Eric S. Raymond: Open Source Initiative's letter of comment on W3C's proposed RAND policy

Oct 09, 2001, 19:05 (23 Talkback[s])

Date: Tue, 9 Oct 2001 14:20:13 -0400
From: Eric S. Raymond
Subject: OSI letter of comment on W3C's proposed RAND policy

The Open Source Initiative has considered the World Wide Web Consortium's Patent Policy Framework working draft at

http://www.w3.org/TR/2001/WD-patent-policy-20010816/

and the W3C's Response to Public Comments at

http://www.w3.org/2001/10/patent-response

The recent public disclosure of these plans to endorse patented Web standards has ignited a major public controversy. The Open Source Initiative's charter requires us to speak out for the interests of open- source developers, and this statement is intended to be our contribution to the debate.

Recommendation for Change in the RAND Proposal
We believe that there is merit to many portions of this proposal. In particular, we unhesitatingly endorse the patent disclosure requirements in section 7. One does not need to settle the general question of software patents to affirm that all developers (and not merely open-source developers) are entitled to protection from unethical "submarine patents".

We further endorse in a general sense the existence of well-defined tests for determining whether a proposed W3C standard meets the W3C's transparency requirements (sections 4 and 5). Clarity on these questions will better serve web developers and the public, and the W3C is to be applauded for attempting to document and regularize its procedures.

However, we must reject the specifics of the proposal for a RAND ("Reasonable and Non-Discriminatory) Licensing mode. Clauses 1, 2, 3, and 6 are unexceptionable; clause 4 is fair; but clause 5 "may be conditioned on payment of reasonable, non-discriminatory royalties or fees" is unacceptable as written.

While opinions may differ on whether the so-called "RAND" policy with clause 5 is "reasonable", there can be no question but that it is discriminatory. Clause 5 would discriminate against open-source developers, who operate in small volunteer project groups and have neither the wherewithal to pay license fees nor the shelter of a fictive legal entity with which a license might be negotiated.

OSI suggests that the proposed clause 5 can be rescued by adding a suitable interpretation of "non-discriminatory". The language in the State of Maryland UCITA exempting open-source projects from warranty requirements provides a suitable model:

The payment of royalties under a RAND license shall be waived for any licensor of a computer program that is provided under a license that does not impose a license fee for the right to the source code, to make copies, to modify, and to distribute the computer program.

This interpretation of "non-discriminatory" would leave patent-holders the option of continuing to collect license fees from implementors with plans to charge for the secrecy of their software.

Principles and the Larger Context
The above is a narrow technical prescription, but there are larger issues here. The OSI would be derelict in its duty if we failed to address those. We stand with those who see the RAND proposal as an invasive and corrupting attempt to hijack the standards process, to tilt the Web's relatively level playing field in favor of players with legions of lawyers and lots of money. That other standards organizations who should have known better (such as the IETF) have let this particular camel's nose under the tent is no excuse for the W3C to repeat that error.

RAND discriminates not merely against open-source developers, but for entrenched monopolies and against the consumer and small busineses -- who under this proposal would be systematically deprived of open-source alternatives and have fewer choices (all of them expensive and proprietary).

RAND would also tend to stifle innovation. The history of the Web and the Internet amply demonstrates that in today's world innovation does not come to us from the entrenched giants of technology and media -- indeed, they experience genuine innovation as an unwelcome disruption, upsetting their strategic plans and obsolescing their cash cows. RAND would permit -- nay, it would invite -- abuse of supposedly "non-discriminatory" license terms to suppress innovation and lock in customers.

Web innovation springs from precisely those independents and open-source developers whom RAND would systematically freeze out of the process. If the W3C enacts Clause 5 as written, it will be an act of betrayal against both the Web's history and its future. The W3C will deserve the schism and developer revolt that would follow -- and it will deserve the fate to which this bad decision would inexorably doom it.

In the past, "standards" that are actually manacles could be foisted on a relatively powerless developer and user community -- because both computing technology and the software development that went with it were expensive and centralized, requiring financial mass and corporate management structures. In that world, there was no alternative but fetter, and customers could only try to choose the least weighty.

The OSI exists today because those days are passing. The combination of the Internet, personal-computer technology, and distributed open-source development has empowered millions of users and developers, and indeed proven that it can produce dramatically better results than the old system. Software developers and users are increasively refusing to accept the kind of suppressive shenanigans that corporate-dominated standards bodies have used in the past as market-control tactics.

The time is ripe for confrontation. Thousands of web developers have already made their voices heard against RAND. The largest IT consultancy in the world has just recommended that IT customers drop proprietary webservers in favor of the open-source Apache. The increase in adoption of open-source operating systems on web servers continues exploding, propelled both by pressure on capital spending and the increasingly onerous license terms attached to their closed-source and technically inferior competitors.

If the W3C persists in its present course, it risks having its tea dumped in Boston harbor as the first move in a revolution that will vest effective control of Web standards in open-source groups like the Apache Software Foundation and entirely out of the ambit of the W3C and its sponsors. OSI would do what we can help lead that revolt.

Open Standards Require Open Source
But W3C does not have to be on the losing side. We urge the W3C to take this opportunity not only to affirm its royalty-free-only license policy, but to institute a requirement that no proposal may become a W3C standard unless it is backed by an open-source reference implementation on an open-source platform, with patent grants sufficient to ensure that the reference implementation remains unencumbered.

That requirement would truly create a level playing field for all parties -- users, developers, content providers, and all of the Web's millions of stakeholders. It would also serve as a reliable reality check on the tendency of paper standards to effloresce into unimplementable monstrosities.

Your organization is supposed to be committed to a World Wide Web that is open to all. The OSI urges W3C to remember this, and to center its standards policy around one simple truth: open standards require open source.

                        Issued by and for the Board of Directors of OSI
                               by Eric S. Raymond, President
                                   9 October 2001

Related Stories:

• CMP: Discussion: W3C and Patents (Oct 09, 2001)
• CNET News.com: Will W3C mean dollar signs? (Oct 06, 2001)
• MoonGroup: The W3C Fork: Coming soon to a net near you!(Oct 05, 2001)
• The Register: IBM risks billion dollar Linux strategy with W3C RAND demands (Oct 05, 2001)
• HP to W3C: We Support Royalty-Free Standards for the Web Infrastructure(Oct 04, 2001)
• The Register: We'll fork the Web to keep it Free - Perens (Oct 04, 2001)
• Debian Weekly News - October 3rd, 2001(Oct 04, 2001)
• CNET News: Gartner Commentary: Standing in way of standards(Oct 03, 2001)
• The Register: Web standards schism "terrible" - W3C patent policy boss (Oct 03, 2001)
• W3C's Response to Public Comments on the Patent Policy(Oct 02, 2001)
• CNET News: W3C Patent Plan Draws Protests(Oct 02, 2001)
• Followups on the W3C Patent Policy from Richard Stallman, Eben Moglen, and Microsoft(Oct 01, 2001)
• Alan Cox: W3C Patent Policy: Bad for the W3C, bad for business, bad for users(Sep 30, 2001)
• Community: W3C and the Promotion of Fee-based Standards for the Web (Sep 30, 2001)