HP has made a statement in support of royalty-free web standards and urging the community to comment to W3C.
HP has made a statement in support of royalty-free web standards and urging the community to comment to W3C.
Bruce Perens, firstname.lastname@example.org
W3C, the World Wide Web Consortium, recently released a Patent Policy Framework draft which provides for two forms of patent licensing that may be mandated by web standards: either RF for Royalty-Free, and RAND which stands for Reasonable and Non-Discriminatory patent licensing. When the proposal belatedly came to the attention of the web community, there was a storm of protest against RAND patent licensing and in favor of RF. There are many reasons to dislike RAND, but the one I focus on as HP's Linux and Open Source Strategist is the fact that a required patent royalty is incompatible with Open Source software. Open Source is by definition royalty-free. Thus, "non-discriminatory" patent licensing actually does discriminate against Open Source.
Over the past several years, all successful new software standards have had one thing in common: an Open Source implementation. In order to maintain the many benefits to the public provided by Open Source software, we must prevent it from being marginalized by interoperability "standards" in which Open Source may not participate. Thus, we must insist that our interoperability standards be free of any legal encumbrance that would prohibit an Open Source implementation.
HP's policy regarding RAND may have been mis-interpreted by the public and the press, because the name of an HP attorney appears on the Patent Policy Framework draft. However, that attorney was not a major contributor to the draft, and he asserted to the committee upon HP's behalf that royalty-encumbered standards would not be successful. While HP has probably been the leading voice in this, they have not been alone. I am told that Sun, and surprisingly, Microsoft, have been supportive on the issue of avoiding patent encumbrances. However, one friend of Open Source may still need a nudge in the direction of supporting royalty-free standards.
So that HP's policy is clear to all, the following is an announcement from Jim Bell, HP's representative to the W3C advisory board:
HP Supports Royalty Free Standards for Web InfrastructureThanks, Jim. It's interesting to note that Jim's previous position at HP was director of our Open Source and Linux Operation, which paved the way for HP's entry into the Linux business.
With the extension until October 11 of the deadline for comments on the W3C's draft patent policy, now is the time to speak up if you have not already done so.
HP has been a leader in the advocacy of royalty free infrastructure standards for the Web. At the W3C Patent Policy Working Group, HP has been the most vigorous proponent of the importance of avoiding patent encumbrances on W3C Recommendations. As HP's representative to the W3C Advisory Committee, I have personally written to every W3C Member Company's representative on this topic. We have backed up our words with actions. For example, HP resigned as a co-submitter of the otherwise excellent Web Services Description Language (WSDL) proposal to W3C solely because other authors refused to let that proposal be royalty free.
As noted in Tim Berners-Lee's message initiating the current W3C discussion of royalties:
We and the W3C Team have assumed that Web Services technologies, such as a description language, would be considered common infrastructure and a basis for much exciting future work. We therefore believe that the draft Patent Policy, which holds that it is especially important that the Recommendations covering lower-layer infrastructure be implementable on an RF basis, is applicable. In this belief we are encouraged by certain members such as Canon, HP and Oracle. On the other hand, the other proposal drawn up by other members involves the possibilities of royalties being payable on RAND terms for Web Services technology.
There are sharp differences among companies on patent licensing for standards. For example, IBM's response to the current W3C draft policy states, The policy of licensing patents under RAND terms and conditions has allowed our best technical individuals to work together without becoming burdened by patent issues. HP feels that the prospect of subsequent discussions of potentially royalty bearing licenses is itself an unacceptable burden on standards discussions.
I would encourage you to let your voice be heard by submitting your comments on W3C's Patent Policy Framework Working Draft at http://www.w3.org/2001/10/patent-response .
Hewlett-Packard's position is that the fundamental standards for the Web should be royalty free.
W3C Advisory Committee Representative
Director of Standards and Industry Initiatives
Agreement on royalty-free standards does not end this discussion. The licensing of patents embedded in standards must be compatible with the GPL license that is applied to the Linux operating system kernel, the MIT-derived license that is applied to the Apache web server, and a number of other software licenses. Because of the many thousands of copyright holders who have already contributed to existing products under those licenses, those software licenses can not be changed - the patent licensing mandated by W3C standards must accommodate them.
I join Jim in urging all of you to make your opinion be known by submitting your comments on W3C's Patent Policy Framework Working Draft at http://www.w3.org/2001/10/patent-response .
Senior Strategist, Linux and Open Source
Nota Bene: For the purposes of this discussion, Free Software is a logical subset of Open Source.
The master copy of this document is at http://perens.com/Articles/HP_And_W3C_Standards.html . You may copy that version of the document and republish it. You may translate the document to another language or reformat it to fit your presentation. Do not change the information in the document in a way that would misrepresent our opinions or policies.
The discussion is not focused on the important point of the situation : the software patent.
Patent shouldn't be allowed for Software and Mathematics. (cf. Bastia)
Why the current battle is not focused on software patent ?
The patent blocks current VRRP implementation (vrrpd exists but without any idea of which troubles they can get from cisco...) for OpenBSD in kernel (user-space implementation exists).
So, the issue has to be more focused on patent for software, that's the real problem...
adulau: Arguments from the axiom "Software patents are immoral" are unlikely to be taken seriously, since software patents are, for better or worse, legal and there isn't a common W3C position on what is ethical. Replies that say that royalties are inherently discriminatory, since open source software cannot comply with them, are likely to be taken seriously. That is why people don't disguise their arguments as rants against software patents.
I've been kicking this idea around for about 4 years and even pitched it to some IPLs within IBM. This may be a religious issue but I have no problem with someone patenting something and saying "I invented this technology," and in many cases I think it's even a good thing.
What drives me nuts is the agreements, licenses, royalties, etc.. that go with it. I can't just walk up to IBM and ask to use arithmatic encoding, I'm going to have to poney up some cash, agree to pay a royalty, agree to give them any IP I may have or develop or similar. The point is washed with free software because there isn't going to be any IP, or royalties and I'm not going to have the money to pay up front.
So what I was thinking was of a license to protect IP with OSS. Say you allow free use of patents and redistribution of implementation of them so long as the implementation is GPLed. It may not make a lot of sense for some companies but I think others could benefit from it and then the open source world would get to also.
For example, suppose MS buys the rights to patents something that makes active-x faster. They probably fork out a few million dollars implement it and sell windows with this better technology for components. The folks building KDE and GNOME cannot use that so unless they develop something better their component model won't be as good. This is a poor example but it gives MS an unmatched advantage. Now since GNOME and KDE do give their code away for free then maybe their could be an exception. Suppose they are allowed to use the same technology then they are going to be at least as good as the non-free competition in that respect (or have the ability to be that good) that forces MS to have to be better in other ways and it also makes the license they paid that much more valuble to holder because the use of that technology is part of the "free entry level to competition benchmark" suppose HP develops an OS that has something similar, then they would have to pay for it to be as good as the free guys or there is always the alternative; they GPL their code.
Sorry this is getting long and rambling, but if the W3C allowed uses of licenses like this I would think that it would be kind of a win-win
situation. IP holders could potentially make more money for their IP and the public at large would be protected because there would be free implementations available or the possibility of making them provided that they stay free. What do you guys think? It's kind of the same idea behind releasing the PC as a set of open standards.
Nelson: I like the idea very much. I think by itself an exclusion for GPL'd code would make almost any other provisions acceptable to the OS community. Some thoughts:
- LGPL would be even less discriminatory than GPL, since for some OS projects it is important to be able to link against propietary code (think of all the GPL exceptions in the GNU CLASSPATH project).
- The scope of patents should be clear in any case: ie. no "moving targets", no incomprehensible documentation of the implementation, etc. I wish this kind of thing is a given, but it isn't.
- We need to be clear about worst case scenarios: what if the W3C decide RAND is good enough, even though open source projects can't comply. Can we at least ensure that in W3C proposals that they are discriminatory against OS, and are not suitable for core W3C infrastructure?
One thing to be clear about: there's been no RAND working groups created yet, and when there is, you'll know about it. It'll be in the charter for the working group. That's the bright side of this proposal. Under the old proposal, a specification can get all of the way to Recommendation without a formal review of the encumberances on the specification. The longer this policy gets put off, the longer the W3C will not have a standard policy of letting everyone know the IPR policy of a group.
I'm a member of the SYMM working group (SMIL 2.0), which was a beta tester of this policy. We ended up pushing for RF, and got it. The thing that was painful was making the decision at the end of the life-cycle of the group, because we didn't have a lot of latitude to change the technical details of the specification at the end of the process.
The W3C is a consensus-based organization, and as such, it's going to be very difficult to get a blanket policy that RAND is unacceptable in all cases. However, this policy is about as good as it gets as a tool for determining when a group is heading down the wrong road, and so, on a case-by-case basis, it'll be much easier to talk in concrete terms about why RAND is bad for a particular specification.
So what I was thinking was of a license to protect IP with OSS. Say you allow free use of patents and redistribution of implementation of them so long as the implementation is GPLed.
This idea, while an interesting one, would shaft Mozilla, which is a consistent implementor of W3C web standards on the client side. Mozilla is under the MPL, which is not the GPL. I don't know if you consider this fact significant :-)
gerv, robla: Both very good points. Gerv, the Mozilla project could have dynamically linked plugins written in GPL code, or libraries written in LGPL code, without compromising the license.
Robla, I think the important point is that the "non-discriminatory" should be dropped from RAND, since it is misleading. I think I shall argue this to the W3C process.
does anyone have an answer to satisfy the W3C wrt stupid companies that wish to engage in patent wars (i've got one hundred patents i can xxxx you over with: how many have _you_ got. if it's greater than one hundred, i'll back off, because the legal bills are too high - i.e. into 5 figures)
the W3C basically doesn't want to waste their time developing technologies that some xxxxer patents and xxxxs them over with.
whilst i hate patent abuse - especially software patents - and do not think it is in anyone's interests to use patents (and agree with bruce wholeheartedly), the issue of keeping the wolves at bay has not been addressed.
the W3C seems to think that they can protect themselves from patent wars by becoming involved and patenting technologies they develop.
the issue is to convince them that this is _not_ in their interests, and also to give them an alternative defense strategy against xxxxers who wish to indulge in time-wasting, money-wasting, anti-competitive patent wars.
Calling your HTML/XML renderer a "dynamically-linked plugin" would be violating certainly the spirit, and almost certainly the letter, of the GPL. :-) GPLed code, unless a special exception is made, cannot be linked with MPLed code.
If the only exception is for GPLed software, Mozilla is in serious trouble. Luckily, I think that this is not a path they will go down. They will either do a UCITA-type thing, as esr, recommends (which is very general), or give up on RAND completely.
New HTML Parser: The long-awaited libxml2 based HTML parser code is live. It needs further work but already handles most markup better than the original parser.
Keep up with the latest Advogato features by reading the Advogato status blog.
If you're a C programmer with some spare time, take a look at the mod_virgule project page and help us with one of the tasks on the ToDo list!