Well here's an unexpected surprise. A lawsuit brought by the world's largest patent troll, Intellectual Ventures
, and handled on appeal (as are all patent cases), by the notoriously awful
Court of Appeals for the Federal Circuit (CAFC) may have actually killed off software patents
. Really. Notably, the Supreme Court deserves a big assist here, for a series of rulings on patent-eligible subject matter, culminating in the Alice ruling
. At the time, we noted that you could read the ruling to kill off software patents, even as the Supreme Court insisted that it did not. In short, the Supreme Court said that any patent that "does no more than require a generic computer to perform generic computer functions" is not patent eligible. But then it insisted that there was plenty of software that this wouldn't apply to. But it's actually pretty difficult to think of any examples -- which is why we were pretty sure at the time that Alice should
represent the end for software patents, but bemoaned the Supreme Court not directly saying so, noting it would lead to lots of litigation. Still, the impact has been pretty widespread, with the Alice ruling being used both by the courts and the US Patent Office to reject lots and lots of software and business method patent claims.
But this latest ruling, from the very court that upended things nearly two decades ago in declaring
software much more broadly patentable than anyone believed, may now be the nail in the coffin on software patents in the US. The headline, of course, is that the patents that Intellectual Ventures used against anti-virus firms Symantec and Trend Micro, were bunk, because they did not cover patent eligible subject matter. But the part that has everyone chattering
is the concurring opinion by Judge Haldane Mayer, that says it's time to face facts: Alice killed software patents. And Mayer is not some newcomer. He's been at the Federal Circuit since the 1980s and was actually the chief judge in the late 90s/early 2000s when CAFC was at its worst in terms of expanding patent law. And it appears he's been born again into the anti-software patent world. It's... quite a conversion.
, Judge Mayer pointed out that the First Amendment
says that such patents should not be allowed. The whole concurrence is worth reading, but we'll highlight some key points, starting with the First Amendment argument -- which is kind of fascinating in that it goes well beyond what most people had talked about in the past concerning software patents.
“[T]he Constitution protects the right to receive information
and ideas. . . . This right to receive information
and ideas, regardless of their social worth, is fundamental
to our free society.” Stanley v. Georgia, 394 U.S. 557, 564
(1969) (citations omitted). Patents, which function as
government-sanctioned monopolies, invade core First
Amendment rights when they are allowed to obstruct the
essential channels of scientific, economic, and political
Wow! That's actually great to see -- and the kind of argument that we'd hoped to see around copyright. But we'll take it on patents. Here, Judge Mayer notes, the real issue is that this patent would basically harm free expression on the internet, making it ineligible to be patented.
Just as the idea/expression dichotomy and the fair use
defense serve to keep copyright protection from abridging
free speech rights, restrictions on subject matter eligibility
can be used to keep patent protection within constitutional
bounds. Section 101 creates a “patent-free zone”
and places within it the indispensable instruments of
social, economic, and scientific endeavor.... Section 101, if properly applied, can preserve
the Internet’s open architecture and weed out those patents that chill political expression and impermissibly
obstruct the marketplace of ideas.
From there, Judge Mayer notes that if everyone just recognized that the Supreme Court outlawed patents with Alice, the First Amendment questions wouldn't even come up at all.
Most of the First Amendment concerns associated
with patent protection could be avoided if this court were
willing to acknowledge that Alice sounded the death knell
for software patents. The claims at issue in Alice were
directed to a computer-implemented system for mitigating
settlement risk.... Although the
petitioners argued that their claims were patent eligible
because they were tied to a computer and a computer is a
tangible object, the Supreme Court unanimously and
emphatically rejected this argument.... The
Court explained that the “mere recitation of a generic
computer cannot transform a patent-ineligible abstract
idea into a patent-eligible invention.”... Accordingly,
“[t]he fact that a computer necessarily exist[s]
in the physical, rather than purely conceptual, realm is
beside the point” in the section 101 calculus.
From there, Judge Mayer makes the point (that some of us tried to make post-Alice), which the Supreme Court refused to say outright, and which many patent lawyers refused to admit: under the test in Alice, basically all software is unpatentable. And that's fine because software is "a form of language" and we don't patent language.
Software is a form of language—in essence, a set of
instructions.... It is inherently abstract because
it is merely “an idea without physical embodiment,”... Given that
an “idea” is not patentable... and a generic computer is “beside the point” in the
eligibility analysis ... all software
implemented on a standard computer should be deemed
categorically outside the bounds of section 101.
And, finally, it appears that a CAFC judge recognizes (citing a bunch of great amicus briefs and papers about how patents have little to do with incentivizing software development) what many in the software field have been saying for decades: software succeeds in spite of patents, not because of it:
Software development has flourished despite—not because
of—the availability of expansive patent protection. See Brief of Amicus Curiae Elec. Frontier Found. in
Support of Respondents, Alice, 134 S. Ct. 2347 (No. 13-
298), 2014 WL 828047, at *6–7 (“EFF Brief”) (“The software
market began its rapid increase in the early 1980s
. . . more than a decade before the Federal Circuit concocted
widespread software patents in 1994. . . . Obviously,
no patents were needed for software to become a $60
billion/year industry by 1994.”); Mark A. Lemley, Software
Patents and the Return of Functional Claiming,
2013 Wis. L. Rev. 905, 935 (2013) (“Software patents . . .
have created a large number of problems for the industry, particularly for the most innovative and productive companies.
. . . [T]he existence of a vibrant open source
community suggests that innovation can flourish in
software absent patent protection.” (footnote omitted));
Wendy Seltzer, Software Patents and/or Software Development,
78 Brook. L. Rev. 929, 930 (2013) (“Seltzer”)
(“Present knowledge and experience now offer sufficient
evidence that patents disserve software innovation.”); Arti
K. Rai, John R. Allison, & Bhaven N. Sampat, University
Software Ownership and Litigation: A First Examination,
87 N.C. L. Rev. 1519, 1555–56 (2009) (“While most small
biotechnology firms that receive venture financing have
patents, the available empirical evidence indicates that
most software start-ups that receive venture financing,
particularly in the first round, do not have patents.”).
But Mayer goes even further in this discussing four separate problems with the whole concept of software patents, including the fact that the scope of the patents greatly exceeds the importance of what they disclose. Second, he notes that "they provide incentives at the wrong time" -- recognizing a key point we've made for years: that an idea is basically worthless when compared to the actual execution (to me this applies to more than just software patents):
they are typically obtained at the “idea” stage,
before any real inventive work has been done, such patents
are incapable of effectively incentivizing meaningful
advances in science and technology. “A player focused on
patenting can obtain numerous patents without developing
any of the technologies to useful levels of deployment
or disclosure, leaving a minefield of abstract patent claims
for others who actually deploy software.” Seltzer, 78
Brook. L. Rev. at 931. Here, for example, it took no
significant inventive effort to recognize that communications
should be screened for harmful content before delivery.
The hard work came later, when software developers
created screening systems capable of preventing our email
boxes from being overrun with spam or disabled by viruses.
Granting patents on software “ideas”—before they
have been actually reduced to practice—has created a
perverse incentive scheme. Under our current regime,
those who scamper to the PTO early, often equipped with
little more than vague notions about using computers to
automate well-known business and social practices, can
reap hefty financial dividends. By contrast, those who
actually create and deploy useful computer-centric products
are “rewarded” with mammoth potential infringement liability.
The third problem he discusses is the fact that the system is overwhelmed with software patents, "most of which are replete with broad, functional claims" which makes it "virtually impossible to innovate in any technological field without being ensnared by the patent thicket." Nicely put. And because of that:
Software patents impose a
deadweight loss on the nation’s economy, erecting often
insurmountable barriers to innovation and forcing companies
to expend exorbitant sums defending against
meritless infringement suits.
Finally, he notes that software doesn't deserve patent protection because "generically implemented software invariably lacks the concrete borders the patent law demands." As he notes, a patent system only functions when it's clear what the boundaries are of what's covered. But that's not the case at all with software patents.
Software, however, is akin to a work of literature or a
piece of music, undeniably important, but too unbounded,
i.e., too “abstract,” to qualify as a patent-eligible invention.
From there, he suggests that the courts (and I guess the Supreme Court) should just stop punting on the issue and declare software patents dead:
Declaring that software implemented on a generic
computer falls outside of section 101 would provide muchneeded
clarity and consistency in our approach to patent
eligibility. It would end the semantic gymnastics of trying
to bootstrap software into the patent system by alleging it
offers a “specific method of filtering Internet content,”....
The opinion is a great read overall -- and it's the kind of arguments that plenty of folks in and around tech and software patents have been making for years. But to see it come out of a judge's pen, in a patent case, and from CAFC
, is what's really incredible. Of course, as a concurring opinion, rather than the majority opinion, by itself the opinion holds no precedential value. That's too bad. But it does suggest that even CAFC judges are recognizing how ridiculous software patents are these days. It will be interesting to see if Intellectual Ventures tries to kick this up a level to the Supreme Court, where it might risk SCOTUS actually agreeing with Judge Mayer.