Patents and Web Standards Town Hall Meeting

December 19, 2001

Michael Champion

On 16 August 2001, the W3C’s Patent Policy Working Group released a working draft of a new framework governing the potential use of patented technology in W3C Recommendations. This draft described a process that would mandate that working groups specify in their charter whether they would operate in "RF" or "RAND" mode with respect to patented technology. The "royalty-free" (RF) mode working groups would refuse to consider any technology encumbered by patents that would not be freely licensed to all. RAND mode groups would consider technologies whose owners agreed to license them on "Reasonable And Non-Discriminatory" royalty terms to all. The proposed policy drew almost no response until two days before the deadline for comments from the public. On 28 September, Adam Warner wrote a piece for Linux Today describing the proposed policy and warning that there was little time to get in comments. This was picked up by Slashdot and triggered an unprecedented response on the W3C’s public comment mailing list, almost all of it vehemently opposed to the working draft.

The XML 2001 Conference and Exposition in Orlando, Florida sponsored a Town Hall meeting on 11 December that was billed as providing "an opportunity to listen to speakers describe the issues at stake in this controversy and to participate in debate over the best way for the Web community to respond." Despite vigorous efforts by conference chair Lauren Wood, the only person brave enough to face the public on this contentious issue was Daniel Weitzner, technology and society domain leader at the W3C and chair of the Patent Policy Working Group. Weitzner gave a 20-minute summary of the issues, the response from the public, and the reaction of the W3C Advisory Committee, then took comments and suggestions from the audience.

Weitzner began by summarizing the objectives of the W3C Patent Policy:

1. To clarify the requirements on W3C member companies to disclose information in a timely manner concerning patents that they hold that could affect the implementation of specifications under development by the W3C.

2. To revise the W3C process document to define the terms "royalty free" and "reasonable and non-discriminatory" with respect to patents on Web technology, and to ensure that a working group decided which mode it would operate under at the time it develops its charter.

3. To ensure that all W3C Recommendations that depend on patented technology can be implemented on at least RAND license terms.

4. To define a dispute resolution process for patent and licensing issues.

Companies holding significant patent portfolios were heavily represented on the original Patent Policy Working Group, and not surprisingly, the working draft released on 16 August described a policy that would allow (but my no means mandate) W3C working groups to consider technology that is only available on RAND terms. Daniel Weitzner outlined three specific reasons for this decision:

• Most other standards bodies can and do standardize technologies that are protected by patents that are offered on RAND licensing terms.

• Precluding the use of technology available only on RAND terms could deny the "best" (presumably patented) technology to the Web.

• Regardless of a W3C policy against technology available on RAND terms, various parties could still claim that implementations of a W3C recommendation infringe on their patents. It is arguably better to have a policy that deals with patent claims and licensing policies in an orderly manner than to sweep such issues under the rug until after a specification reaches Recommendation status.

After 28 September, it became quite clear that the overwhelming number of people responding on the patent policy comments list did not agree with the reasoning in favor of allowing W3C working groups to operate on a RAND basis if they so choose. Weitzner described three points made repeatedly in the responses that struck home with the Working Group:

• The development of the World Wide Web over the last 10 years would not have been possible, or at least would not have resembled the course it did take, if implementers of HTTP, HTML, URIs, and the XML specifications had to pay royalties to patent holders.

• The open source development model that has powered so much Web and XML development is simply incompatible with specifications that require royalty payments to implement. The open source community made very clear that the RAND policy would cause the Web to "fork," that is, split along royalty free/royalty required lines, to the great detriment of interoperability.

• The RAND policy would "tilt the playing field" in favor of the large companies with deep pockets and extensive portfolios of patents to cross license, thus stifling the innovation by small companies and even individuals that has so famously driven the evolution of the Web.

The draft patent policy and the public outcry against it apparently dominated the W3C Advisory Committee meeting held on 6–7 November in Nice, France. Details are W3C member-confidential (the whole point of the confidentiality policy is to prevent member companies from using unpopular positions taken within the W3C against their competitors!), but the "buzz" around the XML 2001 conference indicated that the discussions were quite heated.

As it turned out, Weitzner passed on news that was quite well received: The Patent Policy WG had been directed by the W3C Advisory Committee to rework the draft policy to have a strong preference for technology that is not encumbered by patents, or whose owners agree to make them freely available on a royalty-free basis. The Working Group has been expanded to include several invited experts from the open source and public software license community. The plan is to have a patent policy in place by May 2002.

This went over quite well with the audience. There were numerous questions and a vigorous, but polite and well-reasoned, discussion centering around the question of whether the W3C should issue any Recommendations that may require patent license payments in order to implement.

Some (heavily paraphrased) questions and answers that came up during the discussion included:

Q: How will the W3C prevent [various abuses of the patent system with respect to Web technology]?

A: The issue here is what the W3C imprimatur on a specification should mean, not how to prevent things that it cannot control.

Q: Many of the public statements against the draft patent policy made reference to notorious "bad" patents. Does the W3C envision supporting legislation or litigation to improve the patent system or oppose specific patents that are not in fact original inventions of the patent holder?

A: The Patent Policy Working Group concluded that the W3C cannot meaningfully affect the Patent and Trademark Office, the Congress, or the courts. The goal is to develop an effective strategy for working in what is clearly a compromised legal environment. The W3C did go to considerable expense to get legal advice on the claim by a member company (Intermind, but Weitzner was too diplomatic to mention it by name) that it held a patent on the technology necessary to implement the proposed W3C Platform for Privacy Preferences (P3P). The analysis concluded that P3P did not infringe on Intermind’s patent, but the W3C did not try to challenge the patent itself. Legal fees are certainly a deterrent to doing this kind of thing in the future, so the W3C prefers to fine-tune its process to clarify patent claims up-front rather than challenge them later.

Q: What does the Advisory Committee’s guidance favoring RF as a patent policy mean for the Voice Browser Working Group and the heavily patent-encumbered VoiceXML technology it is considering?

A: Voice technology is an area with many patents and little leeway for the Working Group to operate in RF mode. On the other hand, accessibility of the Web to people with physical limitations is a key part of the W3C’s mission; many consider VoiceXML a high-level application, but blind people may consider it a key infrastructure component. This creates a genuine dilemma for the W3C, and even if the patent policy heavily favors RF, it may allow this activity (and perhaps others in a similar situation) to operate in RAND mode.

Q: Explain the Patent Policy Working Group’s notion of "reciprocity."

A: The Working Group considers this to be one of its real achievements. The W3C patent policy will not allow "bilateral," reciprocal licensing of patents on Web technology; that is, one company can not license its patents on an RF basis to a company in return for an RF license back from the same company. Under the W3C policy, extending RF reciprocity to one company automatically extends it to the entire Web.

The discussion also produced some rather memorable observations by audience members, including these (again, heavily paraphrased) nuggets:

Simon St. Laurent (challenged by Weitzner to choose between the horns of the VoiceXML dilemma): I’d choose simple and royalty-free over powerful and RAND.

Tim Bray: There is less of a dilemma here than meets the eye. The W3C has developed such a powerful brand name that its imprimatur on a technology is extremely valuable. If the W3C takes a strong line favoring only RF patents, the patent owners will toe the line.

Lauren Wood: People from small companies who worked long hours to produce the XML standards don’t appreciate the idea of some big companies coming along and getting royalties from specifications that would be worthless without the XML infrastructure.

These comments seemed to characterize the general sentiment of the audience. A straw poll confirmed this, with only a handful of dissenters. In response to a call from Weitzner and the moderator for the minority opinion to be voiced, Michael Sperberg-McQueen offered: "If I lost my sight, I would rather pay a few pennies in royalties to Motorola and Nokia than endure an inferior standard for voice technology."

And so ended the W3C Patent Policy Town Hall Meeting. The participants appeared to leave happy that the W3C planned to heavily favor technologies that could be licensed on royalty-free terms, but thoughtful about the possibility that this would limit the access of Web users to the best technology in specific application areas. The overwhelming consensus was that Daniel Weitzner had done a superb job in explaining the challenges the W3C faces here.

The author served as chairman of the Patents Town Hall Meeting at XML 2001.