Microsoft's Open Standards Fairy Tale
Published 13:12, 15 March 12
Regular readers of this column will know that I often write about the issues of open standards and FRAND vs. RF licensing. One particular column that explored this area appeared back in October 2010.
I was interested to see that it was quoted in a paper written by Stephen Mutkoski, Global Policy Director, Microsoft Corporation. That paper was referenced and discussed briefly in a new post by Mutkoski entitled "Public Procurement of ICT: Debunking Myths around Open Source Software and Royalty Bearing Standards" (provocative, lui?). You will not be surprised that I have a few comments to make on these.
Here's the abstract of that post:
Patent royalties on standards are not restricting the market for open source ICT solutions and government policy makers should resist calls to “level the playing field for open source” through procurement policies that seek to eliminate royalty bearing standards from consideration in public procurement.
Well, I have to give Mutkoski points for honesty is admitting that Microsoft does not want a "level playing field for open source" - something most of us rather suspected. But his main argument is in the paper referred to, so I shall chiefly address that here.
Part I is about "Debunking the Myth: FRAND “Discriminates against” FOSS Developers", and this is the key argument:
if the FRAND conflicts with FOSS claim were true, we would expect to see a “dead zone of implementation” in the FOSS world around these areas of standards (and that “dead zone” would likely extend far beyond merely the domain of standards, since there would be other areas of innovation not essential to the implementation of standards where FOSS developers would be unable to tread). But more importantly, we would also expect to see detailed impact assessments from the world’s numerous FOSS distributors, identifying the specific standards that they are unable to implement and in fact do not implement as a consequence of the claimed deadlock between FRAND and
FOSS. But we see neither a lack of implementation of standards by FOSS developers nor any articulation of specific standards (or non-standardized areas of innovation where patent royalties must be paid) where the numerous successful FOSS companies are precluded from implementing technology.
Such "dead zones of implementation" aren't seen for the simple reason that open source is intimately bound up with the Internet - it has both powered it and been powered by it. Crucially, the Internet is a FRAND-free zone - the W3C insists on RF licences for its standards. So open source has rarely been faced with FRAND standards that it might want to implement.
The rapid growth and amazing power of the Internet is in fact one of the strongest arguments for RF rather than FRAND, and proof that areas where FRAND is more have suffered as a result. That's why it's crucially important for interoperability standards to be RF not FRAND - so that the same rapid innovation can take place. That's an issue because the success of open source means that it is now being used in more traditional areas where FRAND licensing is more common. It's also why there haven't been many cases so far of where FRAND has been an issue. But the point about framing policy is to prepare for the future, not lock tomorrow's innovators into the past.
Part II of the paper claims to offer "Growing Evidence that FRAND and FOSS are Compatible", and cites:
several instances where FOSS distributors have paid patent royalties, either specifically on standards or on other non-standardized technologies that are incorporated into their FOSS offering(s), which clearly undermine the claim that FRAND precludes FOSS implementation of technologies:
Linux distributor Red Hat has entered into royalty bearing licenses on at least two occasions, with patent holders Firestar and Acacia, to cover distributions of Red Hat’s FOSS products.
Linux distributor Canonical pays a patent royalty in connection with its license for the H.264 standard that it implements in its Ubuntu operating system.
Roku, Inc., pays a royalty in connection with its license for the H.264 standard that it implements in its Linux-based set top box devices.
There are numerous patent infringement lawsuits related to the Android platform, which Herman predicted would be resolved via royalty bearing patent licenses.
Let's look at these in turn.
I have discussed the first of these in my October 2010 column. The key point is that Red Hat paid a lump sum to cover all open source projects, not just its own use. In other words, Red Hat acted as a Fairy Godmother to the open source community. I'll return to this point at the end of this post.
The Canonical licensing only applies to pre-installed copies of Ubuntu: that means a per-copy licence can be paid, since these are limited and countable. The deal does not apply to Ubuntu in general, and therefore is not an example of how royalty-bearing licensing is compatibility with free software, but to do with the licensing for hardware sales.
The Roku is a also hardware system, and is similar to the pre-installed versions of Ubuntu that come with an H.264 licence.
The most interesting example that Mutkoski cites is the last one, because, of course, those "numerous patent infringement lawsuits related to the Android platform" are largely Microsoft drawing up deals with various Android manufacturers. As I've pointed out elsewhere, we don't know the terms of those deals, which could, theoretically, even involve Microsoft paying the manufacturers, at least in kind. Why would it do something as crazy as that? Precisely so that it could claim - as here - that Android infringes on Microsoft's patents, and also, as a bonus, that free software is compatible with such licensing.
But just because Samsung is willing to accept whatever secret terms Microsoft is offering, doesn't mean that in general free software can always be reconciled with royalty-bearing licences - its a completely unjustified extrapolation.
The final section of Mutkoski's paper is entitled "FOSS Advocates Concede the “FRAND discriminates against FOSS” Mantra is Not True" - which in my case at least, is false.
Here's the key passage:
The question that never seemed to be resolved though is whether there is any real licensing conflict between so-called FRAND terms (including a royalty) and the many FOSS licensing models or whether by contrast this is just a matter of philosophical objection to software patents (arguably a widely held belief among FOSS developers. When pressed on this point, Moody concedes in a blog post “let me be clear: by that I do not mean legally incompatible.” Katz makes a similar concession in his Twittersation with Moody: “it depends on what you mean by RAND licences[-] It is possible to draft a compatible licence.” Moody went on to suggest that even if most or all FOSS licenses are legally compatible with FRAND, FRAND is “incompatible in practice” with FOSS.
So let's look at what I wrote:
Again, as I and many others have pointed out, the BSA solution - “a reasonable fee” - is simply not compatible with leading free software licences. And let me be clear: by that I do not mean legally incompatible, I mean incompatible in practice, because it is not possible to account for the uncountable because unknown downstream users.
What I meant there was even assuming that "a reasonable fee" was legally compatible - and IANAL, so I'm willing to grant that it might be possible through careful framing - it simply isn't in practice. And you don't have to be a lawyer to see why: the unrestricted distribution model of open source means that it is impossible to keep tabs on how many copies have been made, and so it is impossible to pay a per-copy licence. That is just simple, inarguable logic, which is why I concentrated on this much stronger statement, rather than one about the legality of such arrangements, which are ultimately for courts to decide. But there can be no doubt that FRAND discriminates against free software.
Mutkoski goes on:
Moody spent some time elaborating on what he meant by “incompatible in practice” and his comments suggest that his real gripe with FRAND is that “Smaller companies are ... at a disadvantage in this situation when they cannot afford to pay licences as easily as big companies.” But this has nothing to do with FOSS and is just a replaying of arguments against patents per se that have been repeatedly raised and rejected. More importantly, there are many successful companies, large and small, that have licensed patents, whether in connection with FOSS products or products that they distribute under traditional commercial licenses.
These statements are powerful evidence that, when pressed for explanation and proof that FOSS and FRAND cannot co-exist due to requirements in FOSS licenses, key advocates of this claim have admitted it is simply not true. That is an important point: all these FOSS advocates are suggesting is that patents appear to be an impediment for some developers who cannot pay royalties (which is true whether the developer uses FOSS licenses or not) and conceding there is no absolute bar for FOSS developers as a result of restrictions in FOSS licenses.
But free software projects are qualitatively different from smaller companies or developers using conventional software licences. For the latter, the licence fees can be reduced to a level such that they can pay; that's not possible for open source - there is no level of per-copy payment that can be accommodated because open source allows software to be copied freely, and hence uncontrollably.
It's revealing that once more Mutkoski displays his contempt for "levelling the playing fields", and emphasises his belief that large companies should be allowed to use their established positions of power to shut out innovative newcomers. That's probably to be expected from someone working for a company that was found guilty of abusing its monopoly position on both sides of the Atlantic.
The EU anti-trust action is very revealing - and relevant - in this regard. Here's what happened:
In 2004 the European Commission found Microsoft guilty of monopoly abuse in the IT marketplace and demanded that complete interoperability information be made available to competitors. Microsoft objected to this decision and was overruled in September 2007 by the European Court of First Instance (CFI). The CFI found Microsoft guilty of deliberate obstruction of interoperability and upheld the obligation for Microsoft to share its protocol information.
The Samba Team has decided to make use of Micrsoft's obligation under the European judgements. Through the Protocol Freedom Information Foundation (PFIF), network interoperability information has been requested and a one-time access fee of 10.000 EUR is being paid to give Samba team full access to important specifications.
This is a perfect example of why this is not just about big companies versus little companies. Samba is not a company, and has no funds, and yet it has created and continues to develop one of the most widely-used pieces of software in the world. According to Mutkoski, it should be perfectly acceptable that this group of public benefactors - for that is what they are - should be denied access to key information held by a company that was found to have abused its monopoly, simply because that group has no funds. And that, in general, is what will happen if open standards are allowed to be FRAND, and not RF.
And to the argument that Samba did, indeed, obtain that key interoperability information, which proves the current system works, and doesn't need changing, consider this. Samba obtained that information only because, once more, two things happened: a one-time access fee was charged, and a Fairy Godmother appeared to pay it.
But the point is, a belief that companies will always grant one-time fees, and that Fairy Godmothers will always magically turn up in the nick of time to save open source projects that otherwise will be excluded from key sectors, is not a basis for European policy making. The European Commission must plan on the basis of reality, not fantasy. The only rational way of protecting open source projects and allowing them to continue to make their contributions to society is to insist on RF, not FRAND licensing for open standards.
Of course, there is an alternative which Mutkowski may like to consider: that Microsoft commits irrevocably, perpetually and unconditionally to take on the role of Fairy Godmother by covering all FRAND fees that may be demanded from any open source project for implementing open standards.