Blog exclusive:

Hallelujah!

I wrote a story for my publication, BNA's Patent, Trademark & Copyright Journal, that assessed the arguments in CLS Bank International v. Alice Corp., a case the Federal Circuit will hear en banc on Feb. 8. In that short article, there was no way I could do justice to three briefs from the parties and 22 briefs of friends of the court.

I want to use this blog entry as an opportunity to highlight one small portion of a brief filed by former Solicitor General Paul D. Clement on behalf of International Business Machines Corp. I want to say, "Hallelujah."  IBM made a distinction about computer software that everyone should understand before making a blanket statement that all software should be patent eligible or all software should not be patent eligible.  And that's what I want to discuss now.

First, understand: I spent a good portion of my life programming computers and another good portion of my life managing people who programmed computers. And for someone with that kind of background, it has been annoying, to say the least, to hear arguments from patent attorneys that lump all software into a single technology.

IBM argued otherwise. See the brief here. Look at the chart on page 20. Software programs live in layers between computer hardware and end users:

• In the layer "above" the hardware, you can see firmware and device drivers.
• Above that, the operating system.
• Then "middleware" such as webservers, databases, and Java.
• And finally, what the user sees, applications.

Unless you go back to inexplicably muddled Supreme Court decisions in the 1970s (Benson and Flook), virtually every complaint that software should not be patent eligible uses a questionable application as definitive proof. But there are three other bullet points -- a group IBM calls "system software" -- to consider as well.

What does system software do? In intent at least, every example of a program implemented as system software makes a computing machine work better. Faster. With more flexibility. Whatever. But better.

We give patents to improvements that make machines work better. You can look it up: "Whoever invents or discovers any new and useful … machine, … or any new and useful improvement thereof, may obtain a patent therefor …." 35 U.S.C. §101. I see no reason why we would give a patent for making a harvesting machine work better but not give one that makes a computer work better. 

But then there's application software, and I'll charge IBM's brief with being a bit facile here. "Inventions residing in system software layers or throughout most of the application layer should not be susceptible to patent eligibility issues," according to IBM. But to say "most" of the layer means there must be some distinguishing aspect of the application programs that merit patent eligibility and those that don't.

You can judge for yourself whether the brief does that well. But I think it starts from a faulty premise. To make its point, IBM uses as examples an extremely small sample of application software: "This software includes, for example, word processors, spreadsheets, email programs, web browsers, and computer games."

Of course, an e-mail program can't exist outside the context of one or more computing systems. Same with a web browser. It's hard to argue that any of those programs are not "new and useful improvements" of a machine, or of networking technology, or both.

But the bulk of application software is of a different kind. The vast majority of application software does not make an improvement in computing. Rather, computing makes an improvement in the application. The "technology" field being developed by most application software is not computer science, it is whatever field the application lives in.

In the CLS Bank case, the technology being improved is bank transaction risk management. The proper Section 101 question is not whether the computer makes it patent eligible, but whether it is an advance in the field of what Prof. John Duffy calls "financial engineering." Those of you who want to limit software patenting will undoubtedly say that term is an oxymoron. But whichever side you're on, your argument is not about the software, it is about whether a technology (other than how a computer works) or an idea is being improved.

Claims such as those in CLS Bank are, in fact, a closer call than most application software. It's pretty easy to argue that financial engineering doesn't exist without computers. Not so with most software applications though.

The genesis of the most typical application is a customer telling a salesperson that X ought to be computerized. X is a technology being improved. Let's see what then happens.

The marketing department then asks whether multiple customers will buy X if it were available. If so, it prepares a requirements document, which soon leads to a functional specification. A system designer specs out technical details, and then the programmer writes the code.

What I think most patent stakeholders miss is that X is effectively enabled before it gets into the hands of the system designer. Or at least, from my reading of most software patent claims, the limitations track what exists in, at best, a functional spec, and probably a requirements document.

So what, in fact, is X at that point? It is still a marketing idea.

And it is the poster child for analysis under Mayo v. Prometheus. The "inventive concept" is the marketing department's idea of what would be a cool application that would make the company a lot of money. The acts that turn that cool idea into something that operates on a computer comprise "well-understood, routine, conventional activity."

You, dear reader (not to mention I, dear reader), may not like what the Supreme Court said in Mayo v. Prometheus, but they have, after all, the final word. And until a better explication of Section 101 standards makes sense of the opinion, we all have to live with it.

I live with it in the following way. I interpret the opinion to breathe life into the word "new" in Section 101, distinguished from "novelty." I think what they were trying to say was that the newness of a patent claim must be embedded in an inventive concept, rather than surrounding it. And that inventive concept had better be something technological. And however you want to define technological, marketing ideas (and statistical correlations) aren't it.

And under that interpretation, I disagree with IBM that an application is rarely defeated under Section 101. Yes, some application software is patent eligible, but probably in the reverse percentage that the IBM folks believe. It boils down to exactly what X is.

When the impetus is "X should be computerized," the "innovator" is not saying X is new (in the Mayo Section 101 sense), but that subjecting X to computerization is new. And that is what Mayo said was conventional. And if X is describable only as a marketing idea, it doesn't pass the inventive concept smell test.

When X is of the class of an email system or a web browser, the innovation lies in X, is embedded in X. Computer games, sure, those too. Make a case for word processing and spreadsheets and I'll probably agree.

Unless Mayo is overturned, that is the best I can do. Instances of patent eligibility of application software are likely rare. Make your case to the PTO, but don't base it on the fact that a computer is involved.