TAG Weekly Telcon

11 Sep 2008


See also: IRC log


Stuart Williams, Jonathan Rees, T.V. Raman, Henry Thompson, Ashok Malhotra, Dan Connolly, Dave Orchard, Noah Mendelsohn
Noah Mendelsohn (partial), Dave Orchard (partial), Norm Walsh
Stuart Williams
Ashok Malhotra



No comments on agenda

Resolution: Minutes from Sep 4 approved http://www.w3.org/2008/09/04-tagmem-minutes

Next week call a risk. Regrets from Tim. Stuart may not be able to make it

HT: Use the time to read our documents

Cancel next week's meeting

<jar> +1 use the time to read

Next meeting f2f

Raman: If we are serious abt this, all TAG members should read the HTML spec

DanC: Please let's finish reading list and Agenda for f2f

binaryXML-30 (ISSUE-30)

<Zakim> ht, you wanted to acknowledge my EXI actions

HT: I will read these on the 'plane and make a recommendation on what we should do

DanC: Last, we said tell us how you are better than gzip

HT: That's where we are, the ball is bak in our court.

SKW: We will discuss this again at our FTF.

UrnsAndRegistries-50 (ISSUE-50)

<DanC> (though Dec sounds wierd... I thought our request was since Dec)

HT: I'm working on a new document. Shd have it ready middle on next week

<DanC> close action-167

<trackbot> ACTION-167 S to start a thread on non-DNS authority resolution on www-tag closed

DanC: What's happening with XRIs?

SKW: Summarizes situation

We have not had a formal proposal saying would you be happy with ...

SKW: We had a discussion on how the discussion was going

<Zakim> DanC, you wanted to check whether skw meant it when he said "prefix", since DNS names go least-significant-first

<DanC> does either Booth or Bradley advocate an actual prefix?

<Zakim> ht, you wanted to say there's one thing we will need to chase no matter what

<ht> Abstract Identifier document: http://wiki.oasis-open.org/xri/AbstractIdentifierArchitecture

<jar> saying "http://xri.*/*" are XRIS is same as saying "http://*/ark:*" are ARKs ...

SKW: That's not the proposal (to <jar> above)

<ht> And I think there is _some_ room to argue that both of these are OK, if not ideal

Self-Describing Web

Noah has incorporated feedback from Norm and SKW:

SKW: Norm and I would be supportive of publication

<Zakim> ht, you wanted to ask a question

ht: In a discussion with a student I realized ...

<ht> I believe the following: "FYN works iff every party to the story is a) publically accountable

<ht> and b) aware of the dependency of the FYN story on their part of it.

<ht> "

<DanC> I think you can follow-your-nose into policies and such that aren't world-readable

DanC: I would not say 'publically accounatable"

HT: The parties have to be publically accountable

SKW: The draft does not say this

HT: I would like to discuss this

SKW: Pl. send comment

DanC: I disagree for 3 reasons

<DanC> (I ran out after 2)

<DanC> (1) need not be world-readable

<DanC> (2) the URI for text/plain isn't actually critical path

<DanC> ... currently

<DanC> (though it's nice that the text/plain full URI is in an RFC)

HT: I will send mail on this

<Zakim> DanC, you wanted to think about whether RDFa is critical path: if we leave it aside, what's the audience/purpose? and to

DanC: How can we finish without RDFa story?
... I'm not sure story holds up

SKW: can we document missing link and encourage them to put it in place.

<skw> http://www.w3.org/2001/tag/2008/09/f2fkc-agenda

F2F Agenda

SKW: Talks about the f2f agenda. Thanks Raman for his help

DanC: I would like to negotiate the reading list now

<DanC> I hear from skw: urnsregs, binaryxml, html*,

<DanC> digest of ?

SKW: Should read binary XML specs, HTML spec, collected digest of refernces from Raman's thread

<DanC> self-describing web draft

<DanC> passwords in the clear

Self-describing Web, Password in Clear, Versioning

<DanC> versioning revision from david

Need two readers for Binary XML, HT is one.

URNsAndRegistries-50 ... HT writing paper. Due Tuesday. Shd be read by f2f

<DanC> * tim's bit

HT: We should all have read Tim's paper

<skw> also had an explicit request from David for Jar's formal treatment...

<DanC> "the document"... one document on versioning?

<DanC> DO nominates JAR's formalism

DaveO: What is new is Jonathan's formalism. Recommend people read this by f2f

<DanC> DO: key chapter is ch5

DaveO: Please review Chapter 5. That is new and is key

<DanC> HT nominates SVG and HTML thread from public-html... a dozen messages

HT: Read SVG and HTML thread. Read 10 msgs and get a feeling of the context

<DanC> TVR 2nds... long thread... read for motivations

<DanC> (looks like TVR's agenda input subsumes HT's suggestion to read a thread)

TVR: Read HTML spec with a view thru the structuring lens I proposed

JR: Is there a document that tells why W3C got involved in html5

<noah> Are you discussing reading list?

<jar> http://wiki.oasis-open.org/xri/AbstractIdentifierArchitecture

<jar> ?

DanC: I can point to formal mataerial but that's not what you want

<DanC> on mime types... a section of the html spec

<DanC> pwinc fri

<noah> Friday's OK if short, I think.

<DanC> (thanks; I was just gonna ask for irc convirmation)

Noah: Are we all supposed to read whole HTML spec?

<DanC> nm nominates thread on meeting goals

Noah: Please read thread on Tag Soup

HT: Norm is not coming to Kansas City

<skw> I think that the thread Noah referred to is based at: http://lists.w3.org/Archives/Member/tag/2008Aug/0019.html

DanC: I will send mail before EOD after editing agenda page

Possible topic GenericResources-53

Content negotiation and Abstract Documents

Not on agenda currently. You can lobby me.

TVR: Steve said he was pulling in my TPAC proposal

abbreviatedURI-56 (ISSUE-56)

SKW: Asks abt status of CURIE comments

<DanC> (anybody have a summary of the comment? the subject line was a generic "comments on X")

Noah: That's for responder to say

SKW: Summarizes comments

Editorial: QNames never intend as attribute values. Some discussion on this

<DanC> (pls promote that "main substantive comment" to the subject line)

SKW: Definition of XML Schema datatype

<ht> Please remember that we have already fed back on this point, see http://lists.w3.org/Archives/Public/www-html-editor/2008JanMar/0014.html

AM: Noah you had a comment on lack of clarity between CURIE and URI where there is ambiguity

Noah: I sent this as a personal comment. If no objection, I can add to my note

<jar> the whole point of safecurie was so that they can be put in uri contexts

<DanC> yes, now that I understand the comment, it seems to miss the point of safecuries

<noah> Well, it hijacks the use of [ in everyone's languages.

Raman: I'm uncomfortable with this. We need to allow new syntax in old contexts

jar: If there was no intention of extensing URI content there would be no SafeCURIEs

<jar> RDFa already would violate a prohibition on safecuries. It's too late to prohibit safecuries

HT: We should be careful abt distinguishing between CURIE's and SafeCURIES

<DanC> <ht> Please remember that we have already fed back on this point, see http://lists.w3.org/Archives/Public/www-html-editor/2008JanMar/0014.html

HT: We should not go back on that advice

TVR: The way Noah phrased it it sets a very high bar for new syntax

<jar> Two questions here! (1) CURIEs in URI contexts? (No.) (2) SafeCURIEs in URI contexts? (RDFa requires.)

<Zakim> noah, you wanted to say implying safecuries can be used in existing languages where URIs are expected hijacks the use of [ in those languages.

Noah: Explains his POV ... I should open my spec to other syntax

<jar> relative URIs can start with [, yes?

They should make clear that these things are not URis

DaveO: Supports Noah. CURIEs cannot be wedged into existing specifications

<jar> I repeat: There are two questions here! (1) CURIEs in URI contexts? (No.) (2) SafeCURIEs in URI contexts? (RDFa requires.)

<DanC> jar, does RDFa use <a href="[safecuri]">? I see deployment problems there.

<skw> http://www.w3.org/mid/48B810F4.60807@aptest.com

DaveO: Must specify how CURIEs and URI are disambiguated

<jar> no, but it allows safecuries in other uri contexts, I believe. Will check.

<DanC> ok. deployment considerations for a/@href are somewhat special

TVR: XSLT uses { } is attribute value templates. Example of use of a special character

<jar> ok, URIorSafeCURIE only occurs in attributes that are newly added by RDFa

<noah> I did propose text to Shane on 8/29:

<noah> <proposed>

<noah> CURIEs and safe-CURIEs map to IRIs, but neither a CURIE nor a safe-CURIE

<noah> <italic>is</italic> an IRI or URI. Accordingly, CURIEs and safe-CURIEs

<noah> MUST NOT be used as values for attributes that are specified to contain

<noah> only URIs, IRIs, URI-references, IRI-references, etc. Specifications for

<noah> particular attribute values or other content MAY be written to allow

<noah> either CURIEs or IRIs (or URIs, etc.). The specifications for such

<noah> languages MUST provide rules for disambiguition in situations where the

<noah> same string could be interpreted as either a CURIE or an IRI. One way to

<noah> do this is to require that all CURIEs be expressed as safe-CURIEs,

<noah> implying that all unbracketed strings are to be interpreted as IRIs.

<noah> </proposed>

TVR: I'm mostly OK with this.

<DanC> x:y

JAR: I'm bothered by saying "CURIES are not IRIs". There are strings that are both.

<DanC> noodling... "neither every CURIE nor every safe-CURIE <italic>is</italic> an IRI or URI"

Noah: I will put this in a note to the TAG list and people can comment

<noah> So, Stuart, what's the next step on the response.

SKW: Let's conclude on email.

<noah> SKW: Noah to redraft considering Stuart's proposal on intent of QNames and add 8/29 draft text on using CURIEs where URIs expected

SKW: DanC, any progress on 171

Dan: No.

<DanC> p.s. any hosting issues?

<DanC> hmm... decisions decisions...

<DanC> collect all preparation materials in one place in the agenda...

<DanC> or tuck them under the relevant items?

<DanC> I lean toward tucking, so far

<DanC> hmm... how to do a crawl-and-zip...?

Summary of Action Items

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.128 (CVS log)
$Date: 2008/09/17 10:21:11 $