W3C

TAG telcon

28 Jun 2005

Agenda

See also: IRC log

Attendees

Present
Henry, Ed, Vincent, Noah, DanC, Roy, DOrchard, TimBL
Regrets
Chair
Vincent
Scribe
Ed

Contents


 

 

<scribe> Scribe: Ed

Meeting to order

Administrative topics

Approval of minutes from F2F ?

<DanC> (found http://www.w3.org/2005/06/15-tagmem-minutes ; marking it not world-readable)

Dan: where working minutes marked draft? or deleted?

Resolved: Cambridge f2f edited version have been approved

Approval of last teleconference?

Dan: they say draft

<DanC> (also set http://www.w3.org/2005/06/16-tagmem-minutes.html,access to member-only)

Ed would rather that they didnt say draft..

Vincent: its difficult to change because they're in the archives

Ed, I think its ok.

Dan: If you're going to mail them out, please don't say 'draft'.

Resolved: last teleconference minutes accepted.

:)

<DanC> note 4 July holiday in the U.S.

Topic. Next teleconference

regrets from Roy, Noah

Scribe for next week: Dave

Agenda for today

<noah> Please note that Noah has sent regrets for both July 5 and July 12. I expect to be back on the 19th (by which time it may well be my turn to scribe.)

Vincent: we may change the discussion around schemeProtocols-49 some, based on email.

no other additions/changes noted.

Vincent: Paul Cotton noted that we did not publish a Qtrly report since last march

<scribe> ACTION: Vincent to write something based on what was presented at the AC meeting earlier this month and will circulate to the mailing list. After comments he will publish this. [recorded in http://www.w3.org/2005/06/28-tagmem-minutes.html#action01]

<Zakim> DanC, you wanted to wonder if I'm supposed to take care of tag-announce

Dan: Tag-Announce; has anyone been sending anything to this?

Vincent: this is something we should do. We should post to this mailing list for any significant announcements etc.

<DanC> "TAG of minutes, meeting summaries, findings, new issues, resolved issues, and drafts of architecture documents"

Dan noted that we are missing minutes and http-range 14 to this list.

Invitations

more information on new subjects; Grid and web apps.

Noah proposed name from IBM

Ed proposed name from HP

Vincent; We have two canidates, both seem interesting. Should we invite them both?

and when should we do this?

Ed thinks both would be good

Dan thinks web apps charter may be more urgent

<DanC> "Proposal: we will have calls through 19th of July, then take a break with next call being 23rd of August."

<DanC> -- http://www.w3.org/2001/tag/2005/06/14-16-minutes.html#item07

Noah: we would give some guidance around what we're looking for
... what is grid, there appears to be some confusion (as an example)

Vincent: I'm not sure we're asking them anything formal.

Noah: Propose grid on the 19th, if possible
... lets us do web-apps sooner

Vincent will contact both parties and try and make arrangements

<scribe> ACTION: Vincent to organize the experts on the grid, if possible both on the 19th of July [recorded in http://www.w3.org/2005/06/28-tagmem-minutes.html#action02]

Vincent: We'll do web applications next week

Dan: the charter is available is our goal to review the charter? Should we be part of the membership review or..?

Timbl: it would be good to have the TAG internalize the architecture for web applications and discuss them.

Vincent: the discussion was not just about the charter, but also about longer term architectural implications

<DanC> > I see it was announced 23 May to the AC

<DanC> > http://lists.w3.org/Archives/Member/w3c-ac-members/2005AprJun/

<DanC> > 0041.html

<DanC> > --> http://www.w3.org/2004/07/webapps/webapps

<DanC> (I don't want to charter a W3C WG to do much design in the webapps space. I think the best bang-for-the-buck is QA. testing and refinement of existing designs)

Vincent: while hearing the charter discussion, I'm less and less convinced its not a good time to discuss the charter itself.

TimBL: We're in the phase of informally discussing thats all.

<Zakim> noah, you wanted to say that having open app models for the web is important

Noah: I see some items in the charter which give me some pause. On the other hand, expectations for what applications do are evolving and I think its critically important that we find that there were open architectures deployed.
... I think its important if this charter is close enough to the mark to make that happen.

Vincent: The plan was not to have an in-depth discussion today, just to get organized to have the discussion in the future.

<timbl> Questions which seem open: Should the interface be language specific to be more powerful and usable? If this is going to be ECMAscript why not say so? If not then say that. Or aim to have shared library between JS and Java?

Vincent: We'll try and have the discussion next week.

<noah> Noah also noted at least one more detailed concern: the draft charter seems to mandate a solution to client side storage. That seems like a deeply complex area. I have no problems with a group taking a look at it, but I'm concerned that we're committing to a deliverable before we know what if anything meets a practical need.

<scribe> ACTION: Vincent to invite Dean for the 5th of July to discuss web applications [recorded in http://www.w3.org/2005/06/28-tagmem-minutes.html#action03]

<timbl> Dean Jackson

<noah> Roy mentioned that security is important and is missing from the charter.

Names, namespaces and languages

<noah> +1 from Noah to Roy's concern on security.

Henry to present work.

http://lists.w3.org/Archives/Public/www-tag/2005Jun/att-0059/names.html

Henry: There are two candidate best practice notes at the bottom

1) As a language evolves, use new expanded names for new things, don't recycle old ones

2) Don't use the same expanded name for two different things of the same sort in different languages under your control

<DanC> (hmm... maybe better to finish this discussion thread 1st...)

concern around versioning and calling 'new things' new as opposed to a new version of the same thing.

<Zakim> DanC, you wanted to respond to the new/old thing...

<Zakim> noah, you wanted to suggest that maybe we need a broader definition of namespace

Noah: when you access a status code 200, that had better be an information resource.
... Namespaces are borderline, we haven't really said what they are. Norm said we should come down on the side of an information resource.

<noah> "A namespace is a set of terms and associated information, all qualified by the same URI. In the case of XML Instance documents, namespaces consist of the constructs specified in "Namespaces for XML", along with supporting information such as schemas, stylesheets, etc. In the case of XML Schema, a so-called "Target Namespace" consists of named type definitions, element declarations, attribute declarations, and so on."

Noah: proposed text.

<DanC> (I don't see the issue... in what way might a namespace *not* be an information resource?)

<Zakim> timbl, you wanted to suggest that a namespace is a set of URIs which share a common prefix, and a namespace document is the document which is identified by that URI

<DanC> (and why does "namespace" need a definition any more constrained than "information resource?")

timbl: we're getting a little tangled in this. You can talk about an html name space just being a set of names

TimBL: A namespace document is a document if you want a list of that.

<ht> Information about ways of using the namespace

<ht> == Information about using qualified names of the form < [the namespace URI], some local name>

<Zakim> ht, you wanted to disagree strongly

Henry: The specific question that the document i wrote came back with a specific negative. The namespace URI just allows you to have a list of expanded names.

<noah> Noah and Tim went back and forth: I think the net is that Tim is suggesting that the namespace URI actually identifies the namespace document or information resource. Noah is happy with that. Noah would have been concerned if the URI identified both the list of names as a resource and the namespace document as a resource.

Henry: We need not to talk so much about namespace documents but aound language documents

<ht> yes

<ht> yes to XHTML anyway -- RDF, like XML itself, is tricky because it's so meta

Dave: We need to think about the relationship of fragment identifiers and namespace documents

<noah> Noah is mainly pushing that we have a clean answer to the question: what is the information resource identified by the namespace URI, and why can we justify status code 200 returning RDDL with things like stylesheet references. Tim's proposal seemed to point a good direction for that.

<Zakim> TimBL, you wanted to say that if Henry wants to break apart the allocation of names in namespaces and the allocation of them to things they denote in different circumstances, then

<timbl> DO: http;//whatever/sdhfghjasdfgjgf#interfaceFfoo

<timbl> What happens if there is RDDL document there, which doesn't tell me what interfaceFoo is?

<DanC> "nothing in the doc says what interfaceFfoo is"? well, by reference, it seems to.

<Zakim> timbl, you wanted to say that if Henry wants to break apart the allocation of names in namespaces and the allocation of them to things they denote in different circumstances, then

<timbl> Seems that David is concerned that one won't know how to get from eth RDDL document to the other document which one actually needs.

<DanC> "WSDL has to give a property for use in RDDL before they're done". well, in a way, I suppose. But XML Schema did the RDDL stuff after-the-fact

<DanC> (the possibility was always there; namespace documents don't introduce it)

<DanC> (the possibility being: that meaning of a doc might include another by ref)

<noah> Dave. From: http://www.w3.org/TR/webarch/#frag-coneg "representation providers must not use content negotiation to serve representation formats that have inconsistent fragment identifier semantics. " Doesn't that cover your concern about WSDL and RDDL?

Henry: I meant that simply that its possible to mint a namespace and that there is an infenant set of names in that name space.
... one of the reasons we like http: URI's is because we know who's job it is to answer that

<noah> ...and then Noah notices that Dave isn't on IRC

<Zakim> DanC, you wanted to note that by the "minimalist" reading, "z" is in every namespace, e.g. <http://www.w3.org/1999/xhtml> and by another notion (that perhaps merits a new name) "z"

Dan: for purpose of discussion, I need another name for namespace

<ht> Indeed, that's what I was just saying to Timbl

<ht> HTML tells us that P is the name of something in that language

<ht> Yeah, I was going to use 'application'. . .

HT: I agree that you think the web is simple enough that the document element is good enough to tell you what you want.

<DanC> GRDDL, for example, doesn't define a document element.

<DanC> XSLT literals results do *exactly* what timbl says shouldn't happen.

timbl: if its in a given name space and you use any other name, then I feel the self describing nature breaks down if in anyplace in the doc itself it changes what type of document it is.

<timbl> XSLT literals do exactly what I say should happen iff you adot the XML fucntions architecture.

<timbl> (In other words, XHTML says functions are allowed, and XSTLT namespace defines what happens when yo umeet some of it in a foreign language)

<DanC> the XSLT spec says that you can start an XSLT transformation with <xhtml:html>

<ht> the XSLT spec says you can start a transformation with _any_ XML element

<DanC> http://www.w3.org/TR/xslt#literal-result-element

<DanC> (yes, DO's issue is one that has come up a number of times; is it in the current draft of whatever we've got on namespaceDocument-8? if not, anybody want to take the ball?)

<Zakim> noah, you wanted to suggest that Tim's distinction between namespace doc and namespace may point to a solution to Henry's concern

<ht> barenames are fine if there's only one sort of thing in your language/application/whatever

<ht> And that's the case for RDF

<ht> It's not OK for XML or Schema or WSDL

<DanC> noah, welcome to the xml-names thread from 1999, all about whether namespace names are actually pointers at all

<ht> This is a use/mention confusion, but I agree it needs to be clarified carefully

Henry: I'll update my document, if it fits into a finding that's great.

<DanC> yup. this is the key to extensibility too.

Dave: This has been very useful.

Vincent: lets move on

shemeProtocols-49

Noah sent a message right before the call.

<ht> When I write <xs:element ..../>, I'm _not_ using http://www.w3.org/2001/XMLSchema in anything _like_ the same way I'm using it when I write <xhtml:a href="http://www.w3.org/2001/XMLSchema">...</xhtml:a>

<DanC> ah... not xml-names thread, but xml-uri http://lists.w3.org/Archives/Public/xml-uri/

<ht> DanC, thanks for good memory and good pointers!

Noah, I'm working on refining the findings. Please provide feedback asap as i'm occupied for the next two weeks

namespaceDocument-8

Vincent: We'll return to this when Norm returns

Review action items;

Vincent: everyone to review from agenda.

Close meeting

<noah> When Henry wrote above: "When I write <xs:element .../> I'm not" that was exactly the point I was making on the call. I'm trying to rationalize the use of the same URI for both.

<noah> Tim says the URI identifies a description document for the NS.

Summary of Action Items

[NEW] ACTION: Vincent to invite Dean for the 5th of July to discuss web applications [recorded in http://www.w3.org/2005/06/28-tagmem-minutes.html#action03]
[NEW] ACTION: Vincent to organize the experts on the grid, if possible both on the 19th of July [recorded in http://www.w3.org/2005/06/28-tagmem-minutes.html#action02]
[NEW] ACTION: Vincent to write something based on what was presented at the AC meeting earlier this month and will circulate to the mailing list. After comments he will publish this. [recorded in http://www.w3.org/2005/06/28-tagmem-minutes.html#action01]
 
[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.126 (CVS log)
$Date: 2005/07/11 16:32:08 $