IRC log of tagmem on 2002-01-07

Note: This IRC log has been edited, primarily to remove interactions prior to discussion beginning in earnest, and after the meeting was adjourned.

Timestamps are in UTC.

16:13:17 [Ian]: Introductions
16:13:32 [Ian]: TimBL:
16:13:40 [Ian]: Welcome!
16:13:42 [Ian]: Questions:
16:13:44 [DanC]: Agenda: http://lists.w3.org/Archives/Member/tag/2002Jan/0001.html (member-confidential. :-{)
16:13:48 [Ian]: a) What do you bring to the TAG?
16:13:57 [Ian]: b) What burning issues would you like to discuss?
16:14:37 [Ian]: TimBL: I'm interested in clean semantic web infrastructure.
16:15:01 [Ian]: ...most important question for the TAG: if I give you a URI, how can you figure out what it stands for?
16:15:25 [Ian]: ...my feeling is that we haven't got the chaining of specs articulated. Hand-off from http -> mime -> xml namespaces not well-defined.
16:15:40 [Ian]: Roy:
16:16:04 [Ian]: - I started messing around with the Web in 1993 to find interesting things about New Zealand.
16:16:25 [Ian]: - Approached from the standpoint of specs; worked with NCSA software.
16:16:46 [Ian]: - Wrote libwww-perl
16:16:57 [Ian]: - Got involved in IETF first for relative URLs, then for HTTP.
16:17:35 [Ian]: - Finished dissertation last year (UC Irvine). Since then been working "in the small" client projects.
16:17:55 [Ian]: - I conclude I prefer working in the large.
16:18:22 [Ian]: - Burning issue: Figure out what W3C WGs have been working on while I've been out of touch...
16:18:45 [Ian]: David Orchard:
16:18:56 [Ian]: - Got involved mid 1994. Vancouver, CA
16:19:13 [Ian]: - Did one of first govt. web projects: maps.
16:19:30 [Ian]: - Joined IBM in 1996; we started a lab all about the Web.
16:19:40 [Ian]: (grew from 2 - 300 people).
16:20:06 [Ian]: - Lead architect for that lab during four years there; always with customers.
16:20:23 [Ian]: - 1997: On a panel with Tim Bray, who suggested that I look at XML...
16:20:50 [Ian]: - At BEA Systems; very interested in Web services
16:21:38 [Ian]: - Burning issue: No coherent Web architecture document that stitches the pieces together, for either audience - new user / developer, or normative audience.
16:21:46 [Ian]: ...such a document would allow us to identify holes.
16:22:10 [Ian]: Ian:
16:22:33 [Ian]: - Writing resource for this group.
16:23:55 [TimBL]: Zakim, who is here?
16:23:56 [Zakim]: I see TimBL, Roy, Chris (muted), Ian, Norm, DanC, Bray, DaveO, Cotton, Stuart
16:24:09 [TimBL]: Chris, you are next
16:24:21 [TimBL]: Zakim, unmute Chris
16:24:22 [Zakim]: Chris should no longer be muted
16:24:24 [Ian]: - I will be writing E&O pieces.
16:24:30 [Ian]: Chris:
16:24:47 [Ian]: - Got involved in Web when I telneted into CERN a long time ago.
16:24:55 [Ian]: - Got interested when there was a client that could display images.
16:25:21 [Ian]: - When GIF patent hit, I was involved in producing PNG spec.
16:25:32 [Ian]: - Roy and I had lunch at WWW4. :)
16:26:08 [Ian]: - I got involved in fonts and css, but am happy to be doing less of that now and doing graphics.
16:26:12 [Ian]: - My focus is client-side.
16:27:00 [Ian]: - Burning issue: How do we make our specs so that we don't use broken html clients 10 years down the road?
16:27:01 [Ian]: Norm:
16:27:02 [TimBL]: Zakim, mute Chris
16:27:02 [Zakim]: Chris should now be muted
16:27:29 [Ian]: - 1994: Involved with DocBook at O'Reilly.
16:27:47 [Ian]: - Moved to ArborText; consulted on SGML and XML
16:28:02 [Ian]: - Worked closely with Paul Grosso and Eve Maler there.
16:28:13 [Ian]: - Then did some xslt at ArborText before moving to Sun.
16:28:46 [Ian]: - Burning issue: Not yet articulated. Mostly I'm concerned with the general complexity with the direction of things. A better model might help us simplify.
16:28:57 [Ian]: DanC:
16:29:24 [Ian]: - Background - comp sci and math, then went into industry doing online documentation at a supercomputer company.
16:29:36 [Ian]: - In 1991, met with with a group that become the DocBook folks.
16:29:53 [DanC]: http://groups.google.com/groups?selm=6487%40cernvax.cern.ch
16:30:04 [Ian]: (original web announcement)
16:30:12 [Ian]: - Started studying standards, DTDs, etc.
16:30:27 [Ian]: - Disappeared into a startup, came out a year later. Went to first WWW conf.
16:30:42 [DanC]: HTML Dialects: Internet Media and SGML Document Types
16:30:42 [DanC]: 6 March 1996, Dan Connolly
16:30:42 [DanC]: http://www.w3.org/MarkUp/WD-doctypes
16:30:44 [Ian]: - starting in 1995, chair of HTML WG (3.2, 4.0).
16:30:47 [Chris]: Chris and DanC were both on the IETF HTML WG
16:31:10 [Chris]: as was Jon Bosak
16:31:45 [Ian]: - Then early xml discussions
16:32:07 [Ian]: - Now semantic web. Formalizing w3c process interesting. Bane of my existence is doing something the computer should do for me.
16:32:16 [Ian]: - Burning issues:
16:32:31 [Ian]: a) Bank web sites are broken. They use post instead of get, they use javascript when they should not, etc.
16:33:03 [Ian]: ...I want to use the web for mundane tasks. This is possible in some cases, but only with some clients. I want to do mundane tasks through a variety of clients.
16:33:12 [Ian]: b) Value of formal systems in standards process.
16:33:22 [Ian]: Tim Bray:
16:33:36 [Ian]: - I'm on old database and search geek. Discovered the web in mid-1994.
16:33:53 [Ian]: - Thought it would be a cool thing to do full-text search. Lycos and Infoseek popped up.
16:34:03 [Ian]: - 1996; Went independent, discovered XML WG.
16:34:30 [Ian]: - I got some visibility during this time since I was independent and could talk to press without restriction. :)
16:34:32 [Ian]: - WASP
16:34:37 [Ian]: - Startup in Vancouver today.
16:34:39 [Ian]: Burning issues:
16:34:48 [Ian]: - Coherency and simplicity.
16:34:54 [daveo100]: daveo100 has joined #tagmem
16:35:30 [Ian]: ...I think it's reasonable at W3C to want to make the Web do more things. At the same time we want to allow people reasonably to figure out what's going on by looking at source code and specs.
16:35:42 [DanC]: hear hear, on "you should be able to bootstrap yourself by ViewSource". Self-describing web!
16:35:57 [Ian]: - Concerned about hype in the industry (e.g., not sure what web services is yet).
16:36:05 [Ian]: Paul Cotton:
16:36:14 [Ian]: - Old database and search geek, too.
16:36:28 [Ian]: - In early 1990's working for a full-text retrieval competitor to company of TimBray's.
16:36:43 [Ian]: - Got involved with query languages.
16:36:54 [Ian]: - Been doing computer stds work since mid-1980's
16:37:23 [Ian]: - Worked at IBM, then Microsoft.
16:37:37 [Ian]: - In 1998, went to XML re-org meeting.
16:37:47 [Ian]: - Moved to Microsoft in 2000.
16:38:08 [Ian]: - I spend nearly 90% of my time in W3C (Chairman of XML Query WG, former AB, now TAG).
16:38:11 [Ian]: Burning issues:
16:38:24 [Ian]: - Composability.
16:38:36 [Ian]: ...I'd like to make sure that the specs we have work together.
16:38:39 [Chris]: I agree about the composability
16:38:54 [Ian]: - Also interested in how XML and Web services will go forward.
16:39:06 [daveo100]: daveo100 has left #tagmem
16:39:08 [Ian]: - Also interested in relationship between TAG and other organization units inside and outside W3C>
16:39:25 [Dave]: Dave has joined #tagmem
16:39:37 [Ian]: ...how will TAG interact with WAI, I18N groups?
16:39:48 [Ian]: - See request for liaison with ebXML work:
16:39:56 [DanC]: '98 XML reorg meeting: http://www.w3.org/XML/Group/1998/08/ftf.html [member-confidential]
16:40:12 [Ian]: Stuart:
16:40:42 [Ian]: - Embedded systems work
16:40:48 [Ian]: - Maclab bridges.
16:41:09 [Ian]: - British Telecom participation in standards (802.1 part D)
16:42:02 [Ian]: - Moved to HP labs in Bristol. started working on protocols for small devices.
16:42:25 [Ian]: - Spent 5-6 years (around early years of web - '92-2000) working on infrared protocols and deployment in small devices.
16:42:48 [Ian]: - Some of that is about devices coming together with little knowledge of each other and to do useful things for users.
16:42:55 [Ian]: - Related to web services and semantic web.
16:43:05 [Ian]: - Relative newbie to web as a contributor.
16:43:13 [Ian]: Burning issues:
16:43:31 [Ian]: - In specs, people start with syntax rather than model.
16:43:37 [Chris]: yes, bottom-up specs are a problem
16:43:51 [Ian]: Stuart: ..I like being able to create pictures in my mind.
16:43:54 [Chris]: "if byte zero is E3 then .."
16:44:06 [Ian]: - I'd like to see an articulation of what something is trying to accomplish rather than imperatives about how to do it.
16:44:22 [Ian]: - XML Protocols: split between bottom-up and top-down styles.
16:44:44 [Ian]: ....HTTP as transfer protocol v. SOAP as extension.
16:45:05 [Ian]: ....honoring the arch. history of the web as moving forward into web services.
16:45:14 [Ian]: =====
16:46:21 [Ian]: Meetings and time commitments.
16:46:46 [Ian]: TimBL: Should we start with weekly meetings?
16:46:47 [Dave]: Dave has quit
16:46:57 [Ian]: [No objection to weekly meetings, at least to start.]
16:47:07 [Chris]: Chris (muted) agrees with weekly meetings
16:47:19 [DanC]: 90min: 2nded
16:47:23 [Ian]: TimBL: We could make these meetings longer than an hour. 90 minutes may be a good length.
16:47:26 [Stuart]: 90 mins is fine
16:47:26 [Chris]: Chris could not manage the next hour, but can do the preceding hour
16:47:31 [DanC]: 90min every other week could be good
16:48:05 [DanC]: I could also join earlier; 30min earlier, anyway
16:48:16 [Ian]: Proposed: 10:30-12pm ET
16:48:18 [Chris]: yes, 30 min earlier id fine for me
16:48:36 [Norm]: I think meeting every week is probably good for a start
16:49:00 [Ian]: RF: No time constraint, but I prefer later...
16:49:13 [Chris]: If we met fortnightly we would never get started, i feel, so i agree weekly 90 minutes is good
16:49:34 [Ian]: Tim Bray: Should I go to France next month? Will there be work to do?
16:50:33 [Ian]: My understanding: Weekly meeting, 90 minutes, starting 30 mins earlier.
16:50:46 [Dave]: Dave has joined #tagmem
16:50:55 [Chris]: OK so I will have to leave at the top of the hour
16:50:57 [Ian]: TimBL: We'll meet 30 extra minutes today.
16:51:19 [Ian]: ===========
16:51:23 [Ian]: Architecture
16:51:32 [Ian]: TimBL: Some people want top-down description.
16:51:38 [Ian]: ...other people thing about the set of invariants.
16:52:21 [Ian]: The relationship between the two is not obvious. Web services are not new top-down modules; just a different axis.
16:52:41 [Ian]: TimBL: Formal specs: should we write these things down in formal language?
16:52:51 [Ian]: TimBL: Should TAG focus on past or future?
16:53:16 [Ian]: Roadmap for web services or getting understanding of http?
16:53:40 [Ian]: Tim Bray: I don't think there's a sufficient body of knowledge about what web services will be to draw a roadmap.
16:53:49 [Chris]: I prefer designing the future to reverse-engineering the past
16:53:54 [DanC]: I like looking at the Web w.r.t. Engelbart's 13 requirements. http://www.w3.org/Architecture/NOTE-ioh-arch
16:54:13 [Ian]: TimBL: Might need a web services planning group.
16:54:39 [Ian]: RF: I have an interest in trying to make sense of the various technologies that we're developing software (apache foundation) for.
16:54:52 [Ian]: ...part of our charter is that the web infrastructure is never owned by any single company.
16:55:47 [Ian]: ...One problem I faced in recent past - Web services as means of communicating RPC over web connection.
16:55:59 [Ian]: ...I'd like to see a better description of the web services architecture.
16:56:17 [Ian]: ...more consistency with existing web architecture.
16:56:53 [Ian]: Stuart: In terms of consistency with web architecture; I think there's a lot written on web architecture but it's spread out.
16:57:15 [Ian]: ...I don't think we have a collective reference about what web architecture is. We should have a statement about what it is.
16:57:21 [Ian]: TB: Read RF's dissertation.
16:57:33 [DanC]: how big is Roy's dissertation? I've seen presentation materials on it.
16:58:25 [Ian]: RF home page:
16:58:26 [Ian]: http://www1.ics.uci.edu/~fielding/
16:58:36 [DanC]: I'm seeing quite a bit of consensus forming around RF's dissertation...
16:59:40 [Norm]: roy's dissertation: http://www.ebuilt.com/fielding/pubs/dissertation/top.htmhttp://www.ebuilt.com/fielding/pubs/dissertation/top.htm
16:59:44 [DanC]: RF's thesis http://www1.ics.uci.edu/~fielding/pubs/dissertation/top.htm
17:00:15 [DanC]: this channel is irc://irc.w3.org:6665/tagmem
17:01:53 [Ian]: TimBL: Holes in architecture?
17:02:08 [Ian]: DanC: My issues tend to be focused around other groups (e.g., RDF core, HTML WG).
17:02:54 [Ian]: In addition to the production of Recommendations, the TAG will help resolve technical issues having architectural impact.
17:02:55 [Ian]: The process for issue resolution is likely to evolve over time. The initial process is:
17:02:55 [Ian]: ]]]
17:02:55 [Ian]: -- Technical Architecture Group (TAG) Charter
17:02:55 [Ian]: http://www.w3.org/2001/07/19-tag
17:02:56 [Ian]: Thu, 19 Jul 2001 13:29:12 GMT
17:03:00 [Ian]: [Process follows]
17:03:17 [DanC]: RF's thesis is 152 pages, btw.
17:03:22 [Ian]: TimBL: I'm inclined to start with one or two documents as a way to go: find the contentious bits.
17:03:46 [Ian]: TimBL: What bits of RF thesis should be required reading for next week?
17:04:09 [Ian]: Tim Bray: Large part of the dissertation is a taxonomy for describing architectures.
17:04:17 [DanC]: chapter 5 on REST: http://www1.ics.uci.edu/~fielding/pubs/dissertation/rest_arch_style.htm
17:04:26 [Ian]: Or chap 4?
17:04:30 [Ian]: CHAPTER 4: Designing the Web Architecture: Problems and Insights
17:04:52 [DanC]: CHAPTER 4 http://www1.ics.uci.edu/~fielding/pubs/dissertation/web_arch_domain.htm
17:05:36 [Ian]: DanC: Is there a relevant tome about "everything you need to know about HTML?"
17:05:47 [Ian]: ...suppose the web arch is net protocols and data formats.
17:05:52 [Ian]: ...what's good to read on data formats?
17:06:37 [Ian]: What about annotated XML 1.0?
17:06:43 [Ian]: TB: A lot more has happened since then.
17:06:59 [Ian]: ...nowhere has anyone written a tutorial on how descriptive markup differs from other types of markup.
17:07:29 [Ian]: TimBL: What are the invariants? What would you have to get the same if you were redesigning XML?
17:07:30 [Chris]: I have
17:08:18 [DanC]: ... have read Web Architecture from 50,000 feet http://www.w3.org/DesignIssues/Architecture.html that is, right, Chris?
17:08:21 [Norm]: not recently. I'll reread :-)
17:08:24 [Chris]: yes
17:08:24 [Ian]: http://web3.w3.org/DesignIssues/Architecture.html
17:08:34 [DanC]: $Id: Architecture.html,v 1.56 2000/09/08 17:38:37 timbl Exp $
17:08:58 [Ian]: TimBL: I'm interested in comments on the early principles. I think there's more disagreement about later principles.
17:09:13 [Ian]: (e.g., dereferencing namespaces)
17:09:21 [Ian]: s/namespaces/namespace URIs/
17:09:39 [TimBL]: http://www.w3.org/DesignIssues/Architecture.html
17:09:49 [TimBL]: http://www.w3.org/DesignIssues/Axioms.html
17:10:39 [Ian]: Homework assignments so far:
17:10:44 [Ian]: - RF chapters 4-6
17:10:57 [Ian]: - 50k foot Arch
17:11:14 [Ian]: David: Can we create a CVS-accessible archive that we can work within?
17:11:18 [Ian]: ...just for TAG work.
17:11:26 [DanC]: I think it was RF that asked, not David
17:11:27 [Ian]: Quip: tag.sourceforge...
17:11:33 [Ian]: thanks dc
17:11:45 [Ian]: TimBL: What access control?
17:12:27 [Ian]: TimBL: /2001/tag/
17:13:30 [Ian]: Action TimBL: Find out what kind of access is possible.
17:14:14 [Ian]: TimBL notes that at W3C, when you check into cvs, published to the Web.
17:14:48 [Ian]: TimBL: You can also use www-archive to give something a URI (if you don't want to use www-tag).
17:14:55 [Ian]: www-archive@w3.org
17:15:02 [Norm]: Go for it DanC, maybe I'll be converted :-)
17:15:07 [Ian]: There is also w3c-archive@w3.org (Member-confidential)
17:15:16 [DanC]: "go for it" it== a Wiki
17:15:24 [Norm]: yes
17:15:45 [Ian]: ========
17:15:47 [Norm]: Maybe the truth is we need both, a Wiki for "scribbling" and CVS to control documents we'll publish on a REC track
17:15:53 [Ian]: TimBL: Do we do Web Services or not?
17:16:13 [Ian]: ...TAG not chartered to design web services.
17:16:41 [Ian]: NW: I don't think we should give web services special status.
17:16:53 [Ian]: ...we should examine issues large enough to encompass other work.
17:17:09 [Ian]: DC: If the entire web disappears behind "POST" due to web services, I will be annoyed.
17:17:19 [Ian]: ...I'd like to use the TAG as a lever to prevent this.
17:17:31 [Ian]: NW: Sounds like a general principle: this is a bad thing, not just for web services.
17:17:41 [Ian]: RF: And the counter-proposal to that approach...
17:18:22 [Ian]: DC: Sometimes doing nothing is better than putting w3c name on something that takes us backwards.
17:18:57 [Ian]: -----
17:19:07 [Dave]: the last RF was Dave
17:19:10 [Ian]: So for homework assignments; please identify holes or bits where you don't agree.
17:19:46 [Ian]: PC: We should action people to take a look at input mandate. Try to summarize into a list of things we might want to list or respond to.
17:20:12 [Ian]: Here's a rudimentary grouping:
17:20:17 [Ian]: URIs
17:20:17 [Ian]: - URIs v. URI references [4]
17:20:17 [Ian]: - Resource discovery: Limits of URIs [7] (public comment)
17:20:17 [Ian]: - Relative or absolute? [8]
17:20:17 [Ian]: - Fragments ok or not? [8]
17:20:18 [Ian]: - URNs sometimes, always, never? [8].
17:20:20 [Ian]: - RDF trick of trailing #? [8]
17:20:22 [Ian]: - Non-http schemes [9]
17:20:25 [Ian]: HTTP
17:20:28 [Ian]: - HTTP not well understood [5]
17:20:29 [Ian]: XML
17:20:32 [Ian]: - XML packaging [6]
17:20:33 [Ian]: - XML profiles [6]
17:20:35 [Ian]: - XML processing model [6]
17:20:38 [Ian]: - XML type library [6]
17:20:42 [Ian]: Global
17:20:45 [Ian]: - System architecture document [6]
17:20:46 [Ian]: - Is simplicity a useful architectural constraint? [10]
17:20:48 [Ian]: (It's the first one in Principles of Design [11]...).
17:20:50 [Ian]: [4] http://lists.w3.org/Archives/Public/www-tag/2001Dec/0019
17:20:53 [Ian]: [5] http://lists.w3.org/Archives/Public/www-tag/2001Dec/0020
17:20:57 [Ian]: [6] http://lists.w3.org/Archives/Public/www-tag/2001Dec/0022
17:20:58 [Ian]: [7] http://lists.w3.org/Archives/Public/www-tag/2001Dec/0011
17:20:59 [Ian]: [8] http://lists.w3.org/Archives/Public/www-tag/2001Dec/0030
17:21:01 [Ian]: [9] http://lists.w3.org/Archives/Public/www-tag/2001Dec/0031
17:21:02 [Ian]: -------------
17:22:37 [Ian]: Action PC and IJ: Summarize input, liaison request, panel discussion.
17:23:27 [Ian]: TimBL: Groupings could go into group space.
17:23:37 [Chris]: I note that client stuff (DF, Int) also have architecture, that is not reflected in that list
17:23:43 [Ian]: RF: I would like to look at events schedule for the next 6 months.
17:24:03 [Ian]: ...and where should TAG meet technically v. socially.
17:24:07 [Ian]: TimBL: Also first ftf meeting.
17:25:05 [Norm]: I have constraints on both ends; so I think we're sunk
17:25:16 [Ian]: TB: I think that there's not a compelling case for me to go to France.
17:25:28 [Chris]: I will be there anyhow
17:25:31 [Chris]: being local
17:25:35 [Ian]: DC: I agree, but there will be a session at the plenary (panel) about the TAG. Some subset of TAG will be there.
17:25:54 [Ian]: DC: Next event planned is Hawaii.
17:26:09 [Ian]: Stuart: I don't know whether I will be there yet.
17:26:23 [DanC]: I could also do a meeting in April, perhaps, if I found out about it right away
17:26:49 [Ian]: TimBL: In Hawaii we are likely to get less than a whole day.
17:27:09 [Ian]: AB meeting 4 May. AC meetings 5-7.
17:27:18 [Ian]: PC: We could meet morning of Sunday 5 May.
17:27:49 [Ian]: Action PC: Send proposed meeting time to tag mailing list.
17:29:03 [Ian]: IJ: My understanding of Cannes situation: TimBL not there, difficult for RF and TB. But panel will speak on weds. So not a formal WG meeting but an opportunity to chat and meet socially.
17:29:44 [Ian]: DO: We could ask people what issues are not being addressed specifically enough that the tag should address. And what output (documents) should be produced.
17:30:59 [Ian]: Action DO: Write to chairs asking for input to TAG, suggest to reply to www-tag.
17:31:15 [Ian]: IJ notes that there are likely to be N architecture documents, not just one Architecture Document.
17:32:16 [Ian]: IJ: Should I send minutes to tag@w3.org or www-tag directly (for correction)?
17:32:20 [Ian]: Resolved: Send minutes directly to public list.