Draft minutes of TAG conference of 6th August 2009

Are available at:


and below, as plain text.


- johnk

- DRAFT - 
TAG Weekly
06 Aug 2009

Agenda: http://www.w3.org/2001/tag/2009/08/06-agenda
    Larry Masinter, John Kemp, Noah Mendelsohn, Henry Thompson, Dan Connolly
    Ashok Malhotra, Jonathan Rees, Tim Berners-Lee


  Topics: http://www.w3.org/2001/tag/2009/08/06-minutes.html#agenda

    Convene: http://www.w3.org/2001/tag/2009/08/06-minutes.html#item01
    Minutes approval: http://www.w3.org/2001/tag/2009/08/06-minutes.html#item02
    Admin: http://www.w3.org/2001/tag/2009/08/06-minutes.html#item03
    HTML: http://www.w3.org/2001/tag/2009/08/06-minutes.html#item04
    Architecture of the Web of Applications: http://www.w3.org/2001/tag/2009/08/06-minutes.html#item05
    Metadata: http://www.w3.org/2001/tag/2009/08/06-minutes.html#item06
    Semantics of "resources" in RFC 3986 and RFC 2616: http://www.w3.org/2001/tag/2009/08/06-minutes.html#item07
    contentTypeOverride-24: http://www.w3.org/2001/tag/2009/08/06-minutes.html#item08

  Summary of Action Items: http://www.w3.org/2001/tag/2009/08/06-minutes.html#ActionSummary

<scribe> Scribe: johnk 

Date: 06 August 2009 


NM: we will have a call next week, LMM to scribe 
Minutes approval

DC: will approve 

NM: opposed? 

JK at risk for next week's call 

NM: suggestion in agenda that we co-locate Spring TAG meeting with IETF in Anaheim 

LMM: that was the third of three suggestion to improve coordination between TAG and IETF 

i) invite IETF to a call 

ii) invite IETF notables to TPAC 

iii) joint TAG/IESG/IAB meeting 


^ Larry's email 

NM: would you be willing to put forward in more detail suggestions about coordination between IETF/W#C before TPAC? 

LMM: more overlap between IETF and W3C now than in past, so relationship is very important 
... arch work in IETF group should be coordinated with arch work in W3C 

NM: can you put in email ways in which I could get started on liaison with relevant IETF people/groups? 


NM: (reviews agenda) 
... (discusses item 7,8 on agenda) 

NM: we agreed to read the specification 

<masinter> pointer to assignments? 

<masinter> would be useful to turn these into action items 

NM: does anyone have problems with their assignments? 

DC: I may have trouble meeting the goal 

<DanC> ACTION: dan read HTML spec parts from http://lists.w3.org/Archives/Member/tag/2009Jul/0026.html due 30 Aug 

<trackbot> Created ACTION-293 - Read HTML spec parts from http://lists.w3.org/Archives/Member/tag/2009Jul/0026.html due 30 Aug [on Dan Connolly - due 2009-08-13]. 

<DanC> action-293 due 30 aug 

<trackbot> ACTION-293 Read HTML spec parts from http://lists.w3.org/Archives/Member/tag/2009Jul/0026.html due 30 Aug due date now 30 aug 

NM: (to all) let me know if this isn't going to work for you 

<masinter> ACTION: Larry read HTML spec parts from http://lists.w3.org/Archives/Member/tag/2009Jul/0026.html due 30 Aug 

<trackbot> Created ACTION-294 - Read HTML spec parts from http://lists.w3.org/Archives/Member/tag/2009Jul/0026.html due 30 Aug [on Larry Masinter - due 2009-08-13]. 

NM: close action 288 on Henry to divvy up sections of HTML spec. 

(and does so via web interface) 

LMM: proposes that we look at comments made by an outsider to the HTML5 process 

NM: link? 

LMM: working on it 
... spent time working on the versioning document, and felt as if I made a breakthrough in my thinking 
... part of the trouble we've had is about the confusion between language as specified vs. language as spoken 
... role of version indicators is in... [scribe missed the last part] 
... every spec specifies at least two languages - an appropriate, correct conforming utterance, and what a receiver ought to do to make a correct interpretation 
... is indicator used to indicate the implemented language, or the specified version? 
... given that background, role of inline version indicator is to show which conservative language was intended 
... receiver might ignore it, but indicator is useful in building a _conservative_ receiver, and also in case where liberal receiver might want to distinguish between incompatible language variants 

<Zakim> noah, you wanted to talk about two views of version identifiers 

NM: what problem are we solving with version indicators? 
... point you're making is interesting, robustness principle is interesting but not the common case 
... common case is when receiver and producer expect the same language variant 
... author may know that element means the same thing in multiple versions of the language 
... references his TAG blog entry and examples given there 

<Zakim> ht, you wanted to concur, but qualify the assertion about authors and VIs 

HT: when author puts an indicator in, they _mean_ the conservative interpretation 
... clear that all authors have not thought that through 

NM: what do you mean by conservative? 

HT: [scribe missed the answer] 

LMM: W3C is in the business of documenting what the specified language is 
... every standard has the specified language, and then specifications of implementations 

NM: not necessarily asymmetry between producer and consumer 
... producer will know what a consumer will do under all circumstances to point of when they agree there is an error condition 

<Zakim> ht2, you wanted to ask about walled gardens (again) 

LMM: HTML5 gives implementation advice as well as specifying the language 

HT: most pushback against version indicators has been that they lead to walled gardens 

LMM: one argument is that DOCTYPE is useless - under what situations would DOCTYPE be useful? 

<Zakim> noah, you wanted to talk about terminology 

LMM: [scribe: phone hiccups giving me fits today] 

NM: suggest that both language is a set of text made by the producer, and language is also something to do with the interpretation 

<masinter> I've been trying to be clear about "language" that doesn't agree with what NM is saying 

NM: what we tell authors to write is clearly the language. I would claim that there is a second language, which is what is interpreted by the consumer 

<masinter> and "Language as a set of strings" isn't a good definition 

<Zakim> DanC, you wanted to suggest putting DOCTYPE aside in favor of an attribute 

NM: call one the "producer language", the other the "consumer language" 

DC: DOCTYPE has issues in rendering 

NM: that doesn't make it a different language? 

DC: you will get counter-arguments which are orthogonal to versioning 

NM: why's that - quirks-mode rendering gives essentially a second language? 

LMM: suggestion is that each time someone issues a specification document, they rev the version indicator 

NM: 150 versions of a spec a year, which version(s) should I put in my instances? 

LMM: consumer will mostly ignore then anyway 

NM: then why put it there at all? 

LMM: wanted to redefine language as "not a set of strings" 

NM: undoes work we did in TAG 

DC: we said "set of strings and meanings" 

LMM: I don't think it undoes what we've already done 
... distinction between things we call languages is really important 

NM: interpreted language is a language 
Architecture of the Web of Applications

NM: geopriv vs. Geolocation? 

LMM: Is the TAG going to be proactive about reviewing the geo-location work? 
... if so, we should invite members of the geopriv WG 

NM: what is the timing on this? 
... last call 
... anyone have suggestions here? 

DC: inviting chairs/editors from both committees might be interesting 

NM: what is the substantive issue here 

LMM: IETF has a long history + documentation about carrying privacy policy with instances 

<DanC> one pretty representative technical thread: http://lists.w3.org/Archives/Public/public-geolocation/2009Jun/thread.html#msg98 

LMM: current W3C activity provides no such in-band explicit mechanism 

NM: would all agree on this as a summary of the issue? 

DC: should privacy indications from UA or provider of the location information? 

NM: is this the issue? 

[scribe break due to person at door] 

NM: next steps though? 

what actually is the _architectural_ issue? 

LMM: propose that we tell geo-location people we will review to discover whether IETF group got a sufficient response 
... fine if we simply resolve to review the geo-location spec. 

NM: don't see the value at the moment in making such a resolution 
... I am happy to facilitate liaison between TAG and IETF on this specific matter 

LMM: request was made already 

NM: details? 
... prefer not to limit our freedom at this point, but open to some specific comment 

DC: Larry offered to do something , we should either ask him to undo his offer, or do something to indicate Larry met his commitment 

NM: no problem with what he asked, just as to what, if any, next steps should be 

<DanC> ACTION: DanC to monitor geolocation response to IETF GEOPRIV comments on last call and report to the TAG 

<trackbot> Created ACTION-295 - Monitor geolocation response to IETF GEOPRIV comments on last call and report to the TAG [on Dan Connolly - due 2009-08-13]. 


<trackbot> ACTION-254 Send email to www-tag announcing issue-63 notes added 

close ACTION-254 

<trackbot> ACTION-254 Send email to www-tag announcing issue-63 closed 

LMM: looking for comments on that email 

NM: first goal is to review this email framing the discussion on metadata 

LMM: tried to separate 4 items which have inter-dependencies 

* Metadata model: what is the "data model" for typical metadata applications - the datatypes of the endpoints 

* Metadata serialization: how can metadata be encoded in a representation system, be it RDF or something else 

* Metadata vocabularies: what are appropriate vocabularies for describing various media objects and network services? What is the process by which new vocabularies can or should be developed, described, extended or changed? 

* Metadata linking: What are the various ways in which metadata can be associated with "data" or other resources? Link relationships, protocol elements, mechanisms for embedding metadata in various kinds of data. 

LMM: was trying to create general categories 

NM: we agree to have two issues - 63 being on the metadata itself, 62 being on the linking part 

LMM: Issue 62 is too narrowly defined 

NM: if it doesn't fit in 62, then it's 63.... or 
... broaden 62 to be about metadata access 

<masinter> http://lists.w3.org/Archives/Public/www-tag/2009Jul/0157.html 

LMM: reason for putting these in the same issue is the interdependencies 

NM: would be good thing to update the descriptions of the two issues to explain the relationship between them 

LMM: willing to update 63 to point out that 62 is a narrow part of the general issue 

<masinter> relationship explained in email above 

action on larry to update ISSUE-62 and ISSUE-63 to reference each other 

<trackbot> Sorry, couldn't find user - on 

action larry update ISSUE-62 and ISSUE-63 to reference each other 

<trackbot> Created ACTION-296 - update ISSUE-62 and ISSUE-63 to reference each other [on Larry Masinter - due 2009-08-13]. 

NM: (return to action review) 
Semantics of "resources" in RFC 3986 and RFC 2616

NM: he raises the question of consistency between HTTPbis and RFC 3986 regarding the term "resource" 

LMM: this was discussed on HTTP maillist, those sections of HTTPbis are being updated, and we should wait to review that text 

<masinter> i think i've read them all 

HT: has been lots of discussion, and don't think we can walk away from it 

<masinter> I've proposed a task force which focuses on updating the documents 

NM: is it enough to wait until we see text from HTTPbis? 

<DanC> (I tuned out a long time ago. too philosophical for me.) 

<Zakim> ht, you wanted to point to the scale of the discussion 

NM: OK, let's wait for the HTTPbis work to appear 

DC: Larry suggested a task force 

DC (to LMM): Are you OK to wait? 

LMM: idea was to charter a taskforce to come up with specific wording changes, and taskforce should wait until the updated HTTPbis draft appears 
... rather than handling as a TAG issue ourselves, we actually bring in the relevant people to work on any specific wording changes 

DC: might help to say that we're going to wait for the updated HTTPbis draft? 

NM: I could send a note saying that we have decided to wait for the redraft that we know to be coming? 

LMM: mnot has requested a separate list is created for these discussions 

NM: would suggest that email aimed at getting the TAGs interest should be sent to www-tag - we cannot promise to review other mailing lists 
... is there anything more to do here? 

(no responses) 

NM: should we add to next week's agenda? 

LMM: put it off 

<ht> http://lists.xml.org/archives/xml-dev/200907/msg00157.html 

<ht> http://lists.xml.org/archives/xml-dev/200908/msg00000.html 

HT: people interested in extensibility should look at the thread at the above link 

<masinter> looks really interesting 

NM: can you (HT) send email to TAG inviting discussion "over there"? 

Summary of Action Items
[NEW] ACTION: dan read HTML spec parts from http://lists.w3.org/Archives/Member/tag/2009Jul/0026.html due 30 Aug 
[NEW] ACTION: DanC to monitor geolocation response to IETF GEOPRIV comments on last call and report to the TAG 
[NEW] ACTION: Larry read HTML spec parts from http://lists.w3.org/Archives/Member/tag/2009Jul/0026.html due 30 Aug 
[End of minutes]

Received on Tuesday, 11 August 2009 20:03:54 UTC