See also: IRC log
Approve minutes from 18 Dec?
Next telcon: 15 Jan 2009
Ashok to scribe.
Stuart: Ashok, you said you had progress...
Ashok: Yes. I've started to look at it and I've been thinking about it. I have a question.
Ashok: We're thinking about recommending two different ways to access metadata: via a Link: header and via site data.
... The question is, should these two be consistent?
... Accessing through site data gives you a metadata structure with one element for each type of metadata.
DanC: The way you find that is through a well-known URL.
Ashok: The question is, should those entries be consistent (parallel with) the entries in the Link: header.
... Right now they're slightly different.
TimBL: They're different in that they contain different things. Presumably the data you get back from the Link: header can ...
Ashok: I'm talking about what's in the link header and what's in the metadata structure. One has a couple of extra attributes.
TimBL: One is the HTML link rel=meta, the other is the HTTP Link: header.
... I still think you're misunderstanding each other.
... The two mechanisms are a well known URI or the Link: HTTP header.
TimBL: Oh, I thought you were talking about the HTML link header.
Ashok: No, I'm talking about what you get back from the well-known URI.
<noah> FWIW, I briefly had the same misunderstanding as Tim
<Stuart> FYI: Ashok is reviewing: http://tools.ietf.org/html/draft-nottingham-site-meta-00
TimBL: First of all, I think it's clear that they must not be inconsistent. Inconsistency is an error.
... But after that, it's not clear what kind of information you'd expect to find in each of those locations.
... Sometimes I think the HTTP link header might point me to the well-known URI.
Ashok: But you're ok if one produces a subset of the other?
<Zakim> Stuart, you wanted to remark that: 1) the subject of the assertions is less clear via /site-meta and 2) explicit in Link header.
Stuart: Basically there are two mechanisms and you might get different things. The format in the site-meta reuses some of the link relationships, it's unclear what the subject of any particular relationship is.
... But with the Link: header, you are explicitly getting information about a specific subject.
... The question of consistency is contingent on the question of "consistency to what?"
... At least some in this community, want to find information about mailto: URIs.
... So they want to put information int he site-meta file about more things than just web pages.
<Zakim> ht, you wanted to remind us of JAR's doc't and the connections it draws: http://www.w3.org/2001/tag/doc/more-uniform-access.html
Henry: At the f2f, Jonathan presented a document that points towards the point at which these two approaches converge.
Henry: The one I like, and the one I'd like to remind people of, is that one of the things you might put in your site-meta document is something which says
... In order to get metadata about a particular document on my site, perform the following transformation on its URI and fetch that on.
... It's the information that you get from that onee that you'd like to make consistency statements about wrt what you get from the link header.
<DanC> (issue-36 siteData and issue-57 HttpRedirections aka uniform access to metadata are quite closely connected; I sorta knew that, but AM's review is making it even more clear.)
Stuart: Ashok, did you get the feedback you needed?
Ashok: I think so.
Some discussion of where feedback should go.
<DanC> action-202 due next week
<trackbot> ACTION-202 Review http://tools.ietf.org/html/draft-nottingham-site-meta-00 due date now next week
Stuart: So action-202 is open, not pending review.
DanC: It defaults to next week.
Ashok: Ok, next week should be fine.
Stuart: Jonathan discharged an action, there was some discussion and I wondered if that warranted discussion here.
Noah: To set the context more broadly, we had an extensive discussion at the f2f in Cambridge. I think our conclusion was that we'd like to get it published before the TAG gets new members.
... There were a fair number of suggestions made, some of which seemed to garner consensus, others not so much.
... What I did in the last couple of days was go through the minutes and notes, integrating as much as I could do.
... I didn't have a chance to identify which issues I thought I resolved.
... We could just note that there's a new draft and schedule discussion for next week. If people prefer, I have some notes on the extent to which I addressed the issues.
... One of the things that needed attention was JARs suggestions for removing "semantic web". Dan was ambivalent; Ashok thought it was not helpful.
... On balance, I kind of like the way it is. But I'll adopt JAR's text if that's the direction we want to go.
<Zakim> ht, you wanted to support JAR's suggestions
Henry: I don't want to settle this in Jonathan's absence, in particular because I think they are good suggestions, but I think we should discuss the background to the changes when Jonathan is avialable to take part.
... The basic question isn't should we make those edits, but do we agree with the reasons that caused him to suggest them.
... I'd be at best a poor proxy for Jonathan to make that argument.
Noah: We should establish a direction and then I'll go do it.
Some discussion of Henry's ability to collaborate on two of the examples
Noah: Would an 8-10 day target work?
Stuart: TimBL, I wanted to ask if you'd had a chance to look at Jonathan's feedback.
TimBL: No. I'm prepared in principle to let go of the words "Semantic Web". I think that's more of a marketing question.
... But we should wait until Jonathan is here to discuss the reasons.
Henry: The reasons aren't to remove the semantic web, but to make the discussion more clear
<Zakim> DanC, you wanted to poll
DanC: Poll: who wants to do the edits?
In favor: Henry
Opposed: Noah, Ashok, Dave, Dan
Some discussion of the time constraints
<timbl> I am happy to leave this to the group wihtout me
Stuart: I'd like to wait a week and see if we can resolve this next week.
Noah: I don't think we can vote to publish next week, but I'd like to get resolution on whether or not we accept Jonathan's proposal next week.
... It's going to take some time to integrate the changes.
More scheduling discussions
Stuart: In the next week or two, I expect election results and we may have observers as well.
Noah: My personal preference is to set the bar pretty high wrt bringing newcomers up to speed because we're almost done.
<DanC> action-207 due next week
<trackbot> ACTION-207 Try to wordsmith to get rid of 'Semantic Web' and submit for review due date now next week
<DanC> (ht, http://www.w3.org/2001/tag/group/track/issues/51 doesn't show you in the critical path on this... ah... ok; it's accurate)
<Zakim> ht2, you wanted to go to specific points
Henry: I think most of the changes Noah made are excellent. I think they address most of the issues we raised.
... I have one substantive point which I think we made. The second good practice in section 5 says "RDFa should be used..."
... The next one says "GRDDL should be used..."
... I thought we weren't happy with those because it's too easy to read them as "If your HTML doesn't have RDFa, you're not with the program", ditto XML/GRDDL.
... I think we wish to say, if you wish to use RDF in conjunction with HTML, use RDFa, ditto XML/GRDDL.
Noah: I think you're making a good point that wasn't recorded in the minutes in the minutes of the f2f.
<DanC> (I understood noah saying "if the group wants" to be an implicit poll)
TimBL: We need another verb in capitals which means "it's good" rather than "you should"
<DanC> (ht, can you suggest alternate wording on the fly?)
<Stuart> Suggets reword in form: "In order to do X use Y"
Henry: But English is ambiguous. It can be read more than one way.
<DanC> "Authors of HTML document containing data should use RDFa to ..."
Some suggestion of adding "is good to". DanC objects on the basis that we've trained our readers to expect SHOULD.
<DanC> +1 "In order to..."
<ht> "Integrating XML documents into the self-describing Semantic Web is best done with RDFa"
<noah> RDFa can be a valuable tool for making information conveyed in HTML self-describing.
<ht> "Integrating XML documents into the self-describing Semantic Web is best done with GRDDL"
<DanC> -1 cuz GRDDL is just as good as RDFa
<Stuart> "In order to integrate XML documents into the self-describing Semantic Web it is good to use GRDDL"
<DanC> - "Integrating XML documents into the self-describing Semantic Web should be done with GRDDL"
Henry: I prefer Dan's suggestion to what's in the document.
Noah: I could live with it.
<DanC> (I prefer to settle on words for boxed thingies)
<timbl> I note danc's one incldues 'semantic web' so would be achnaged by jonathan's edits
<DanC> dancs and ht's both, and noahs current draft
<Stuart> "It is good to use RDFa to make information conveyed in HTML self-describing.
<ht> "To make information conveyed in HTML self-describing, use RDFa"
<noah> "use" seems a bit too unambiguous. Seems to discourage alternatives.
<DanC> yes, we're here to STANDARDIZE, which is to discourage alternatives
<ht> Representations provided directly in RDF, or those for which automated means can be used to discover corresponding RDF, contribute to the self-describing Semantic Web.
<DanC> yes, I like "To make information conveyed in HTML self-describing, use RDFa"
<noah> So, Dan, using GRDDL with XHTML is bad?
<DanC> but it's not as good as RDFa
<DanC> in some ways
<DanC> -1 "are a good thing"
<noah> I can live with that.
Scribe observes much wordsmithing.
<Zakim> DanC, you wanted to reiterate ht's concise "To make information conveyed in HTML self-describing, use RDFa"
TimBL: I like that.
<ht> "To make information conveyed in XML self-describing, use GRDDL"
Ashok: That's good.
Norm: Those work for me.
<noah> To integrate HTML information into the self describing semantic web, use RDFa
Stuart: Both of those work for me.
DanC: No preference.
Noah: I have a slight preference for the longer one I typed, but it's slight.
Stuart: I actually don't, but it somewhat depends on the outcome of Jonathan's message.
TimBL: I prefer the shorter one.
Henry: I prefer the longer one.
... The shorter version suggests that HTML w/o these things isn't self describing and that's not what we mean.
<ht> "To integrate information conveyed in XML into the self-describing semantic web, use GRDDL"
<DanC> (just stop with "and I'm the editor, so tie goes to me" ;-)
<noah> That's not my view of the editor's role, but I have the guidance I need.
Consensus to use the longer versions.
<DanC> well, there's a short pause-for-appeal that follows naturally
<noah> ACTION: Noah to change two good practice notes in sections 5.1 and 5.2 of Self-describing Web [recorded in http://www.w3.org/2009/01/08-tagmem-irc]
<trackbot> Created ACTION-212 - Change two good practice notes in sections 5.1 and 5.2 of Self-describing Web [on Noah Mendelsohn - due 2009-01-15].
Stuart: Any more on the Self-Describing Web before we move on?
DanC: Yes, I took an action in parallel to consider how it impacts "x-" tokens. I hope to have contributions soon.
<trackbot> ACTION-211 -- Dan Connolly to draft discussion of x- tokens and squatting for self-describing web finding, if review shows it's not already there -- due 2009-01-14 -- OPEN
DanC: At our last meeting, TimBL had some sort of news that got captured with no action.
<trackbot> ACTION-24 -- Tim Berners-Lee to clarify http://www.w3.org/2003/04/iri , perhaps by using N3 -- due 2009-01-13 -- OPEN
<DanC> recalling from http://www.w3.org/2001/tag/2008/12/11-minutes.html#item05
TimBL: The issue has to do with IRIs that have been percent-escaped using non-UTF-8 characters.
Some discussion of how/where the IRI/URI conversion took place.
<DanC> (tim's action-24 is actually quite relevant... he tries to clarify this using math)
TimBL: I think a TAG document should make it clear that wherever you can you should display the Unicode version. One thing is to say you should allow Unicode within the href inside HTML. But a URI field should be displayed in UTF-8.
... That doesn't work with that set of legacy URIs.
<DanC> ACTION-24 on Tim Berners-Lee to clarify http://www.w3.org/2003/04/iri , perhaps by using N3 - due 2009-01-13, open
DanC: TimBL, you had a math document where you tried to explain this. Do you remember that?
... I think someone else should take this over and finish it up.
<DanC> (the canonical example is the largest chinese search site)
Henry: These existing URIs which have %-encoded bytes in the query string are intrinsically problematic because when you see one, you don't know how to interpret those bytes.
... You can only interpret them if you know something in addition. You must know the character encoding of the document whose form the submit button created that URI.
<timbl> no non on
Henry: If I send you that URI, how can you know what it means?
<Stuart> This is a problem for UA that have to make up the query strings from form entries - yes?
DanC: In general, I can take any arbitrary byte sequence, pickle it into a URI and send it to you.
<DanC> "pickle" meaning %xx-escape it
Henry tries to reframe the question
Henry: If you send me that URI and I do a get from it, the question is, how does the server know what character encoding to interpret those bytes in.
DanC: It doesn't necessarily interpret them as an encoded character sequence. It might be a very small image.
TimBL: The server has a way of interpreting those bytes. The form is made by someone who has a secret agreement with the server who knows what mapping to make.
Stuart: Where are we left here in terms of actions and future work?
<DanC> action-24 due 1 Feb 2009
<trackbot> ACTION-24 clarify http://www.w3.org/2003/04/iri , perhaps by using N3 due date now 1 Feb 2009
<noah> Sorry I had to drop off for a second.
DanC: I'll put Tim's action to sleep until the new TAG members are official present.
<Zakim> ht, you wanted to ask DanC where he has gotten to with trying to reconstruct LMM's proposal
<trackbot> ACTION-188 -- Dan Connolly to investigate the URL/IRI/Larry Masinter possible resolution of the URL/HTML5 issue. -- due 2009-01-13 -- OPEN
<DanC> last change was Wed Dec 17 18:25:26 2008 -0600 (3 weeks ago)
Norm: I've re-engaged with Ted and hope to make some progress.
<DanC> action-205 due next week
<trackbot> ACTION-205 S to begin responding to Marcos asking the question: Why does the spec not say "A conforming spec MUST recommend a hierarchical adressing schems that can be used to address the individual resources within a widget resource from within a config doc, widget, or other constituent of the same widget pkg." due date now next week
<DanC> feel free to set a different due date in the future, h t
Stuart: On 206, I have raised the question with the group. I got no response, I've offlist pinged some folks in an attempt to get a reply. Private response suggests public response will be coming.
<DanC> action-206 due next week
<trackbot> ACTION-206 Query WebApps whether external access to widgets via URIs is ever expected due date now next week
Stuart: I communicated a tentative schedule to the candidates. I don't think we can make progress until we have election results as one of the candidates did push back.
<DanC> PROPOSED: to 3-5 Mar 2009 SFO, contingent on consultation of new TAG members 1 Feb
<DanC> PROPOSED: to meet 3-5 Mar 2009 SFO, contingent on consultation of new TAG members 1 Feb
<ht> HST is happy with DanC's proposal
<dorchard> Seems incredibly obvious that we should have this next week.
<DanC> tim's not here next week
<DanC> you want to make a ftf decision without tim, dorchard ?
Stuart: Proposed: to meet 3-5 Mar 2009 SFO, contingent on consultation with new TAG members 1 Feb
<dorchard> I'm here.
No objections, no abstentions.
RESOLUTION: The TAG will meet 3-5 Mar 2009 SFO, contingent on consultation of new TAG members
<DanC> action-197 due 1 Feb 2009
<trackbot> ACTION-197 Relay 3-5 Mar 2009 SFO proposal to TAG candidates and collect feedback due date now 1 Feb 2009
Stuart: We're adjourned.