TAG Weekly Teleconference
20 Sep 2004

Roll Call

<scribe> Scribe: Roy Fielding

<scribe> ScribeNick: RF

Date: 20 Sep 2004

Meeting records

Minutes 19 July

<DanC> 19July minutes http://lists.w3.org/Archives/Public/www-tag/2004Sep/0116.html

RESOLVED: Accept minutes 19 July

RESOLVED: Accept minutes 13 September

PaulC: will get minutes of last F2F done for next meeting

Accept this agenda

<DanC> (hmm... I have news on httpSubstrate; not sure whether it merits discussion.)

Next meeting

Next meet: 27 Sep, regrets: none, possible regrets: RF (in Basel)

Stuart: meet with QA WG on 27 Sep?

PaulC: This is to deal with last call comments?

DanC: thinks it will be the general topic of finding + last call

<Chris> best to set expectations about which comments are LC and which are for the finding

PaulC: want to set expectations right, have deliverable in hand before we talk

Stuart: we have: 1) set of comments on webarch from QA; 2) general comments on extensibility and versioning
... will make clear to QA that our priority is to address the webarch comments
... will work on that in next week's agenda
... volunteer to scribe next week?

Chris: volunteers to scribe for 27 Sep

<Roy> at risk due to Basel travel

Basel meeting update

<Chris> Roy, is it easier to get to the meeting/hotel from Zurich airport or Mulhouse airport?

<Chris> I can get flights to either

Stuart: F2F meeting planning for Basel

<Roy> Mulhouse is closer if you don't mind the bus or paying for taxi

<Roy> I am happy to invite non-TAG guests

TAG Vacancies

TimBL: in progress, nothing official to report yet

TAG charter

Stuart: no updates

Report on completed LC#2 Actions

<Stuart> http://www.w3.org/2001/tag/2004lc/lc-status-report.html

5. Definition of agent in the Web Architecture

<DanC> [DONE] PaulC: explain to a.p.meyer that we didn't mean any more than we said

PaulC: Meyer was disappointed but expressed acceptance

6. information resource

<DanC> stuart's reply http://lists.w3.org/Archives/Public/public-webarch-comments/2004JulSep/0062.html

Stuart: sent message, action item done

<Stuart> http://lists.w3.org/Archives/Public/public-webarch-comments/2004JulSep/0086.html

<DanC> stuart's substantive reply ^

Stuart: (reads proposal from mail to webarch-comments)

TimBL: doesn't help me a lot because of emphasis on accessibility
... did the TAG see Sandro's proposal?

DanC: last week, it was briefly mentioned

<Chris> That strikes me as a good feature

<Chris> something that implements an electronic protocol is a information resource and is testably 'on the web'

TimBL: web resource definition is a bit circular

<timbl> What strikes you as a good feature, Chris?

<Chris> the emphasis on whether you can access it

<Chris> in particular, its a testable statement

Stuart: Patrick didn't like information resource, suggested web resource because it better fits our definition
... discussion centered around literal interpretation of "information resource" rather than what was in webarch

TimBL: we are trying to get a common understanding for the concept, and find words to do that -- possibly change words later

Stuart: ack that TimBL is not happy with that resolution

<DanC> Sandro proposed text in http://lists.w3.org/Archives/Public/public-webarch-comments/2004JulSep/0057.html

TimBL: Sandro put a lot of effort into that

<Chris> seems a lot of thread to go through on this call

<Chris> I think conflating 'information resource' and 'on the web' is the source of the problem here

Roy: (made a lot of comments generally in support of Patrick's position, but I can't speak and type at the same time)

<timbl> Roy I think said he supported http://lists.w3.org/Archives/Public/public-webarch-comments/2004JulSep/0057.html

<timbl> sorry not that

<timbl> Patrick's message

<timbl> http://lists.w3.org/Archives/Public/public-webarch-comments/2004JulSep/0058.html

<Chris> **punt** it

<Stuart> http://lists.w3.org/Archives/Public/www-tag/2004Sep/0089.html

<DanC> punt it how? leave the doc as is, Chris? I could perhaps accept that.

<Chris> punt it as in, move on to do the other issues.

<DanC> i.e. postpone discussion until later, Chris? Well, I'd rather not do that unless/until somebody in particular takes the ball (i.e. an action)

<timbl> http://lists.w3.org/Archives/Public/www-tag/2004Sep/0046.html

<Roy> ACTION: Roy to digest Sandro's proposal and explain why it is unacceptable to him

Stuart: will post his suggested solution to Patrick's comments on www-tag

7. Comments on Web Arch WD - 2004-07-05

<DanC> let's continue NW's pile of actions in reply to Karl "Comments on Web Arch WD - 2004-07-05"

DanC: suggest continuing actions for NW [general agreement]

8. non-authoritative syntaxes for fragment identifiers

<DanC> Re: non-authoritative syntaxes for fragment identifiers from RF http://lists.w3.org/Archives/Public/public-webarch-comments/2004JulSep/0085.html hasn't produce a reply from the commenter

<Roy> I responded to commenter, no reply yet

11. AWWW, 20040816 release, sections 1 and 2

DanC: yet to do, GK indicates he doesn't feel strongly about them

14. resources/representations

<Stuart> http://lists.w3.org/Archives/Public/public-webarch-comments/2004JulSep/0067.html

<Chris> we can always point out important architectural consequences of particular specs

<Chris> we don't contradict the http spec, merely point out something in it

DanC: related to information resource, will leave pending

15. too positive on extensibility

<Stuart> http://lists.w3.org/Archives/Public/public-webarch-comments/2004JulSep/0068.html

<Roy> Commenter wishes us to include the negative trade-offs as well as positive for extensibility

Chris: agrees with commenter, will draft some text

<scribe> ACTION: Chris to draft text in response to "too positive on extensibility"

PaulC: will this be on agenda for QA-TAG interaction?

<DanC> (which finding?)

<Roy> extensibility?

<Chris> some text to record possible negatives; but text will still say that on balance its better to consider extensibility from the start

<Chris> PaulC: some useful stuff in the actual finding

16. what does it mean to 'take on meaning'

<Stuart> http://lists.w3.org/Archives/Public/public-webarch-comments/2004JulSep/0077.html

<Chris> "But this is a philosophical rathole that I think the document should try to avoid." yeah, what he said

TimBL: You can talk about the meaning of a word, as well as a sentence

DanC: reasonably happy with what we have

<Stuart> The whole quote is "[URI] is an agreement about how the Internet community allocates names and associates them with the resources they identify. URI Scheme specifications define the protocols by which scheme specific URI are associated with resources and take on meaning. "

TimBL: what about URIs that only identify a concept?
... suggests, "we agree with what you said, but can't find better words -- can you supply better words"

<DanC> ACTION: DanC to reply to djw re "take on meaning": yes, agree with your comment, think that's what webarch says. let us know if you think of something better

17. Use of "assign" for URI -> resource

<Stuart> http://lists.w3.org/Archives/Public/public-webarch-comments/2004JulSep/0078.html

17. comments from Larry Masinter

<Roy> I find Larry's comments compelling

DanC: would like to hear what the editor thinks

<Chris> I agree that pruning remaining 'assign' is a good idea

<Chris> "Resources have owners. URIs have users. The owners of resources arrange the resources so that URIs can be used to identify the resources and their related resources."

<Chris> seems very reasonable to me

TimBL: thinks that we added URI ownership because it was needed in several places. Are we going to replace it with new words, or remove the concept?
... need to find all uses in document and see if term needed

<DanC> (I don't know whether we need to speak of URI owners, Chris, but I know for a fact that we do.)

<DanC> e.g. "When a URI alias does become common currency, the URI owner should "

TimBL: architecture is set up so that the machine tells client what the "owner" says the URI means ... [scribe can't keep up]

Stuart: also issue about "owning" versus "renting"

Chris: owner of resource reveals how to construct URI according to scheme in a general sense, not in the URI sense

TimBL: describes ownership in terms of allocating names within a domain

<Roy> It is not the URI that is owned -- it is control of authoritative behavior at that URI

TimBL: DNS names are owned in the sense that ICANN and name delegations are bound by community agreement/standards

DanC: we do treat the concept in detail in webarch

PaulC: we are using ownership as a complex set of things known as "responsibility"

<Roy> URI specs use "authority" (perhaps too much)

TimBL: Larry seems to say that URIs have users, resources have owners

DanC: (quotes from webarch)

PaulC: "URI owner" is useful because it makes the good practice easier to read

Chris: Larry lists all of the cases in the document, do they all work?
... not "resource owner", because that would suggest changing the URI every day would not be harmful

DanC: inclined to leave this as unsatisfied

<scribe> ACTION: Stuart to lead another round of discussion on Larry's comment

PaulC: wonder what Norm's plans are regarding draft before F2F?

<scribe> ACTION: Stuart to find out what Norm's status is for F2F document

18. Comments from the QA WG on WebArch 2nd LC (extensibility)

<Stuart> http://lists.w3.org/Archives/Public/www-qa-wg/2004Sep/0020.html

PaulC: Ask QA if they think draft extensibility finding addresses some of their concerns?

ADJOURN