Trip Report W3C Activity Committee Representatives meeting, Fall 2002 by Lloyd Rutledge The agenda, with links to all slides, is at http://www.w3.org/2002/11/AC-agenda.html This meeting was pleasant and well-run, as usual. The W3C staff does a good job of communicating, maintaining and setting an example for good behavior in how to discuss technical matters in a group and build consensus. While every AC Rep meeting has touchy issues, such as last May's fee hike, this meeting's round of touchy issues did not ruin this general W3C gezelligheid. This meeting's touchy issues, discussed in more detail below, are the new patent policy and the use of XLink in HTML. Philipp Hoschka, in the corridor Philipp and I discussed panels at WWW2003, Budapest. He was involved in proposing two panels with multimodality and universal access as topics. I am on the panel committee, and had been communicating with Philipp on these proposals. Philipp will be at WWW2003, but not at ACM Multimedia 2002. Management Report, by Steve Bratt This past year has had fewer new recommendations than in years past. This is due to a wave of recommendations that started several years into W3C at a good point of W3C maturity for recommendations to start, and which have all finished at the same time over the past couple of years. ERCIM is replacing INRIA as the European contact. This is not "loosing INRIA" but "gaining many new European hosts". CWI was mentioned in this slide along with the other INRIA partners. Tighter financial management is more of a priority now, as with many organizations. W3C lost some members, some perhaps due to last year's fee raise, some perhaps due just to hard times, and some certainly due to mergers, in which two members become one member. Several measures are being followed which have kept FY2003 in balance. There will be, effectively, voting on what activities stay active. I imagine this could result in some campaigning, perhaps competative, by groups for their survival. The November 2003 AC Rep meeting will be in Japan. This means, of course, that my trip next year will be more expensive than usual. During the question and answer session, concern was expressed on funding for parallel activities such as WAI and the Semantic Web. This discussion happens every meeting, as does the financial management in harder times discussion. The line: parallel funding gives W3C free overflow without cost to the members. Skeptical members generally wondered if it went unfairly in the other direction as well. Patent Policy Update, by Daniel Weitzner Daniel introduced this session as "the moment we've all been waiting for". The patent policy is, as expected, rather complex, as evolved over the past year by input and participation from the interested partners. Basically, it allows defensive patents on essential technologies. The policy is primarily "royalty free". Being royalty free instead of RAND (Reasonable Access Non-Discriminatory ... or perhaps "Royalty Affordable"?) is a step toward liberalization beyond the RAND policy standards bodies typically adopt, and certainly distinguishes W3C from these other bodies. The policy also requires that all W3C members state applicable (potential) patents on a project during its requirements phase. This clarifies from a project's beginning that it will not develop into a format that conflicts with member patent claims. Working group members have further restrictions on patents on technologies from that working group participation. The final policy document will be out in February or March 2003. One question was on the exception of people outside of the working group. They have no licensing obligation, but the same disclosure obligation. A straw poll about allowing RAND exceptions was tied. During the question and answer period, I played champion of the naive by requesting a statement on why W3C didn't resolve the issue by setting up a centralized patent-making authority that generates the needed pre-emptive patents with certain phases of each Recommendation development. I compared their policy to taming the Wild West by banning fire-arms but providing instructions encouraging private use of cattle-prods and tear gas ... shouldn't we just hire a sheriff? My multiple choice question was responded to with "all of the above". It is legally, and technically, unwieldy to generate all potentially applicable pre-emptive patents, and even much more so to follow them up when challenged. It is better to have the patents, if any, spontaneously emerge from a free-market-like approach. Furthermore, you can't know what to patent, or even technically be allowed to patent it, since the problematic patents usually aren't about the standards themselves but about "essential technology" required for any implementations of a standard. One panelist suggested that the best pre-emption is early open source implementation of emerging Recommendation features. On the whole, the new policy isn't about setting up or preventing patents. It's more about getting companies to actively consider earlier on what their commercial interests and approaches are for emerging W3C technologies. TAG Report The TAG (Technology Architecture Group) establishes an overall architecture, and a vision for it, for the Web and all its activities. The hottest issue, and most of the questions, was about the XLink/HTML controversy. The usual comments were made, and not much was really resolved. During the question and answer period, I asked about why they made their position on XLink and HTML when the AC reps, and the XLink spec itself, stated explicitly the opposite position. I got a variety of answers from the different members. One answer was that the TAG statement is a "finding", that is, simply a suggestion. Another was that the HTML concern was only about legacy documents, and since XHTML 2.0 is a large departure from many old constructs, and backward compatibility is always maintained through deprecation anyway, HTML could be made XLink 2.0 conformant without conflict. Tim Bray added that adding element nesting is an important syntax design characteristic and that it should be added for HTML linking. Steven and another person from the HTML WG (Sebastian?) also asked questions. The HTML WG concern is that adding element levels is actually bad design. That is, while Tim Bray is an "elementalist", the HTML WG is "attributionist". The division between the two has been debated since the early days of SGML, sort of a neat/scruffy debate. In the context of HTML linking, elementalists want the better-defined complex structure of links that elements provide, while the attributionists want the easy of text editing that attributes provide. Stuart Williams is from the TAG. In the corridor we discussed the XLink/XHTML fiasco. He is interested in further discussion and meetings to resolve the issue. I offered my services in preliminary meetings to get the effort started, but declined participation in any resulting working group. There may be a teleconference in January on the issue. David Orchard of the TAG also approached me in the corridor. He said that my questions and comments to the TAG were accurate, appropriate and helpful ("right on the money", he said). He felt, as many suspect, that the TAG simply screwed up -- that Tim Berners-Lee and Tim Bray had simply overlooked and forgotten the XLink spec statement and the AC Rep decision when they should have remembered it and taken it into account. David felt the TAG should simply admit this, and thought perhaps they will. This paragraph is a potentially inflammatory opinion you can choose to ignore, but in any event shouldn't share. Tim Berners-Lee could very well have genuinely spaced the issue, since he relies heavily on his various lieutenants, and may have been simply following Tim Bray's lead on the issue. Tim Bray, on the other hand, is well aware of the issue, and has been for a long time. He was chair of the XLink WG for a time while I was a member, and he and I debated directly with each other over the issue at the time. Tim Bray is a knowledgeable, essential and well-respected expert in the field, but I've seen several times over the years how he uses his position to sneak his fundamentalist elementalist position into the W3C process when he knows it is far from a consensus position. If there's an issue you disagree with him on, you need to watch him as a hawk would, even if you think consensus agreeing with your viewpoint was thoroughly achieved and documented. Overall, it seems everyone agrees that in the long run an extension of XLink should be used with HTML that allows you to use intricate element structure to define link semantics for HTML attribute-based syntax and then define a mapping between the XLink-define semantics and the HTML syntax. This problem and the necessary solution have been known about for years, even if it is frequently forgotten by many. Although there is consensus on the solution, achieving the solution will be a lot of work, and it is unclear who will do it, and how. Perhaps the HTML WG should be encouraged to continue with HLink, with the agreement that is be such an XLink extension. The W3C Interaction Domain, by Philipp Hoschka Voice Browser and Multimodal are now active, large groups. Timed Text, an offshoot of SMIL that will define captions integrated into multimedia, will start in January and go on for two years. It still needs approval. Communications Report, Q&A, by Janet Daly The most interesting issue raised here is that the press reported a study done that shows only 3% of the W3C members' Websites conform to W3C recommendations. Janet did some campaigning, and got the rate raised to 4%, but clearly this should be made higher still. Steven Pemberton told me that CWI is among the 1% that become conformant during the past year, so congratulations everybody -- now all we have to do is make the CWI site accessibility compliant. The study was done by a graduate student who wrote a simple script that runs the Website for each member listed on the W3C Website through the W3C HTML validator and tallies how many have passed.