome in this case, however. The CAFC's bizzare upholding the DC's dismissal of the case for lack of timely filing of a motion to amend the claims deep six'd all the litigated claims.


Please compare and contrast the findings re Claims 60 - 68 of the '273 patent in

Allvoice Developments US, LLC v. Microsoft Corp. (Fed. Cir. 2015) with

AllVoice Computing PLC v. Nuance Communications, Inc., No. 06-1440 (Fed. Cir. Oct. 12, 2007)

Help us understand.

In one case, the issue was under Section 101, in the other, the issue was under Section 112. Thus, the analyses and outcomes can easily be different.

The comments to this entry are closed.

April 2024

Sun Mon Tue Wed Thu Fri Sat
  1 2 3 4 5 6
7 8 9 10 11 12 13
14 15 16 17 18 19 20
21 22 23 24 25 26 27
28 29 30