Slashdot Mirror
Act Now To Sidestep A W3C Patent Pitfall
Here is what the Free Software Foundation says on its front page about this bug:
"The W3C 'Royalty-Free' patent policy proposal does not protect the rights of the Free Software community to full participation in the implementation and extension of web standards. Please read more on this issue and send a comment to the W3C."
Part of the effort that moved the W3C to its present position was a furious outpouring of comments in opposition to the original proposal of the Englobulators:
- www.w3.org/2001/ppwg
- www.w3.org/2001/10/patent-response
- lists.w3.org/Archives/Public/www-patentpolicy-comment
- lists.w3.org/Archives/Public/www-patentpolicy-comment/2001Oct/thread.html
The fix needed right now is a small fix. But the W3C must again be reminded with what jealous vigor we guard our right to build our Web the way we have built it down to this day, using free software. The bug appears in Item 3 of Section 3, titled 'W3C Royalty-Free (RF) Licensing Requirements,' of the present proposal.
This Item allows for a supposedly free grant to use a patent to be so restricted that a piece of Web infrastructure software might be encumbered if used for some non-Web use. Since the GPL does not allow such encumbrancing, GPL-ed Web software re-purposed for non-Web use could not be legally freely redistributed. Please read the Free Software Foundation's page on this bug.
Here is the official Last Call for Comments.
If you write a comment in your own words, for repair of the bug, it will help. I shall write in, and I shall argue against adoption of the buggy sub-section. I shall also suggest an extension of the deadline for comments.
No, what the FSF wants is for the W3C standard to be defined in a way that GPL'ed software can comply with. If complying with the W3C standard requires the use of any restricted software technology, then there might not be compliant GPL browsers and servers, and everyone loses.
If the patent owners don't want to release their IP, then the patented technology shouldn't be part of the standard. Period.
In the Star Trek evil Mirror Universe, virtuoso cellist Yo-Yo Ma is gangsta hiphop star DJ Yo Ma-Ma.
So I'll paste a bit from the fsf site:
Here's a detailed step-by-step example that shows how this problem could play out:
Programmer P downloads the Konqueror web browser, receiving it under terms of GPL.
P learns of a new web standard that requires exercising a technique for parsing URLs that is patented by Corporation C. C has licensed the patent under an RF, non-exclusive license, but with a "field of use" restriction that says the license can be used to "implement the standard". The standard, as it turns out, covers only what browsers must do with URLs, and says nothing about the server side or clients that aren't user browsers.
P implements this technique in Konqueror, and seeks to redistribute the modified version on his website so that other users can benefit from Konqueror now complying with the standard. If he does, he is bound by the GPL under copyright law, because he is redistributing a modified version.
However, he knows full well of a condition on that code that contradicts the GPL (violating Section 7) -- namely, he knows that C's patent license prohibits folks from taking his URL parsing code and putting it into, say, a search engine. Therefore, under GPL Section 7, he is prohibited from redistribution.
You might think that he can simply assign his copyright to the existing copyright holder of Konqueror let distribution happen from that source. They could distribute under GPL, but they would be granting a self-contradicting license. Nothing (to my knowledge, but IANAL) prohibits someone from distributing copyrighted works under licenses that make no sense and are self-contradictory. However, it is certainly true that those who receive distribution of the works are stuck and can't undertake further distribution or modification themselves.
The way I understand it is that it can create a weird situation. Suppose I write code to parse Yahoo pages to extract information from them. Suppose that Yahoo uses some technology licensed in this way on their site. Now they can stop me because it's not a web browser.
This Item allows for a supposedly free grant to use a patent to be so restricted that a piece of Web infrastructure software might be encumbered if used for some non-Web use.
The issue for me as an architect (I have written IETF, W3C and OASIS standards) is that I don't necessarily own all the IP that I need to address a problem. If I need to get an IP owner to donate their property for the common good my job is easiest if I have to ask for as little as is necessary for the particular purpose.
Equally when I hold the IP I see little point in giving away more than I need to for the purpose of the Web specification even though my company does not regard patent licensing as a revenue stream. The point is that I might need some IP held by a party that deals in the non-Internet world. I will have an easier time negotiating a license for Internet use if I have some bargaining chips.
The FSF reading on the situation is 100% about their ideology and has nothing to do with real needs as far as I am concerned. Open Source software is not imune to the patent system. If you modify any open source software sufficiently you will run into a patent infringement.
If you apply the FSF "logic" you would have to stop distributing gzip because someone could modify it so that it infringed the Unisys LZ patents.
The W3C does not have a monopoly on standards making and in fact is already seeing a lot of the standards work migrating from to OASIS. Dealing with the cumbersome W3C process and formatting conventions is bad enough without additional IP roadblocks being errected.
At the end of the day the IP policy is utterly irrelevant since nobody is going to use the specification unless the IP terms are acceptable and they are going to determine what is acceptable, not Tim and certainly not RMS. To date that has generally meant 'free as in beer' however there are many applications where that is simply not achievable, if you want to do voice browsing you will run into IP issues and your choice will be do something encumbered or don't do it at all.
Don't do it at all may be the FSF answer, but he does not pay $57,500 a year for W3C membership dues. The point that Tim has missed is that the W3C membership is already annoyed on the value for money front, W3C is way more expensive to join than OASIS where we pay $10K. We are also far from happy on the bogus process front, it took me almost a year to get a W3C working group started. I am not happy with a set of document publication rules that are 'standards based' but turn out to mean that you can only edit standards with one editor.
The last thing W3C needs to do at this stage is to reopen the IP issue with the membership.
Looking for an Information Security student project suggestion?
Try http://dotcrimeManifesto.com/
That is not quite right, we never relased any code under the GPL. We released the libwww code as public domain, it is not GPL, it is not BSD, it is public domain.
When GPL was being discussed it was in the context of 'make it free', the GPL was rejected because it did not make the cern code free in the ways we wanted it to be free. We explicitly wanted browsers to be included with computers as a cost free part of the basic operating system. Remember that at the time (91) Mosaic had not even appeared, let alone Netscape. The point is that Tim never wanted the viral aspect and dropped the GNU angle as soon as it was explained to him.
In the end the public domain route was in large part dictated by political expediency. Explaining GPL or BSD to cern management would have taken a lot more time and led to more opportunity for confusion. Putting the code in the public domain was something they could understand - it had already been done with much of the CERN libraries.
The mistake that was made was public domain rather than BSD. If we had gone BSD then Mosaic would have been required to state that it used CERN code (60% of the Mosaic code was code from CERN used without attribution). That in turn would have meant that IE would have a credit. As it was the mainstream media did not recognize Tim as the true father of the Web until about 1996, and then only as a result of a major PR campaign led by MIT.
I would certainly advise researchers to use the BSD license in their code. I would strongly advise against the GPL if you want your ideas to be taken up by industry.
Looking for an Information Security student project suggestion?
Try http://dotcrimeManifesto.com/
Nor does it make it impossible for GPL software to make use of the patents. If you want to use a patent in GPL software, put that routine under the MIT license, which does not have the GPL's language regarding patents, and can link with GPL work. This work-around was suggested by Eben Moglen.
This doesn't mean I support software patents. I think they should be eliminated. But we can't eliminate software patents through W3C - only through legislatures.
I'll end this with a plea to Jay Sulzberger. Jay, you are working to destroy two years of work by myself, Eben Moglen, and Larry Rosen. You didn't participate in the patent policy working group. I didn't see you volunteer. You don't sufficiently understand the issues yet. Please help us get the current W3C policy accepted, so that things will get better instead of worse.
Bruce
Bruce Perens.
Bruce
Bruce Perens.
Regarding W3C, if you think they are insensitive, try IETF's attitude on intellectual property. It mostly comes from these working groups consisting of employees of too many big companies. They are all cross-licensed with another, and don't give a hoot if their standards can't be implemented by any merely medium-sized enterprise.
But we fixed that at W3C. We got a compromise. The community isn't used to compromise.
Bruce
Bruce Perens.