Minutes of W3C Advisory Committee meeting 14 Dec 94

This is a record of the first W3C advisory board meeting held at MIT LCS in room 518 on 14 December 1994. The meeting was open to all members and, exceptionally, to those actively pursuing membership. It was an informal meeting for members to get to know each other and discuss suitable activities for the consortium. See also:

Thanks to Mary-Ann Ladd/MIT for notes upon which these minutes were based.

Tim BL

Introductions

Albert Vezza, Associate Director of the MIT Laboratory for Computer Science, welcomed attendees who then introduced themselves.

The question was raised as to the relationship between the W3C and the European effort being discussed with CERN. Vezza and Berners-Lee explained that

MIT had an informal agreement with CERN that MIT would set up the W3C and would collaborate with CERN;
The exact arrangement with CERN was being worked out. MIT was in favor of the CERN team working closely as part of W3C, The relationship with work done by CERN under proposed funding from the Commission of the European Communities was to be clarified. Vezza was to attend a preliminary meeting in Brussels discussing this matter on the 15th.

There was strong concern expressed that there should not be a separate body working towards standards for W3 based in Europe. To further unity with Europe, approximately 50% of those present said they or a representative would attend a future W3C Advisory Committee meeting in Europe.

Berners-Lee described a little of the history of his attempt to find the optimal host sites for a consortium to ensure a single standard.

Vezza presented some of the work being done at LCS, as an introduction to the lab hosting the W3C. The lab has has 500 members engaged in a variety of research. Vezza showed excerpts from a video about LCS work.

Issues Brainstorming

The meeting continued with a brainstorming session attempting to establish the universal set of issues of concern to members and which the W3C might address. Berners-Lee's outline points were used as a nucleus. (These were largely the same as his talk for WWWF94 but as a number of members had been unable to attend that conference they were shown as reference to bring everyone to some common level). The list produced a list of issues on the whiteboard.

In response to questions, Berners-Lee defined the web as the universe of network-accessible information (data). Berners-Lee: The philosophy of W3 development had always been, and should be, to define the minimum necessary to ensure interoperability without making arbitrary decisions which would bind the web to particular current technology.

Contractual Issues

Paul Powell/MIT (PP) took questions on the contractual side of W3C membership.

Powell stated that from his experience, he would like to keep contracts as standard as possible. Companies who need to make superficial changes to the contract should contact him, but major changes are not acceptable.

Berners-Lee asked for pointers to a 48x48 transparent GIF image of member's logos and preferred URL for each to point to, for sponsor acknowledgement. Email to timbl@w3.org please.

Relationships between W3C and others

Comments from the floor: If W3C is to be a neutral convener in the web issues, how does it stay neutral? Neutrality will be defined by requirement for openness, as well as technical merits. W3C will not endorse propriety standards which benefit one company over another. We must avoid UNIX wars, which caused problems in the user community. The X consortium model is a very good one to take as a basis.

Dave Clark

Dave Clark <ddc@lcs.mit.edu> gave an "after lunch" talk illustrating work in progress at LCS and in the Internet, discussing the future of the Internet from an architectural point of view.

@@@ slides to be provided

W3C process

A list of relevant mailing lists was requested. [See the overview of WWW-related mailing lists in the W3 project information]

The W3C members mailing list is not currently public. It will at some point be trimmed to members who have signed the agreement.

Should W3C's process be opened to include input from persons not member of consortium? The X consortium did not expose much until the end of the process. One can however invite parties who are sufficient knowledgeable to join in process. There is a balance between expectations and availability to deliver.

Members should mail Berners-Lee with coordinates of people in their companies who are interested in particular activity areas.

Some members wanted to be able to decide priorities with colleagues, and would prefer to vote by email rather than settle it in the room. Berners-Lee should send round questionaire in X consortium style with points allocated by members to demonstrate their priorities.

It would be useful to build some collaboration tools for the use of the consortium itself. These could be based on the MIT/AI lab's work on the National Performance review Open Meeting.

Discussion of important project areas

Berners-Lee presented a status report on his current attempts to set up groups to work on standards problems. He asked members to forward names of individuals within member companies with expertise in particular areas. He outlined existing work outside W3C prior to its formation. See Berners-Lee's slides.

Points made not covered in the slides:

HTTP enhancements

We want to run several requests simultaneously much as netscape get good hypertext editor, common protocol

Security

See the security group information.

The next step is to review proposals regarding security

Architecture group

A proposal was discussed to set up a group to study the overall architecture, separately from the W3C AC. The dangers of top-down design and the merits of a wide view were mentioned.

It was decided that a separate architecture group was not to be formed.

Replication and Caching

There is a need for replication but very little work. (see slides)

The caching work should not prevent accurate logs from being kept. This is related to a more general discussion of demographic and personal data.

Intellectual Property: Content status issues

During the discussion, various issues merged related to marking the status of the content of information, including:

copyright and other intellectual property issues
payment
privacy
content rating (suitability for minors, etc). Generally agreed to be important and to be addressed.
the use and abuse of personal and demographic data about readers.

While technically one should define as little as possible, there is a need for a consistent model across all platforms.

Whereas most protocol extensions may be ignored by older agents, a method is required of making the implementation of these features a requirement for completion of a transaction.

W3C must separate mechanism from policy. W3 standards may include conventions for describing policy in these areas, but should not mandate particular policy.

HTML

HTML is important to the web as its lingua franca, and W3C is the place where standardization should take place.

The HTTP provides a marketplace in which new quite different data representations may be introduced by third parties. W3C need not be involved in these introductions.

What was known as HTML+ is now known as HTML levels 3 and 4. The specification remains under David Raggett's editorship.

HTML 2.1 enhances 2.0, but will proceed in parallel with the 3.0 specification.

Style

There is a need to support multiple presentation of the same content as a function of user preference.

Real time

There is a need to define how a hand-over occurs between HTTP and real time protocols, for example for video conferences, or remote logon.

Interactivity on the Web

Should W3C be concerned with it?

Should W3C start a group on collaborative use of W3?

API specification

The W3C should specify Application-Program Interfaces (APIs). The form of the API will have to be adapted per platform to fit platform-dependent conventions.

Reference code

There was discussion on the role of the W3C reference code, bounded by the following extreme positions:

W3C should concentrate on protocols, reference code is irrelevant as everyone will write their own.
W3C should go for test suites instead of code, as writing the code will be impossible, and test suites will be a better use of the resources.
Sample code is essential, as it is only by sample code that standards acquire any use or acceptance
Near production quality code is essential, as it will form a part of member's products.

There was a worry expressed that W3C would tackle too many roles for its resources with resulting lack of quality.

Reference code allows everyone to acquire state of the art functionality without reinvention, and so put their own effort into advanced features. Some members intended to incorporate reference code into their products, and adapt it to their environments. For them, it should be fast and bullet-proof. For comparison, the X code was more of a proof of concept in the early stages, but later developed into a stable product.

With other standards, no one has standard header files, things are opened to interpretation, even if there are written specifications. OMG, although its products are interface definitions, requests demonstrations from proposers and companies are bound to build code based on specifications.

There was consensus that W3C should produce industrial quality reference code for its relaeses but that it may also produce demonstration quality code to furnish specific examples outside the core code library.

Future Meetings

This committee is to cover business (marketing) aspects of W3C. It will meet every 6 months, possibly alternating on E and W coasts. It should meet in Europe at some point, to help ensure one standard. If outside continental US, W3C AC meetings should co-locate with WWW conferences to save travel costs. If within continental US, this would though have the disadvantage of overloading most US-based members.

Technical and business parts should be separate, but not at the same time so one person can attend both.

The decision-making processes of the group is spelled out in the contract, though many decisions will hopefully be taken by consensus without recourse to formal procedure.

There should be a formal charter published which defines the role of the W3C and its bounds. It is important to Send message that this is stable, ongoing endeavor

Should W3C make a statement regarding direction of reference code prior to release? In the case of competing proposals, we cannot do it until proposal has been selected. It is however useful to announce that work will be taking place in specific areas, to prepare industry. W3C's track record, rather than mere statements, will define whether W3 specifications are followed.

There was strong support for the X consortium model of single major releases.