LWN: Comments on "Open Standards in Europe: FSFE responds to BSA letter"

Thanks, Jake

spaetz — Fri, 29 Oct 2010 09:25:32 +0000

Another happy customer of comment filtering here. Thanks :). It would even be more satisfying if I saw the number of filtered comments next to the individuals at http://lwn.net/MyAccount/spaetz/commfilters/

Thanks, Jake

corbet — Thu, 28 Oct 2010 21:19:24 +0000

Filtering is Jake's work, actually...I'll pass your thanks on to him :)

Thanks, Jon

stevem — Thu, 28 Oct 2010 21:06:44 +0000

The comment filtering is a *wonderful* feature.

Open source must be able to deal with patent licenses if necessary

nix — Mon, 25 Oct 2010 22:38:56 +0000

If a legal dispute doesn't make it to court, is it a case at all? I was assuming 'no', but IANAL so ICBW.

Open source must be able to deal with patent licenses if necessary

vonbrand — Mon, 25 Oct 2010 11:02:49 +0000

AFAIU, there is no "lack of cases"; there are plenty of those, they just didn't make it to court (or into the headlines) as they were resolved (in GPL's favor) silently. That in itself is very powerfull evidence for GPL, even more so than a few highly publicized court wins.

Twittersation?

nix — Sun, 24 Oct 2010 23:36:16 +0000

That's not even the most ugly neologism I've seen this month. (But it is pretty ugly.)

Far less elegant

man_ls — Sun, 24 Oct 2010 22:29:32 +0000

Why bother, it's like punching a tar pit (and I've been there). I sure hope that jthill avoids tar pits in his profession with the same elegance that he has done here.

Twittersation?

man_ls — Sun, 24 Oct 2010 22:18:57 +0000

My fellow LWN readers were too kind to let it pass, but I cannot. So you had an interesting conversation on twitter, and you regularly shorten this kind of happening to "twittersation" to make it sound cool. Well, it isn't. Sorry.

Not the implicit red herring again

man_ls — Sun, 24 Oct 2010 22:09:46 +0000

GPLv2 doesn't really do that (only implicitly, which creates uncertainty).

Really? The GPL v2 says right in the preamble:

Finally, any free program is threatened constantly by software patents. We wish to avoid the danger that redistributors of a free program will individually obtain patent licenses, in effect making the program proprietary. To prevent this, we have made it clear that any patent must be licensed for everyone's free use or not licensed at all.

Later clarified in sections 7 and 8.

Open source must be able to deal with patent licenses if necessary

FlorianMueller — Sun, 24 Oct 2010 15:55:50 +0000

What's the point you were trying to make? That they may still be, four years later, in quiet negotiations with Novell?

Open source must be able to deal with patent licenses if necessary

nix — Sun, 24 Oct 2010 15:54:13 +0000

But... the FSF cared about copyright, too, and had enough resources for quiet negotiation with companies to continue throughout the 90s, in preference to a court case, because they would rather have a not-violated GPL than some test case to make Florian happier. So your argument falls at once.

Open source must be able to deal with patent licenses if necessary

FlorianMueller — Sun, 24 Oct 2010 15:40:13 +0000

On such a central issue it is. Everyone knows how much the FSF cares about patents. Also, you can't compare the resources available to the FSF (including its satellites such as SFLC) these days to the ones in the 1990s.

Open source must be able to deal with patent licenses if necessary

nix — Sun, 24 Oct 2010 15:34:08 +0000

Fifteen or twenty years passed between the drafting of GPLv1 and the first successful lawsuit brought under it, but it would not have been valid to say, in the interim, that the GPL was unenforceable and meaningless rubbish because nobody had enforced it.

A lack of data (-> a lack of cases) is not the same as negative data (-> lost cases).

Not OK

FlorianMueller — Sun, 24 Oct 2010 11:39:09 +0000

Of course the patent-related language changed between GPLv2 and (the early drafts of) GPLv3. But the fact of the matter is that a stricter language on patents was a GPLv3 design goal from the outset. So it's perfectly reasonable (except for people who want to create unreasonable doubt) to say that there's a continuum in terms of patent-licensing-related ridigity ranging from GPLv2 to early drafts of GPLv3 to the final version of GPLv3:

GPLv3_final > GPLv3_early > GPLv2

Let X be on said continuum the point at which you can prevent the Novell type of deal:

GPLv3_final > X > GPLv3_early (according to RMS)

Thus, X > GPLv2 (or more likely, X >> GPLv2).

Further empirical evidence -- as if it were needed -- for the latter is that Novell's deal never got challenged in court during all those years, even though the SFLC certainly would have the funds in place to do so.

Let Y be on said continuum the point at which Red Hat can, according to Eben Moglen, do the FireStar type of deal, where Red Hat paid a royalty:

Y > GPLv3_final > GPLv3_early > GPLv2

Thus, Y > GPLv2 (or more likely, Y >> GPLv2).

Not OK

Ze — Sun, 24 Oct 2010 07:17:22 +0000

I would prefer not to have to repeat my remarks on honesty.

Is that because you seem to have a problem with honesty yourself? You've made dishonest claims about GPLv3 and GPLv2.

You claim that GPLv3 is irrelevant because no major projects have adopted it , yet you've been shown evidence to the contrary that the tiniest bit of research on your part would've found.

You've also claimed that GPLv2 is patent compatible using extremely poor logical reasoning. It's spurious logic to claim that just because the text of early drafts of GPLv3 doesn't protect against software patents that GPLv2 doesn't protect against them , when the concerning them between the two had most probably changed. Someone who is so interested in FOSS and software patents surely can't claim that such a basic flaw in reasoning is an honest mistake....

Could the reason you don't want your claims about honesty repeated is because you have a problem with honesty yourself?

Open source must be able to deal with patent licenses if necessary

FlorianMueller — Sat, 23 Oct 2010 23:24:17 +0000

Of course the FSF has copyrights assigned that it could use here if it wanted. And the assumption that FSF couldn't afford it is also wrong. What is the SFLC for? It has an annual budget of several million dollars.

Open source must be able to deal with patent licenses if necessary

anselm — Sat, 23 Oct 2010 23:02:50 +0000

To sue somebody about a GPL violation, one would have to have appropriate standing, i.e., hold copyright in the software package in question. For example, Harald Welte of gpl-violations.org fame gets to go after companies which don't publish source for their Linux-based routers because he has contributed code the Linux kernel, so he is a copyright holder and is entitled to sue.

On the other hand, the FSFE in particular can't really make a big show of »enforcing the GPLv2« even if they wanted to, because that would require them to have released software under the GPL (and software that would be interesting enough for ruthless companies to rip off, at that). However, unlike the original FSF, the FSFE does not appear to actually produce any free software at all – it is mostly a lobbying outfit –, so there isn't anything to enforce.

Open source must be able to deal with patent licenses if necessary

Arker — Sat, 23 Oct 2010 21:35:03 +0000

However, after four years it's fair to say that they don't have a case to enforce the GPL (otherwise they would have done something).

No, it's not fair to assume this at all, I just explained to you why not, and your reply is simply to reässert your ill-founded assumption.

After four years during which ever more such deals have been signed, it's clear that it's GPLv2-compliant behavior.

No, no, no. All that can be inferred from this (and I am generously assuming you are 100% correct on the facts) is that no one has both standing and sufficient resources and motivation to sue over any of these deals yet. Period. Anyone with the slightest familiarity with civil law would understand that.

I didn't say that by tolerating those deals they change the legal status of those deals.

You say that the lack of action proves they are legal, the difference is splitting hairs. Both are wrong, and dangerously so.

Before you bring a suit you need not only a solid legal case as the wrongness of the defendent, you also need standing (FSF cant sue over linux, no, they dont keep copyright on linux code as you so bizarrely asserted a little earlier) and you need a theory of damages that could generate an award sizeable enough to at least pay the lawyers, or some other compelling business case for the expense. This is the reason you havent seen anyone suing Novell - not because what they are doing isnt clearly, demonstrably in violation of the license, but only because those with standing to object have no interest in filing the case.

Open source must be able to deal with patent licenses if necessary

FlorianMueller — Sat, 23 Oct 2010 18:38:16 +0000

The comment to which you replied actually said that they wouldn't have to go to court if it can be avoided. However, after four years it's fair to say that they don't have a case to enforce the GPL (otherwise they would have done something). After four years during which ever more such deals have been signed, it's clear that it's GPLv2-compliant behavior.

I didn't say that by tolerating those deals they change the legal status of those deals. What certainly happens is that companies feel good about doing ever more of those deals.

If they (in this case, FSF's European affiliate) tell politicians that such deals can't be reconciled with the GPL, it's striking that they don't enforce the GPLv2 where they could if they were right.

Open source must be able to deal with patent licenses if necessary

Arker — Sat, 23 Oct 2010 18:33:23 +0000

This is fundamentally illogical Florian. Courts are a last resort at best. Before you go to court over something like this you have to analyse whether it is a worthwhile use of scarce resources. That's a tactical decision, and has little to do with the legalities. It's not like trademark where you have a legal obligation to attack or lose either. A million cases can go by where there are technical violations of the GPL that no one finds worthwhile to sue over, a million more where they sue and settle out of court, not one case actually resulting in a judgement, without it having any direct bearing on case #2,000,001.

The question is more complex than a selected soundbyte from FSFE makes it sound? Sure. But when you turn around and claim the exact opposite, you arel being far more inaccurate than they were.

This once I actually agree with Grokl.. :-)

FlorianMueller — Fri, 22 Oct 2010 17:32:54 +0000

Do you think a statement like this is in any way a helpful contribution to a discussion? If you believe something is wrong, or if you disagree, you can suggest a correction or present a different view. If you do it clearly enough so that one doesn't have to guess what you actually mean, then that's useful. I had to ask question after question to find out what you actually meant. I hope in your profession as a notary public you make it easier for people to understand you.

This once I actually agree with Grokl.. :-)

jthill — Fri, 22 Oct 2010 17:29:31 +0000

Mr. Mueller, I believe you've now spent more words misrepresenting the FSFE's view and "imprecise"ly dwelling on that falsified view than they spent stating their actual view.

I suppose some sort of congratulations are in order.

This once I actually agree with Grokl.. :-)

FlorianMueller — Fri, 22 Oct 2010 02:18:59 +0000

I admit I was imprecise by referring to just the passage he quoted and that one term. There are other references in the published part of that license agreement that are not boilerplate and do make it clear that a payment had to be made.

This once I actually agree with Grokl.. :-)

FlorianMueller — Fri, 22 Oct 2010 02:15:33 +0000

"Paid-up" in that context just refers to the fact that downstream users don't have to pay. Nevertheless Red Hat committed to pay.

This once I actually agree with Grokl.. :-)

jthill — Thu, 21 Oct 2010 20:06:28 +0000

a passage that says "paid-up". So a royalty payment was made.

Search for you grant "paid up". Search for java license "paid up". Search for mysql license "paid up". Search for fedora license "paid up". $0.00 is a common dollar amount on "paid up" notarized documents, says me, a notary public who has notarized well over ten thousand documents and is required to verify that they contain no material blanks, for instance in fields requiring dollar amounts. It's bog-standard boilerplate meaning no future payment is due. It says and implies exactly nothing about any past payments.

This once I actually agree with Grokl.. :-)

Wol — Thu, 21 Oct 2010 19:41:33 +0000

Do you not understand the term "royalty-bearing"? Or are you being deliberately dense?

"Royalty bearing" and "paid up" are MUTUALLY EXCLUSIVE.

Re-read the GP in the light of that statement!

Cheers,
Wol

Open source must be able to deal with patent licenses if necessary

Wol — Thu, 21 Oct 2010 19:33:22 +0000

Precisely.

And the reason for this is that classical physics is couched in maths. It is self-consistent, built on a bunch of axioms, and can be shown to be mathematically correct.

Unfortunately one of the KEY assumptions (if not THE key assumption) is the axiom "parallel lines never meet". Which has been empirically proven false - Einstein's 1919 solar eclipse experiment was the event.

Cheers,
Wol

This once I actually agree with Grokl.. :-)

FlorianMueller — Thu, 21 Oct 2010 14:52:29 +0000

Don't be paranoid about being trolled. But it would help if you could make it clearer what your concern is. You quoted a passage that says "paid-up". So a royalty payment was made. Why would this not be possible on FRAND terms? You can make a payment that's FRAND and then give customers this degree of legal certainty. All you need for that is FOSS-compatible FRAND.

This once I actually agree with Grokl.. :-)

pboddie — Thu, 21 Oct 2010 14:49:51 +0000

When practically everyone in the community concerned gets a "perpetual, fully paid-up, royalty-free, irrevocable worldwide license to the patents", that's a perverse definition of royalty-bearing FRAND licensing indeed.

Again, I get the feeling I'm being trolled.

This once I actually agree with Grokl.. :-)

FlorianMueller — Thu, 21 Oct 2010 12:40:24 +0000

I never supported the BSA position either. FSFE now claims I'm "echoing" Microsoft/BSA positions, which isn't true. What I suggest is for FOSS to meet FRAND and FRAND to meet FOSS in a way that works for both. That's as much in the middle of the two positions as it gets.

Concerning "royalty bearing regimes", there are various ways to define royalties and some of them do work with FOSS. One of them (or a combination of several of them) was apparently agreed by Red Hat in the FireStar deal and very likely also in the recent Acacia deal, and possibly also in other deals they never even announced.

If Red Hat wanted to be constructive and truly open, they could present (without providing confidential detail on the amounts they paid) the kinds of terms they've already accepted.

This once I actually agree with Grokl.. :-)

pboddie — Thu, 21 Oct 2010 12:35:48 +0000

BTW, I never claimed that all (F)RAND works for FOSS. But some does, and it's not that there's no way to pay royalties on GPL'd software.

Huh? Red Hat effectively licensed various patents for all downstream recipients, effectively removing any "FRAND" element from the licensing for those people.

And looking again at what the FSFE says to refute the "entirely compatible" nonsense claims of the BSA - it's quite possible to demonstrate a FRAND concept which won't work with Free Software licences - they're referring to a "patent royalty bearing regime" (which explicitly conflicts with various licences).

Really, despite maintaining a degree of patience to evaluate the claims of dishonesty from numerous angles, I'm starting to think I've just been trolled all along.

Not OK

armijn — Wed, 20 Oct 2010 15:49:38 +0000

Karsten Gerloff (FSFE president) has written down how FSFE's open standards work is actually down. It's definitely worth a read: http://blogs.fsfe.org/gerloff/?p=408

This once I actually agree with Grokl.. :-)

FlorianMueller — Wed, 20 Oct 2010 14:13:29 +0000

I want to point out that I didn't provide, let alone clarify, a definition of an "open standard".

This once I actually agree with Grokl.. :-)

tzafrir — Wed, 20 Oct 2010 14:11:01 +0000

I must say that this is an odd definition of an Open Standard. I prefer the one of the w3c. But I'm glad you clarified your definition.

This once I actually agree with Grokl.. :-)

FlorianMueller — Wed, 20 Oct 2010 13:56:58 +0000

The second question about whether we "want" this is, for the umpteenth time, separate from whether it's intellectually honest to claim that patent royalties don't work for FOSS.

In terms of "reasonable" for small companies, they will always have some structural disadvantages compared to large companies in a variety of contexts. Even if you solved this for IP, it would still be a problem in other fields. So it's not an argument with which you can convince politicians to ask IP holders to waive the entirety of their rights just to solve a part of your structural problem.

This once I actually agree with Grokl.. :-)

tzafrir — Wed, 20 Oct 2010 13:43:12 +0000

Is it Reasonable to require the small company I work for to negotiate a separate deal with the patent holder and pay a huge pile of money to get such a broad license?

And do we really want open standards to have such t(r)oll posts?

This once I actually agree with Grokl.. :-)

FlorianMueller — Wed, 20 Oct 2010 12:58:10 +0000

Even with GPLv3. Yes, 3. Eben Moglen, as Groklaw also reported, declared Red Hat's deal with Firestar GPLv3-compliant. BTW, I never claimed that all (F)RAND works for FOSS. But some does, and it's not that there's no way to pay royalties on GPL'd software.

This once I actually agree with Grokl.. :-)

tzafrir — Wed, 20 Oct 2010 12:54:38 +0000

Is that an example of how RAND patent licensing can just live happily ever after with GPLv2?

(F)RAND

rsidd — Wed, 20 Oct 2010 11:51:33 +0000

Please make the suggestion to the right people! Who knows, it may pass through.

Not OK

pboddie — Wed, 20 Oct 2010 11:28:25 +0000

I would prefer not to have to repeat my remarks on honesty.

So would we all.

The "advice to the contrary" you link to is weak tea compared to the fact that it's common practice now in the industry and not legally challenged by the free software movement.

Well, I'm just explaining how the FSFE people concerned can make their particular claims and not go round winking to each other about how they were really "lying" afterwards. And you go on again about how no-one has sued anyone, but again that doesn't have anything to say about whether someone could do so once everyone has "agreed" that there's no underlying problem.

You know, quite a few people had a discussion on this very site about patent-encumbered standards a few months ago. Look it up: it was all about MPEG, Theora, Mozilla and HTML5. There were three principal positions:

"Patents in standards should be avoided because it ends up mandating a form of private taxation on implementers."
"Patents in standards don't matter because even though our code is affected, we don't care and no-one is going to come after us anyway."
"Patents in standards are OK because it's just a matter of paying a fee and then we're covered."

And yes, the discussion involved persuading the people taking the second position that although their code is nice, that perhaps they should consider what the people taking the third position were doing with their code. And no, sadly, those people representing the developers weren't particularly interested.