W3C | TAG | Previous: 17 Feb teleconference | Next: 10 Mar 2003 teleconf

Minutes of 24 Feb 2003 TAG teleconference

Nearby: IRC log | Teleconference details · issues list · www-tag archive

1. Administrative

  1. Roll call: NW (Chair), TB (Scribe), DC, RF, TBL, SW, CL, PC. Regrets: IJ, DO?
  2. Accepted 17 Feb telecon minutes
  3. Accepted this agenda
  4. Next meeting: 10 March.

1.1 Meeting planning

May and Nov meeting dates left open until 10 Mar meeting. IJ to update Member cal when known.

May

[TBray]

Chris can do alternate Budapest dates
DanC: might not be able to make dev-day session
Stuart may have trouble with 24th too
[C3s]
I can juggle my dev day presentation most likely to avoid a clash
[TBray]
No input from TimBL on these dates yet, need to get it

Holding May dates open a little longer...

Nov

[Roy]

okay by me
[TBray]
Need input from DC & TBL
DanC: not aware of any conflicts
Leave these open, try to close on 10th of March
Keio can host us on those dates
[Chris]
we did get confirmation from keio that they could host us
but CL is sure now

1.2 Technical Plenary presentations

Resolved: Review slides Tuesday, 4 Mar in Boston.

1.3 Mailing list management

[TBray]

SW: has action item outstanding to update policy & publish it. Made an interim intervention, which seems to have helped
Dan: post-f2f, we did everything wrong; flameburst following on TimBL's post with TBL on vacaation
SW: wait for my action item?
[Chris]
we can ask for better quoting discipline; three pages of quoted matter without comment is not acceptable
[TBray]
Proposal from someone: ask people not to post more than once per day without reply
[Chris]
prefer leading by example to constraining by rules
[TBray]
TBray: don't like doing by policy, it's an individual judgement. Propose offline intervention with people causing problems
Dan: some people are way out into the territory of wasting everyone's time; perhaps a private email to them?
Norm: SW will finish action item, and when individuals get out of line, it's appropriate for the chair to intervene
DanC: of course this may take days to get to
Norm: of course TAG members could send direct email to chair acting for intervention

1.4 Other stuff

2. Technical (70min)

2.1 Site metadata hook

[Roy]

favicon.ico
[TBray]
with only 3 thiings, not too bad a prob, but this is a slippery slope
[Chris]
reserved urls /robots.txt, /w3c/p3p, /favico
[TBray]
TBL reviews points in his posting referenced above
[Chris]
guys, stop putting technical discussion in /me
is the question : given a uri x, how to get metadata about x?
or is it given a site s, get metadata?
[TBray]
TBray: 1. support adopting the issue
[Chris]
one persons data is another persons metadata
[timMIT]
HTTP DNA domain metadata could well include delegation information giving actual notional "sites"
[TBray]
TBray 2. web arch currently doesn't have notion of a "site" and to the extent it does it's coupled to host (e.g. robots.tx); so this is new but might be good
TBray: recent proposal along same lines from (I think) Roger Costello
TBray: TBL said HTTP "tag" meant header
Roy: robots.txt isn't necessarily a file
Roy: this isn't metadata it's just data about a resource
[Chris]
any resource is not necessarily a file
[Zakim]
timMIT, you wanted to define site in the context of this issue proposal only
[Roy]
no, data about a site
not a resource
[Chris]
ok
[TBray]
Roy: we need to manage this whole area of per-site names
[Chris]
there is no way to give a URI of a site as opposed to a URI for a welcome page for it
hmm... sites are significant resources, no? so they should have URIs.....
[Roy]
/
[TBray]
TBL on lack of distinction between data/metadata
TBL on whole family of interesting metadata you could have about a site
TBL: need a hook to hang this stuff
[Zakim]
Chris, you wanted to talk about subsites, tenants, server sharing etc
[TBray]
No, "/" isn't the site it's the server, they're not the same things
[timMIT]
Server isn't a perfect name eitehr ... tends to be a computer.
[TBray]
Chris: echoing problem of site/server disconnect, bad architecture to require everyone to write one file
Chris: if a Site is an important thing, it should have a URI; right now there's no such thing
Chris: per our axioms
Roy: When robots.txt was invented.. (Chris: everyone had their own server) .. the idea was to knock politely on some part of a naming authority's domain
Roy: haven't seen a proposal yet with equivalent semantics
[Chris]
it has had excellent expressive power at ultra low implementation cost
[Zakim]
timMIT, you wanted to explain to roy where this fits in
[TBray]
TBray: wants to introduce a new notion called "site" a collection of resources (on one server?)
TBray: "Site" has a URI, which could be provided in an HTTP header and an HTML <link>
TBray: could contain robotrs policies, RSS feed, all sorts of stuff
[Norm]
I can't see how you're going to give site a URI independent of the pages on the site...
[TBray]
Roy/TBL: Problem because many sites consider the root URI to be revenue-significant and don't want robots to go there
[timMIT]
A head would work
[TBray]
Roy: but likes TBL's idea
[timMIT]
a HEAD would work.
[TBray]
Roy: wants the issue to be tightly circumscribed
Roy: i.e. we're just solving /robots.txt (but that cat's out of the bag) or more generally, algorithm for determining appropriate metadata for a site
TBL: but doesn't like metadata/data distinction
TBL: how would we design robots.txt if we were doing it now or ina couple years
Bray: propose we accept SiteData-NN
[timMIT]
SiteData-$int(ian++)
[TBray]
Chris: does that include defining notion of a site?
Bray: yes
Roy: rather empower authors to define their own site
Roy: rather than define for them what it is
[Zakim]
DanC, you wanted to support the issue as proposed in Proposed issue: site metadata hook http://lists.w3.org/Archives/Public/www-tag/2003Feb/0093.html
Chris, you wanted to clarify
[Stuart]
Just found "What if I can't make a /robots.txt file?" at http://www.robotstxt.org/wc/faq.html#noindex
[TBray]
Chris just wants to make sure we don't leave undefined terms like "site" hanging
Roy: can we define it reflectively
TBray doesn't understand Roy
[Roy]
aww
[TBray]
Norm: any objection?
Resolved: Accept issue siteData-36.
Chris: owner?
[Roy]
all resources on "site" point to same "site URI"
[TBray]
what roy said
Issue owner: TBL
Action item: proposal to close it
TBL: not till after discussion
[Roy]
next number is 36
[TBray]
I think this is SiteData-36
Action TBL: Summarize discussion & recast issue
Action TBray: Post a strawman proposal

Actions accepted

2.2 namespaceDocument-8

[TBray]

Bray: proposal: http://lists.w3.org/Archives/Public/www-tag/2003Feb/0213
Very minimal, see exegesis in my covering email
Dan: this isn't XHTML, they own the syntax (Chris doesn't agree)
Dan: would prefer a custom XML or RDF language, but not enough to object; would abstain
Norm: you really think that HTML-WG has to approve any attributes in any namespaces
Dan: yes
Chris: flabbergasted
[Norm]
I'm a bit flabbergasted as well
[TBray]
Dan: doesn't like it but thinks that's the way it is
[Chris]
Actually to clarify - they DO own the syntax, no argument; the syntax of the HTML namespace. Attributes in other namespaces they do not own and this was what I objected to in Dan's statement
[TBray]
Bray: what about modularization
Dan: then you have to change the DOCTYPE
Chris: if you want it to be valid
[Chris]
if you want it to be valid you would need to change the doctype and write a driver dtsd for it etc
[TBray]
Bray: Granted
[DanC]
chris, there aren't any XHTML documents that aren't valid XML, are there?
[Zakim]
Chris, you wanted to correct TimB
[TBray]
Bray: not sure what the correct term is
Chris: It's an XHTML-family doc, which is a defined term in the XHTML spec
Bray: in technical terms, it's XHTML + 2 attributes, which is easy to understand and implement
[timMIT]
3
[TBray]
Norm: want to change proposal?
Bray: no
Dan: does proposal want to change DOCTYPE
[Zakim]
timMIT, you wanted to express the concern that teh semantics are notwell defined in rddl
[TBray]
Bray: silent on that subject
Paul: we're open to suggestions
[Chris]
got it
http://www.w3.org/TR/2001/REC-xhtml-modularization-20010410/conformance.html#s_conform_document
[TBray]
TBL: covering letter said RDF wouldn't give semweb people what they wanted; ?
[DanC]
thx, chris. that's new to me. but it does involve changing the doctype... "A conforming XHTML family document is a valid instance"
[TBray]
TBL: suggests that RDDL semantics be given in RDF terms, as classes & properties
[timMIT]
1. The cover note suggets the RDDL document does not meet its SWeb goals. In what way?
[TBray]
Bray: Various RDF instantiations either fail to capture the linkage to the namespace as a resource, or are really complex
Bray: prepared to believe that RDF-defined semantics are a agood idea, who's going to write it down?
Dan: I would, but I wouldn't use XHTML, I'd use RDf anyhow
TBL: if introducing a thing called "nature", if you make it an RDF Class then that explains it to a lot of people and you don't need to say anything more
[Norm]
PC+
[TBray]
Norm: proposal could be left alone and people who wanted to do the RDF definition could do so
TBL: no, interoperability suffers
Paul: pushing back on Dan's thesis that we should use RDF
Paul: we agreed that NS doc should be human readable
Paul: and there were other issues with regards to using RDF in XML
Dan: RDF can be as human-readable as you like
Dan: consumer is a machine not a human
Paul: disagrees strongly
Paul: we have two objectives, hard to achieve both
[Zakim]
TBray, you wanted to say that I don't know what an RDF class is
[TBray]
TBray: can we publish a XSLT or other code that would process a minimal-RDDL and emit the RDF that you'd like to see?
Chris: user-agents, given XML & a stylesheet, typically don't work
[Zakim]
DanC, you wanted to ask for a use case to focus on
[TBray]
Bray: Consider WordML; human perl programmer could dereference namespace name to figure it out
[DanC]
ok, thanks for the use case.
[Zakim]
timMIT, you wanted to wade into this one
[TBray]
Bray: Also the desperate perl hacker could dispatch to code via RDDL to generate postscript etc
Dan: but that's hard, subtle, hard to believe, given the experience of MIME dispatching
TBL: agree that it's usable to have both;
[DanC]
would people please stop saying "we've agreed to X"? I'm quite confident we have resolved *nothing* anywhere near this issue.
[TBray]
TBL: consider high-volume applicatios, apps hitting this thousands of times a second, the architecture has to support this
[Norm]
I'm not sure I agree it's infrequent...
[TBray]
TBL: use case only appeals to fairly infrequent access
TBL: If it doesn't have well-defined semantics people won't use it. Dan & I would both put RDF there.
[Roy]
scenario: human wants info about namespace (I don't care about automation here)
[timMIT]
?RDDL_based?
[TBray]
Paul: perplexed how to handle at technical plenary, this has been going on for a long time and he hasn't seen statements from TBL, DC in public that non-RDF was unacceptable
[Chris]
wondering about proposing reserved paths nsURI/rdf/ and nsURI/schema and so forth
[TBray]
Dan: not saying "has to be RDF" - he's saying he would prefer RDF & would abstain on this proposal
[timMIT]
I don't know what Paul meant by "RDDL-based" of all these various proposals for RDDL
[TBray]

... discussion of technial minutiae of how to make it valid, with DOCTYPE wrangling and so on ...

2.3 Other issues

The TAG is likely to review action items associated with these issues.

2.4 Architecture document

See also: findings.

  1. 21 Feb 2003 Editor's Draft of Arch Doc:
    1. Resolve to request publication of this draft (with modifications?) on TR page?
    2. Action DC 2003/02/06: Attempt a redrafting of 1st para under 2.2.4
    3. Action DC 2003/01/27: write two pages on correct and incorrect application of REST to an actual web page design
    4. Action DO2003/01/27: Please send writings regarding Web services to tag@w3.org. DO grants DC license to cut and paste and put into DC writing.
    5. Action CL 2003/0127: Draft language for arch doc that takes language from internet media type registration, propose for arch doc, include sentiment of TB's second sentence from CP10.
    6. Action TB 2003/01/27: Develop CP11 more: Avoid designing new protocols if you can accomplish what you want with HTTP. DC suggested describing GET/PUT/POST in a para each, then say "if your app looks like that, use HTTP". Proposal from TB to withdraw the proposal.
Ian Jacobs for Norm Walsh and TimBL
Last modified: $Date: 2003/03/25 05:25:43 $