Dan Winship (who needs to blog more often ;) asked in a comment on my last post:
[W]hy is the implying-that-Linux-infringes-patents thing even controversial? I thought everyone agreed that because of the way the USPTO hands out software patents like candy, that basically every program larger than “Hello World” infringes more patents than there are atoms in the observable universe.
This after I said:
[There is a] sense that Novell is now competing on unfair terms; terms that amount to a tacit acceptance that “our” code is patent infringing (if opensuse isn’t infringing, why is Novell paying so much money?)…
Dan’s question is a fair one, because I didn’t explain myself. I think what is at issue is not that Novell tacitly admitted it, but in particular the way Novell admitted it.
Past tacit admissions have always come with one of three disclaimers: (1) this admission demonstrates that the system is insane and should be reformed (e.g., the European anti-software patent movement) (2) this tacit admission comes in the context of a Mutually Assured Destruction pact (e.g., OIN) (3) the tacit admission comes in the context of a discussion of monopoly power in the software industry being compounded by a competitor’s refusal to license patents (e.g., Microsoft’s patents on Office Open XML.) In each case, the disclaimer made it clear that the patent (and not the patent infringer!) was the real source of the problem.
Novell’s admission came with none of those disclaimers, and as a result looks like a tacit acceptance not just of the patents, but of the patent system and the monopolistic patent owner. It can even be read to weaken all of these disclaimers when the rest of us use them: (1) the patent system must be sane, after all, it led to protection of all these nice volunteers, in a nice market-oriented way (2) we don’t have enough weapons for MAD (3) MS is now reasonable and the onus is on commercial GPL users to cooperate.
[This last one, IMHO, is the most damaging one; it will absolutely come out the next time MS has discussions with EU antitrust regulators, and frankly, with justification. Once you grant the legitimacy of the patent system (which antitrust regulators must do), explaining why GPL users shouldn’t negotiate with MS is suddenly damn hard to do if MS is publicly willing to negotiate, no matter how false that willingness is.]
Using these disclaimers, free software potential patent infringers were effectively on the offensive- those counter arguments put the burden of proof was on MS and other large patent holders to prove that the patent system was reasonable. This move has potentially shifted the burden of proof- the question now may well be ‘why isn’t RH signing the same deal? After all, Novell did’. I suspect that MS did this deliberately (given how quickly after the deal they started discussing using this against others) but of course have no proof, and have no idea if Novell thought this through or just got caught unaware.
It will be months or years before we know if all this has actual concrete impact, or if it is just a temporary distraction. But it is there, and I suspect that a lot of people understand this, which accounts for a lot of the anger, even if they can’t fully articulate what it was that Novell ‘sold out’.
[Tangentially, you can argue that the patent system is overall valuable and correct, or at most needs to be corrected, and that hence all of these things are (overall) not a bad thing. This is certainly the position of some of erstwhile free software allies, like IBM. But the Novell agreement is still potentially a bad thing for them, too; but that’s another blog post (quite possibly a blog essay ;) and hence it’ll have to wait for another day.]
I would disagree with your points about weakening. 1) The patent system does not have to be sane for this deal to happen, nor do the parties have to jointly agree it should exist; it just has to exist and present a real threat that has been upheld by the courts. If all software patents were terminated, then this would not be an issue. 2) Novell did not preclude Microsoft from suing Novell. This suggests they still believe that MAD is an effective deterent. 3) There were, and are, a number of commercial (and large non-commercial organizational) potential users who would not deploy any GPL (or frequently other free or open source) software in their organization as they already felt threatened.
What I think has really happened is that it has become clear that Novell, as a company, does support software patents and their management is happy to use patents to create an uneven playing field. I do not think that this has changed one bit from two weeks ago, but previously Novell was not as direct in saying so.
Given,
a. the OIN patent pool is mostly (if not totally?) contributed by Novell (see Miguel’s posts recently);
b. it’s not obvious that any of the patents on their site (except for the first in the list) have been transferred to the independent OIN organisation;
c. Microsoft have bought themselves a patent truce with Novell;
Have Microsoft bought themselves protection from OIN and the mutually assured destruction? Are they now free to persue those who OIN supposedly protects?
And what about going beyond MAD to Dr Strangelove’s strategy?
One could create a doomsday device.
An anonymous wiki itemises all patent violations of all GPL code (which GPL developers can choose to remain blissfully ignorant of).
All GPL publishers submit regular patent violation confessions (encrypted) to a central repository.
Have a publicly assured unknown private key, with doomsday public key and then everyone encrypts their own private key with this. Verifiable.
One chooses the length of key such that it will take ~N years to crack (even given Moore’s law and public distributed computing) where N is around the time one expects patent itemisation to settle down.
When the private key is discovered all patent violations effectively become declared and lawyers take over the world – unless someone’s got a brain cell somewhere.
> Dan Winship (who needs to blog more often ;)
Yeah, yeah. :-) I’ve written three different posts about the MS deal now, and then thrown each of them away rather than posting them, because “I don’t speak for Novell” only gets you so far. (But parts of one turned into the comment I left on your last past, and parts of another turned into this message: http://lists.opensuse.org/opensuse-project/2006-11/msg00027.html.)
Anyway, thanks for the reply. I suspected this was what you were getting at, but I wanted to be sure (and I wanted someone to confirm that I wasn’t crazy, because so many of the other people commenting on the deal seem to be completely in denial about the possibility of patents-in-Linux).
Personally, I think Novell *does* still think MAD is sufficient, although I agree that we are not going out of our way to communicate that at the moment. But now that we’ve addressed all of the potential large customers saying “We don’t want to migrate to Linux because of patent issues”, I suspect it won’t be long before some of them switch to “We don’t want to migrate to Linux if we’re going to be locked into Novell’s offering”. :-)
Alex:
> a. the OIN patent pool is mostly (if not totally?) contributed by Novell
> c. Microsoft have bought themselves a patent truce with Novell;
I’m fairly certain (a) is wrong, although I don’t know for sure. At any rate,
(c) is definitely wrong; Microsoft and Novell have agreed not to sue each other’s
*customers*, but have reserved the right to sue *each other*. Given that Novell has said that it still won’t ship patent-infringing software, if Microsoft did launch a massive patent strike against Linux, it seems like invoking MAD would be the quickest way to deal with it. But who knows.
The nightlight/monster under the bed analogy is pretty decent one. The problem, of course, is that the parents are divorced, the nightlight is built into the wall, and now unless dad buys a nightlight to match mom’s, he’s screwed. And forget Grandma Deb- she can’t afford a multi-million dollar nightlight, so Johnny is never going over there again. Of course, it turns out the nightlight maker is also a bed monster in his spare time, so it is in his interest to create a sense of false security in Johnny :) So yeah… buying the nightlight, in and of itself, is good for Johnny, good for mom, etc. But it has ripple effects that are hard to completely discount.
Tangentially, I wish I were a fly on the wall and had been able to see the discussion with any customers actually scared of patent infringement lawsuits. It’s hard to find any examples of even the worst patent trolls suing customers, and the patent protection Novell and RH were already offering was pretty ironclad to start with. As Mike Dolan (IBM biz strategy guy) pointed out in this comment, the hypo there is like someone suing Hertz if they held a patent on ABS, instead of Ford- it isn’t something I imagine Hertz loses much sleep over. I realize not all customers are sane, and some are more litigation-averse than others, but if this is purely customer-driven, either MS’s FUD has been really, really effective (quite possible) or some of Novell’s customers are really, really paranoid.
[…] The document has moved here. […]