See also: IRC log
<scribe> Scribe: Roy Fielding
<scribe> ScribeNick: Roy
VQ: Next week telecon is okay? [yes]
... scribe for next week?
Ed: I'll scribe next week
VQ: and the following week Henry is due to scribe
DC: Please add discussion of F2F diagrams to the agenda
VQ: Okay, added to end of agenda
... Have enough people read the minutes? [no]
... We will defer accepting until next week, so please read them
<DanC> ACTION: DanC to link parts of ftf minutes to meeting page [recorded in http://www.w3.org/2005/10/04-tagmem-irc]
DO: I disagree with the TAG's rough consensus on 47
<DanC> (so the decision didn't stick. oh well; nice try.)
DO: reasons are that, 1) giving guidance to EPR
minters that they should use the address field suffers from the problem that
the W3C has not provided sufficient technology for them to do this, 2)
... 3) stateful services is more than EPR, such as HTTP cookies as session ids, people use them in almost all significant server apps, and criticizing EPRs without acknowledging that people are creating stateful services seems wrong
... TAG could do nothing at all, say there is no issue, and ack that they are separate architectures not on the Web
... or TAG could work on stateful services as a separate architecture
... or TAG could work on creating a solution that places these services on the Web
<DanC> (ah... we do seem to have a recorded action "David to write material on state in distributed application design [12 jul 05]")
<Zakim> DanC, you wanted to note we don't seem to have an issue to collect the cookie/state discussions; perhaps we need one
<noah> +1 to Dan's suggestion
DC: I don't think you need to convince us ... our decision was contingent on consensus and you seem to disagree, so we should work on a solution
DO: I worked on a SOAP to HTTP binding that made SOAP resource addressable as URIs, so we can work on educational material to improve this situation. Should we also press to create technology?
<DanC> (remind me what technology we asked for? I thought we just asked for primer text)
<Zakim> noah, you wanted to ask how fast we need to have feedback to wsa
DO: we may have missed the train
NM: the plan makes a lot of sense, but WSA folks are late in the process and if we are going to do this in a finding then we should do so before they have completed the process
DO: They are already in CR, so it is perilously close to being too late already
<DanC> (double-checking... yup... ws-addressing is at CR as of 17 August 2005)
DO: we could ask for a binding of EPRs to URIs (?)
DC: Have you sketched this out somewhere?
DO: yes, I'll look for link
<DanC> (hmm... title doesn't connect to ws-a "WS-REST continued: do we need an HTTP Transfer SOAP binding and simplified WSDL?")
<DanC> (ah... but ws-a is in the 1st screenfull)
NM: maybe we could make a fast-track approach to figure out what we would like to happen and give a direction, so that we can then decide whether we need to coordinate with WSA progress
VQ: what can we do in the next few weeks? should we get back to this in the next telecon?
DO: I could write up something elaborating the issue and potential solutions
DO: by next telecon
<Zakim> DanC, you wanted to float the idea of inviting the ws-a chair next week.
<scribe> ACTION: DO to draft something indicating the issues with EPR and potential solutions [recorded in http://www.w3.org/2005/10/04-tagmem-irc]
DO: am working on web services for versioning ... will shuffle that back on the pile if EPR is more time critical
<DanC> David to write material on state in distributed application design [12 jul 05] [CONTINUES]
<DanC> ACTION: David to write material on state in distributed application design [12 jul 05] [CONTINUES] [recorded in http://www.w3.org/2005/10/04-tagmem-irc]
<Ed> Phone is dead.. retrying :(
VQ: just an update
DC: this goes all the way back to original
media type specs, how to version PostScript
... I think this is an issue for the CDF wg
NM: At a F2F in Boston at the airport Hyatt, I believe we had some discussion of the use of media types to type abstract bit streams, such as those that might have been encoded in an XML Infoset (e.g. in a DOM).
VQ: do you think it would be helpful to check with CDF WG where they are on this issue?
DC: hold on.. looking
<DanC> CDF issue 18
<scribe> ACTION: VQ to check with CDF WG to see how they have solved issue and if they have more details [recorded in http://www.w3.org/2005/10/04-tagmem-irc]
<DanC> hmm... they have another issue, "Identification, MIME type and Accept headers" that's still open. http://www.w3.org/2004/CDF/Group/track/issues/33
VQ: checking June 2004 minutes the TAG resolved
to move issue to pending state
... is this something we need to record in issues list?
DC: Norm handed the ball to XML CG
<scribe> ACTION: Norm to check on current status of issue xml11Names-46 with XML CG [recorded in http://www.w3.org/2005/10/04-tagmem-irc]
<Zakim> noah, you wanted to report on schema status
NM: Schema WG is also looking at this
... may or may not have coordinated that with XML CG
VQ: recall NM was not sure it was worth spending more time on
NM: I thought it was an area worth working on
when I was new to TAG
... whether and to what extent the URI scheme determined the protocol and how that effects the nature of resources in that scheme
... In June I said I wanted to write something new ... spent the summer toying with that.
... it is not a rush, it is not clearly broken, let's postpone it until we deal with P2P issues and have further discussion on how the resolution process should work.
DC: Tim Bray had written something about making
RSS a one-click, so I was thinking about how we could do that. Perhaps a
SUBSCRIBE method is needed in general for reloading feeds, web pages, etc.
... would like to find a general statement that says "webcal" and "feed": please don't do that
NM: there is an attempt to explain the confusion in the draft
DC: problem is that it is easier to make the misunderstanding greater than it is to find a precise description of what is actually in the architecture
RF: problem is that schemes are not uniform at
all in their prescription of resource and representation characteristics;
writing a simple description can be misleading because it is the finer
subtleties of the architecture that make it so extensible.
... HTTP, for example, is at its essence a layer of indirection, and because of that it is capable of far more than the simplified view of "here is a sequence of octets"
NM:I will pull together a rework of the draft in time for consideration at the Dec. 2005 F2F (or else, I suppose, to report that I've decided that proceeding is a bad idea after all).
<noah> The group implicitly agreed to spend some time at the F2F reviewing such a draft and considering again whether we might in fact do something useful in a finding.
<noah> Noah has to go...I will be on the call next week. See you then.
DC: as I recall, HT still has the ball on this issue
DC and NW: waiting on Henry to say he is done, please review
<scribe> ACTION: VQ to check with Henry on progress on URNsAndRegistries-50 [recorded in http://www.w3.org/2005/10/04-tagmem-irc]
<Zakim> DanC, you wanted to say I very much enjoyed the circles-and-arrows discussion of versioning, and to ask if maybe it would help to try to formalize schemeProtocols likewise and to
DC: I would like to see more diagrams that can be run though RDF-style XML and inferencing engine
DO: I'll look into my tools to see if they can generate XML
<scribe> ACTION: DO to update extensibility finding with the result of Edinburgh F2F discussion and related diagrams [recorded in http://www.w3.org/2005/10/04-tagmem-irc]
<DanC> violet UML diagram tool
<DanC> DanC gets some encouragement to supplement EDI minutes with diagrams, provided he gets some review, e.g. from henry
DC: will try to provide diagrams for minutes of F2F before next week
VQ: next is Dec 5-6 in Cambridge. We scheduled two more at Edinburgh with details available on the TAG website.
<DanC> (what VQ says about meeting schedule is consistent with http://www.w3.org/2001/tag/#deliverables Date: 2005/09/30 15:20:22 )
VQ: Ed should check that with his schedule and confirm