See also: IRC log
SW: Propose to approve 4 October minutes
DC: OK
<DanC> +1 approve http://www.w3.org/2001/tag/2007/10/04-minutes
RESOLUTION: Minutes of 4 October as distributed
DO: Regrets
RL: Regrets
NW: I can scribe
... Regrets for 25 Oct and 1 Nov
<Stuart> http://www.w3.org/2007/11/07-TechPlenAgenda.html
DO: Tantek Celik and Ian Hickson have agreed to be on the panel
DO: TimBL has agreed to be on the
panel, I'd like one more TAG member
... Three issues that are asking for discussion: microformats
-- are or are not URI-based extensibility?
DO: another topic was the
question of how namespaces should be added to XHTML
... and deprecating attributes which are in the way of
extensibility
... Ian Hickson will pbly mostly say we aren't going to get
there
... I have begun to think maybe there's something to this --
easy-to-author and URI-based extensibility are pretty much in
conflict
SW: Around the table to see what people feel like
DO: My preference would be to ask DC, with his HTML WG hat on
DC: OK, I'll do it
<DanC> trackbot-ng, ACTION-46 is done
SW: Norm, state of Action-46?
<Norm> Closed
NW: No further comments
SW: Lots of threads on www-tag,
on httpRange-14 and httpRedirections-57
... One on alternatives to 'resource', converging on
'thing'
... Are people disposed to make this change?
HST: No -- compare
'referent'
... but I didn't read the thread
DC: But there's a previous
understanding of the word, so it's not a good choice
... you can allocate resources, or deallocate them, they can be
accessed. . .
HST: OK, I see
DC: In a tutorial context, maybe
we don't need a name at all -- perhaps the editor of the Cool
URIs spec. could try to avoid a name altogether
... failing which 'thing' has some good properties
... Who has the ball? Us or the editor?
SW: Don't know
DC: I'd like to have a clear
expectation of how the process is going forward -- we make
comments and get quick turnaround, or the editor has gone away
for a few weeks and will be back for a draft, or . . .
... What I'd like is two or three weeks of low-latency
interaction with the editor(s) -- if not this two-or-three,
some other is fine
<scribe> ACTION: Stuart to contact editors of Cool URIs to determine when/how to carry our interaction forward. . . [recorded in http://www.w3.org/2007/10/11-tagmem-minutes.html#action01]
<trackbot-ng> Created ACTION-64 - to contact editors of Cool URIs to determine when/how to carry our interaction forward. . . [on Stuart Williams - due 2007-10-18].
SW: Rhys, where are we wrt HTTP Resources and Endpoints?
RL: I've been letting this rest for a bit -- I didn't feel I had much support for trying to pin down what was in common between all 200s and at least 303s. . .
HST: You convinced me!
DC: 'resource' vs. 'thing' is
editorial, but there's a technical distinction between 'denote'
and 'identify' which Pat Hayes is trying to make
... Anybody get that?
<DanC> (I'm clear on 'access' vs 'identify', but not 'identify' vs name/denote)
SW: My understanding of Pat's point is that there are two senses in which we use the word 'identify' wrt URIs -- one meaning 'access' -- the URI as a way of getting at something, vs. when a URI appears in a piece of text we're using it like a name, w/o any interest in accessing, or any implication that it can be accessed
DC: I don't hear anyone using 'identify' in the 'access' sense
[scribe missed SW's reply]
SW: The intention of the web
arch. is that URIs denote, and sometimes are a means of
access
... another point is that a name can refer to many different
things, which is at odds
DC: I still don't hear how
identify isn't a synonym of denote
... and then there's the 'rigid designator' idea
SW: That's what he's offering . . .
HST: 'rigid designator' for me turns out to be an unsatisfiable concept -- we don't have any in practice
DC: But 'Paris' works, and so do URIs
HST: Agreed, they just don't satisfy the literal interpretation of 'rigid designator' as something which always and everywhere denotes the same thing
DC: What about this thing in the new Schema 1.1 draft -- described as a URI for the Schema language, but there's a document at the end of it. . .
HST: I think this is accidental, not the result of a carefully thought through analysis
DC: I would like to suggest they add a #language -- would that work for all of us?
HST: I think I prefer not --
perhaps we should use this as a good concrete case to work
through as an example of what we want the cool URIs doc't to
say
... I would prefer 303
SW, NW:similar
... We'll put this one up on a whiteboard at the November
f2f
HST: Schema group are not focussed on this right now
SW:
http://lists.w3.org/Archives/Public/www-tag/2007Sep/thread.html#msg148
... Which URI should be persistent when redirects are used?
<DanC> Misha tells the story behind "which should be persistent" http://lists.w3.org/Archives/Public/www-tag/2007Sep/0154.html
DC: IPTC are making up URIs for
the Mona Lisa
... Again, this is the audience we hope will read Cool URIs. .
. --- what should they do?
HST: Did the thread converge?
DC: No -- I suggested a # solution, we want back and forth on that vs. 303, no resolution. . .
<DanC> "The IPTC is likely to approve the NewsML-G2 specification at our meeting in Prague in mid-October."
HST: Sounds like we've missed
their cutoff. . .
... Until we have a document, threads will fail to converge and
that's to be expected
SW: Yes, and I have an action to move that document along
http://www.w3.org/2001/tag/doc/nsDocuments-2007-10-05/
<raman> need to drop off in 3 minutes ...
<DanC> (stuart and I are using the due dates as "next time you want this to be discussed in a tag meeting")
HST: I produced the revision
based on the informal presentation at the f2f
... One outstanding question I'd like advice on -- is
'natureKey' a DatatypeProperty or an ObjectProperty ?
DC: Not sure. . .
DC: I guess the name natureKey suggests string rather than thing
SW: I feel like the 'purpose' relation runs the wrong way
SW: Consider the first diagram
<DanC> the blank node is a nature:Object
DC: Look rather at the instances -- schemaValidation, normativeReference
DC: schemaValidation isn't an instance of Purpose
SW: Yes it is
HST: Please check the ontology, DanC
<DanC> (subproperty)
SW: Take this to email
<scribe> ACTION: Norm or Henry to fix the first two diagrams [recorded in http://www.w3.org/2007/10/11-tagmem-minutes.html#action02]
<scribe> ACTION: Henry to look at the name of the Purpose property class to see if we can't do better [recorded in http://www.w3.org/2007/10/11-tagmem-minutes.html#action03]
<DanC> (I'm not optimistic about improving on "purpose"; I suggest an ed note to say "these words are a little awkward due to the awkwardness of 3-ary relations in RDF. life goes on.")
<Stuart> Henry may find SWBPD writing on N-ary relations useful: http://www.w3.org/TR/swbp-n-aryRelations/
SW: I picked up another problem -- 'identify' gets used a lot
DC: Also, this document isn't done until the first ednotes is addressed
HST: Wrt SW's point, until WebArch changes I don't see the issue
SW: Talk about this now?
DC: So minutes suggests the next
step is a report
... any progress towards a f2f meeting?
RL: Report is out: http://www.w3.org/2007/06/mobile-ajax/report.html
DC: Ah -- report bucks next steps squarely into lap of W3C. . .where will the resource come from ?
SW: Anything on the plenary
agenda about this?
... Lightening talk?
HST: That's a good idea, if they're not full
DC: People there from Dojo and the other big library builders, right?
RL: Besides DoJo, other library builders are Prototype, Scriptaculous, Yahoo, Google, Microsoft
DC: IBM contribute, or have their own?
RL: I don't know if what they've
got is packaged as a library
... IBM's secure mashup approach (SMASH) was discussed
DC: RL, are you tracking this stuff personally?
RL: Yes, I chair the Mobile
taskforce at Open AJAX Alliance -- think of that as rather like
an XG
... We're doing a whitepaper giving an intro to what Mobile
AJAX is
<DanC> local access
RL: and an API/library to give local access to for example device capabilities
DC: Logistics of the taskforce?
RL: All email and telcons -- f2f is only plenary, for whole Alliance, so next one in the spring some time
DC: Any of the AJAX library guys going to be at the tech plenary
RL: Not sure how many are even
W3C members -- see http://www.openajax.org/ for the
OpenAJAX membership
... I did ask the question at the plenary whether they thought
WebArch was OK from their perspective
... perhaps half had read it, and they were pretty much OK with
what they found there
... This was particularly in the area of state for URIs
HST: The whole client-side
persistence thing is mushrooming fast
... perhaps we should look at this at some point
RL: Indeed the mobile ajax taskforce is looking at that as well
DC: I hope we can just trust them
to get it right
... so we can leave it be for a while
HST: I agree we can leave it, but for a different reason
<DanC> (IETF/W3C telcon seems to be 2 weeks from 9 Oct, so it can wait a week)
HST: We've already tacitly agreed to ignore the architectural impact of the fact that most web pages today are constructed on the client by Javascript anyway
DC: We haven't ignored it, we wrote the Least Power finding
HST: Which people are ignoring
DC: No they're not, Christian Heilmann wrote a book about Unobtrusive Javascript, and it's having a big impact
<DanC> See also behavioral separation by Jeremy Keith
HST: Gee, not from what I see coming at my browser
DC: Yes, well, Sturgeon's Law applies ("Ninety percent of everything is crap")