W3C

TAG
26 Jul 2004

Attendees

Present
Norm, DanC, Chris, PaulCotton, Roy
Regrets
IJ, SW, TimBL
Chair
NormW
Scribe
DanC

Contents/Agenda

see also: Agenda, IRC log


Administrative: roll call, review agenda, record

review of 19July minutes postponed until they become available

agenda seems OK

next week: NW regrets. PC regrets.

we don't expect SW is available 2 Aug

RESOLVED: to cancel 2 Aug telcon. next meeting: ftf 9-11 Aug

Ottawa meeting update

<Norm> http://www.w3.org/2001/tag/2004/08/09-11-tag.html

NW: toward having a good meeting in Basel, "Our goal for this face-to-face is to leave with a technically complete second Last Call working draft."
... aiming for publication [n]th of [month?] last call for about a month.

<Norm> one month

ChrisL: have we started negotiating with peer groups about LC schedule?

NW: no; haven't started

PC notes I18N WG's recent inquiry about LC schedules

ACTION NW: respond to I18N's inquiry about LC schedules, noting TAG's evolving plans

PC: we're hoping to be able to edit the webarch doc during the meeting

NW reviews daily schedule in http://www.w3.org/2001/tag/2004/08/09-11-tag.html

NW: we're considering inviting DaveO to participate by phone...

PC: I was in contact with DaveO; he's considering it among various obligations
... monday might fit his schedule better

NW: I'm open to monday if others are

DC: likewise

NW: I'll follow up.

PC: I expect DaveO to reply to your earlier message

PC asked about possibility of remote participation by IJ. DanC was thinking we encouraged him to focus on other things.

thanks Paul and NW for preparing the agenda.

Action Item List

Action NW: 2004/07/12: Write XMLChunk-44 as a finding. continues

<Zakim> DanC, you wanted to speak to XML Schema action

Action TBL/RF: 2004/05/13 Write up a summary position to close httpRange-14, text for document (need to reschedule httpRange-14 when TBL available-single issue telcon? guest?). CONTINUES.

DanC: I asked the XML Schema WG for telcon time; haven't heard back yet...

PaulC: shall I call the chair?

PaulC is excused to call the XML Schema WG chair for a few minutes...

Web Architecture Document Last Call

NW: let's continue sorting into OBE, LC-critical, open

-- nottingham1: Second bullet doesn't make sense 1.2.1. Orthogonal Specifications

NW: doesn't look OBE
... relevant text is still there, though moved

<Chris> I agree thatperformance is the reason in practice (parsing all content to look for headers)

PaulC reached Ezell, who has now seen the request and intends to answer presently, after consulting some XML Schema WG members.

NW: I'd like to be available Thu, but I see that I'm not. so I still prefer Fri.

<Norm> http://www.w3.org/2001/tag/2003/lc1209/issues.html?view=wg&closed=1&expert=1&editorial=1&clarification=1&stateAgreed=1&stateDeclined=1&stateSubsumed=1

<Norm> http://www.w3.org/TR/webarch/#orthogonal-specs

<Norm> perhaps

CL: suggest open. webarch makes a good point here.
... commentor says "it's not deployed because of performance"; perhaps so, but it's also a problem w.r.t. architecture

NW: so... nottingham1 open?
... so... nottingham1 open.

<Roy> oops, was trying to say it was not a bad idea -- there is a lot more history involved

roy, is that re nottingham1?

<Roy> yes

<Norm> You've fallen off the phone, will you be able to come back, Roy?

do you want the TAG to discuss nottingham1 further?

<Roy> later

we can stick it in LC-critical for now if you like.

i.e. schedule it for discussion later

<Chris> we are saying that even if perfrmance was great, its still a level-breaking architecture problem

<Roy> yes

ok, nottingham1 is LC-critical

-- klyne7 Use other schema than mailto as example

ACTION NW: take klyne7 as editorial.

-- klyne9: Add stronger language on not permitting unregistered URI schemes

CL: yeah... "is discouraged" isn't clear enough. "should not"

ACTION NW: treat klyne9 as editorial

PC: yeah, that text is still there.

-- klyne12: Proposal to drop paragraph on inconsistent frag ids

http://www.w3.org/TR/webarch/#fragid

http://www.w3.org/TR/webarch/#media-type-fragid

NW: odd; greek letters are still in 3.3.1. CL: yes, see proposed text from my action. NW: good!
... we've re-written this; it's now in 3.3.2

(CL, you're welcome to write him individually)

NW: klyne9 is OBE.

<Chris> yeah okay

NW: klyne12 is OBE. [rather than klyne9]

-- klyne17: Worth pointing out value of RDF descriptions depends on URI persistence?

CL: commentor not sure or something... NW: open, at least; we've re-written some.

NW: klyne17 open.

-- klyne20: Say something about relationship between Hypertext Web and Semantic Web?

NW: Ian 8Jun rev seems to deal with this. DC: yes, 4.6.
... klyne20 OBE.

-- klyne21: Add statement about scalability concerns

CL: fair point; hmm... I have an action

NW: klyne21 is LC-critical, to review CL's action

ACTION CL: Draft text to explain that there's a tradeoff in this situation. continues from 14 May 2004

-- klyne25 klyne25: Add reference to RFC3117, section 5.1?

"On the Design of Application Protocols" http://www.ietf.org/rfc/rfc3117.txt

sec 5.1 Framing and Encoding

DC: I'm interested to look at it

NW: klyne25 LC-critical.

ACTION DanC: report on study of RFC3117, section 5.1

PC: note BXXP is in the same design space as SOAP... CL: yes, there are probably lots of things written about "Why we did X with XML".

<Chris> The pain of recreating this social infrastructure

<Chris> far outweighs any benefits of devising a new representation. So, if

<Chris> the "make" option is too expensive, is there something else we can

<Chris> "buy" besides XML? Well, there's ASN.1/BER (just kidding).

-- manola17: "Agent" that includes "people" source of confusion

DC: I think this is OBE

NW: manola17 is OBE.

-- manola27: Provide examples of mistaken attempts to restrict URI usage

CL: yes, fair point... e.g. "we assume HTTP" in a format spec would be bad.

NW: manola27 is LC-critical

ACTION CL: draft example ala manola27: Provide examples of mistaken attempts to restrict URI usage

-- i18nwg5: Discussion of content-type header hint

PC: looks like nottingham1

DC/scribe: it's LC-critical.

NW: i18nwg5 is LC-critical, like nottingham1

-- i18nwg8

PC: looks worth discussion. CL: yup

RF: Ian's dealt with this, yes?

PC: yes, but let's look again.

NW: yes, let's look again
... i18nwg8 is LC-critical

-- i18nwg16: Good practice on URI opacity impossible to follow for humans.

NW: we've changed to "SHOULD NOT"... OBE?

DC: either way...

NW: i18nwg16 is OBE

-- i18nwg19: text/foo+xml considered useless?

CL notes recent Internet Draft relevant to this.

scribe: deprecates this.

<Chris> rfc3023 revision: http://lists.w3.org/Archives/Public/www-tag/2004Jul/0016.html

DC: worth reflecting in webarch? NW: I think we do already. OBE.

NW: i18nwg19 is OBE

<Chris> http://www.ietf.org/internet-drafts/draft-murata-kohn-lilley-xml-00.txt

<Chris> Major differences from [RFC3023] are deprecation of text/xml and

<Chris> text/xml-external-parsed-entity, the addition of XPointer and XML

<Chris> Base as fragment identifiers and base URIs, respectively.

-- i18nwg20

NW: hmm... same slug as 19...

CL: I can see how readers could come to wrong conclusions...

ACTION CL: propose text based on i18nwg20

NW: i18nwg20 is LC-critical

-- rosenberg3: Reuse appropriate URI schemes (and protocols)

"On the use of HTTP as a Substrate" http://rfc.net/rfc3205.html

NW: we have an issue on that...

<Roy> /me I'll check the status of 3205 at next week's IETF

<Chris> http://rfc.net/rfc3205.html

<Chris> Says its a BCP

DC: ah; IJ did this.

NW: rosenberg3 is OBE

<Chris> http://rfc.net/rfc3688.html

ACTION NW: incorporate reference to RFC 3688 per rosenberg

NW: and perhaps bump httpSubstrate up in priority for ftf discussion

<Chris> If the registrant wishes to

<Chris> have a URI assigned, then a URN of the form

<Chris> urn:ietf:params:xml:<class>:<id>

<Chris> will be assigned where <class> is the type of the document being

<Chris> registered (see below). <id> is a unique id generated by the IANA

<Chris> based on any means the IANA deems necessary to maintain uniqueness

<Chris> and persistence.

-- rosenberg5: Proposed reference to IANA registry for namespaces and RFC 3688

NW: rosenberg5 is LC-critical. [cf action above]

--------- TRIAGE DONE! -------------

-- schema12: [3.6.1] [3.6.1] Good practice: Available representation. Too preferential to dereferencable URIs

DC: note telcon negotiations in progress.

ADJOURN.


Minutes formatted by David Booth's scribe.perl 1.81 (CVS log)
$Date: 2004/06/08 14:14:43 $