See also: IRC log
<scribe> Scribe: Dave
<scribe> ScribeNick: dorchard
<Norm> Agenda: http://www.w3.org/2001/tag/2007/10/04-agenda
<DanC> (I just updated action due dates; if http://www.w3.org/2001/tag/group/track/ shows an overdue action, please give a new ETA)
proposal: tag meet in vancouver for 2 1/2 to 3 days, Feb 25-28th 2008
... tag meet in vancouver for 2 1/2 to 3 days, Feb 26-28th 2008
<DanC> so RESOLVED.
RESOLUTION: TAG meet in vancouver for 2 1/2 to 3 days Feb 26-28th 2008
<DanC> (skw has an action to follow up on tim's schedule. ACTION-61 continues.)
TimBL: meeting starts at 9am on Feb 26?
Discussion of possible local venues
<DanC> yes, let's meet within an hour of YVR
<timbl_> I woudl travel on Fri
<Rhys> I would travel Friday too
<ht> Last same-day flight appears to be 1330 departure, via Chicago on United
<DanC> Rhys to scribe 11 Oct
Next meeting Oct 11th
Rhys to scribe
<ht> So I don't think leaving Thursday departure gives us 2.5 days for Noah/TimBL
proposal: meet for full 3 days
RESOLUTION: full 3 days, Feb 26-28th
RESOLUTION: F2F of Sept 19th etc. approved
http://www.w3.org/2001/tag/2007/09/17-tagmem-minutes.html and http://www.w3.org/2001/tag/2007/09/18-tagmem-minutes.html and http://www.w3.org/2001/tag/2007/09/19-tagmem-minutes.html with summary http://www.w3.org/2001/tag/2007/09/17-agenda
RESOLUTION: Minutes of September 27th approved. http://www.w3.org/2001/tag/2007/09/27-minutes
DO: some panelists have accepted, some thinking it over...
danc: if popular tools required @profile usage.
timbl: how about writing a microformat validator that checks the @profile?
... it could report "you don't have a valid microformat" how about adding the @profile..
report has issues list management section... more on that.
what URI is "right" for issue 14?
<scribe> ACTION: Stuart update http://www.w3.org/2001/tag/issues to say "this is old; see http://www.w3.org/2001/tag/group/track/ " [recorded in http://www.w3.org/2001/tag/2007/10/04-minutes#action01]
TimBL: can we add links back? or both ways?
DanC: any of us can add a note in the new system to point to the old entry. that suffices, for my purposes. Yes, Dom could do it programmatically, but it seems to me that it if it matters, one of us can do it ad-hoc.
timbl: it was talked about by tech plenary planning committee but it didn't make the cut..
ht: michael smcq will attend exi
... people could say EXI begins with an xml declaration, and then some more that I couldn't captcha
... the TAG set a bar, and if the TAG won't follow up who will?
danc: looked but couldn't find the TAG's note on setting the bar.
<DanC> (reviewing records... http://www.w3.org/2001/tag/issues.html?type=1#binaryXML-30 )
<DanC> (deferred on 12 May 2004 ...)
<DanC> (ah... # TAG opinion on XML Binary Format , 24 May 2005. ... http://lists.w3.org/Archives/Public/www-tag/2005May/0044 )
the 4th chart compaction: "both" application class looks to show that gzip is way worse than exi on the low end
<DanC> hmm... "6. Summary and Analysis of Test Results" http://www.w3.org/TR/2007/WD-exi-measurements-20070725/#results
I didn't find that there was enough "middle" analysis between the detail and the summary that "exi is great".
<DanC> For example, a target might be that "in
<DanC> typical web services scenarios, median speed gains on the order of 3x in
<DanC> combined parsing and deserialization are deemed sufficient to justify a
<DanC> new format."
DanC: the combination of a use case and hard numbers in one sentence would be particularly handy
... some mobile folks have indicated there is a bar, say 7 times smaller, that means it could ship in a handset
... I'd like to get the real numbers from the mobile community
stuart: have exi met the bar the tag set?
they may have met it but I can't tell.
<DanC> (we might be reading too quickly, but the typical W3C tech report reader is likely to have even less attention span than the average tag member)
<timbl_> I would like a scatter-plot of gzip vs efx
rhys: we could ask them whether they have met the bar?
in general, the tag doesn't seem sure.
<timbl_> They decided not to put an initial 4 cvhaarectyers on e file format for eg the unix 'file' command as it would add 4 bytes
<timbl_> So they are not likely to add am encoding stirng
ht: in constrained environment, start with exi and fail over to xml rather than start with xml and fail over to exi
timbl: you could vector on the "2 bytes" they start with
... first byte is between 80 and DS then they know it's binary
... can vector between with a single instruction
ht: when the 2 parties are well synched, that raw binary xml is perfectly acceptable.
... binary xml should start with an xml declaration that says the exi encoding
... and then xml spec says that xml declaration can be omitted if both parties "know" what the encoding is
timbl: everybody needs to change their processors to accept binary xml without any external labelling?
ht: if they specify it the way i suggested, then the door is open
<Zakim> DanC, you wanted to note that the SVG case is interesting, in that .svgz is widely deployed
ht: if lots of people started wanting shift-JIS, then people would add that encoding with no change to xml
danc: note that the SVG case is interesting, in that .svgz is widely deployed
... I wonder why svg thinks about exi?
stuart: our own next step?
dorchard: i can look at the end of October.
<Rhys> Rhys just notes that mobile networks sometimes silently compress materials on the way to the device
<scribe> ACTION: dorchard to review EXI measurements against the TAG bar, due Oct 31st [recorded in http://www.w3.org/2001/tag/2007/10/04-minutes#action02]
<trackbot-ng> Created ACTION-63 - Review EXI measurements against the TAG bar, due Oct 31st [on David Orchard - due 2007-10-11].
ht change action 48 to Oct 11th
<DanC> ACTION-38 now due 18 Oct http://lists.w3.org/Archives/Public/www-tag/2007Oct/0021.html
<DanC> on story... http://lists.w3.org/Archives/Public/www-tag/2007Oct/0021.html
my phone just gave low battery warning
<ht> At its best, this is the TAG equivalent of " 'Snow is white' is true just in case 'snow' is 'white' "
<ht> That is, it's a statement that what the world understands as extensibility is a good thing, using our careful definition of extensibility
Discussion of first bullet, specify that the language is extensible.
<Stuart> "...provide clear extensibility points in the design of your language." ?
danc: one way of thinking about this is "if you are designing a language for a distributed environement provide extensibility"
<DanC> If you're desigining a language for use between parties in a distributed system, think about extensibility
<DanC> ok, I'd like the bullet points to stand alone... so...
<DanC> * to enable forwards compability, start with an extensible syntax
json isn't worried about compatible versioning so they don't need extensibility
<Stuart> "If you anticipate future variation in your language, provide clear extensibility points in the design of your language."?
I had worried about defining distributed
we've gradually stripped out the motivation material to do just the "what you need to do"
<timbl_> FYI, on magic numbers: less /usr/share/file/magic
DanC: Re: ACTION-38, I declare victory on my action item.
stuart: Will keep as "pending review" in order to schedule future discussion.
ht: Re: ACTION-48, aiming to have a new draft of xhtml modularization using subst groups ready for next week
<DanC> not for discussion next week
rhys: goes through the minutes
... w3c and openajax may have lists of things to work on..
... ibm had a solution in the client space, some level of secure interaction. Dave Raggett may be looking at a workshop on security aspects
... the IBM solution has to deal with secure communication between parts of the app
relationship between frames, script
rhys: it's a sandboxing solution.
<Zakim> DanC, you wanted to note Ian Hickson on 21 Sep "proposal for offline web app API " http://lists.w3.org/Archives/Public/public-html/2007Sep/0445.html
ht: there are 3 or 4 different offline web app APIs
danc: exactly. He's trying to standardize
how about that for timing of my phone battery dying!