- From: Tim Berners-Lee <timbl@w3.org>
- Date: Mon, 17 Dec 2001 16:11:35 -0500
- To: <www-tag@w3.org>
TAG members: welcome. First of all, a big thank you to everyone who allowed themselves to be nominated for offering their time and wisdom. To those elected, welcome, let us start. Thanks to Ian Jacobs, TAG team contact, for taking on issue and schedule tracking and some writing work for this group. I'd like to start by a little brainstorming over issues to be discussed. I suspect that this will be important but may be frustrating in the long term, as we will unearth so much to do that we will have no hope of covering it. The normal brainstorming method, which seems to apply here, is to first collect thoughts relatively randomly, and then to look for structure, so I will resist the urge to impose my preconceived structure at this point. (As you probably know, the architectural issues which I have found it personally necessary to note down over the years are in http://www.w3.org/DesignIssues. Of these, http://www.w3.org/DesignIssues/Architecture.html "Web Architecture from 50,000 feet" is an attempt at an overview, which was originally requested in 1998 by the cross-working group meeting. So that is where I would start, although it is now out of date.) References to other architectural documents would of course be very useful. The TAG members will met first (apart from on this list) in a teleconference Ian is scheduling. At that meeting I would expect us to start to lay out a framework in which to slot these issues, so as to lay out a plan to make the best use of our time. Before the meeting, then, please send your thoughts on the issues we should address to this list. I have noticed that there are two very connected ways in which the word architecture comes to be used: one is as the set of global invariants which ensure the operational properties of the system; the other is of a module diagram of the technologies or work items. The connection is that the architectural invariants create interfaces which make the modularity possible. But if there is a choice between defining invariants and defining module structure, I thing our job is more of the former. I would like to focus here on the fundamental issues, the ones which will be described in some core documents which the TAG will produce. There are indeed many urgent but less fundamental issues arising in current work in W3C and elsewhere, but I would like to keep those separate. We will also have to distinguish between issues which we think are well understood and in dire need of documentation, from those which are not widely understood, or not widely agreed. I hope we find most are in the former category! Tim
Received on Monday, 17 December 2001 16:11:35 UTC