See also: IRC log
<Stuart> scribe: Jonathan Rees
re news items: ashok not here yet.
stuart: dave o not here, so we may not get to last 2 items
ht: request time for ns docs
RESOLUTION: accept minutes for 29 may
norm: regrets for next week
stuart: taking up agenda item 2 - defer for a week?
... (item 2 was f2f minutes approval. deferred)
ht: persistent identifier draft is in the works
(UrnsAndRegistries-50)
scribe: an xhtml module that also makes use of curies
... is in last call
stuart: action items - carry over to next week
stuart: XRI ballot result: failed. >25% no votes
... See agenda for selected messages bearing on this
... Communications problems on both sides
... XRI TC feels it has requirements that can't be met using http:
timbl: A question of trust perhaps: relative reliability of their organization vs. DNS, for example
... Pushback in blogosphere against XRI may have to do with business model
stuart: There are patent issues too
... xdi.org arrangement
... don't have exact pointers right now, but there is some dispute
... What proactive measures should we be thinking about?
timbl: (considering options)
ht: Make goals of proposed discussion clear
stuart: Figure out joint goals
... Maybe aim to clearly elaborate the TAG's goals (to start to work out differences)?
timbl: Nondereferenceable ids are a threat
ht: Folks are flirting with the handle scheme. Boeing ships a locked down desktop - should be nervous about XRI
timbl: Will XRIs be used for documents? href="xri:..." ?
ht: editors of a draft ISO standard for ids in language resource business, leaning toward handles
... used Boeing story to show them the problem
[persistent inaccessibility of XRI-named things via http: on Boeing managed laptops]
<Stuart> http://lists.w3.org/Archives/Public/www-tag/2008Jun/0008
<ht> "XRI vote aftermath" email from Marty Schleiff (Boeing)
stuart: Who would like to respond?
timbl: Which argument is the primary one?
[the following list is Tim's priority ordered list of stumbling blocks]
<timbl> 1) Fragmentation of the space of identifiers, where the community benefits from everything being a URI.
<timbl> That is, not using URI syntax ('='') etc
<ht> "You can't click on xri://=drummond"
<ht> "your resolution service is at least as vulnerable as the DNS system"
<timbl> 2) to make a standard with the goal of creating a new registration system which would benefit particular commercial parties, is not appropriate, when system of governance for DNS already exists (and a big public cost and resposnibility).
<timbl> ?
<ht> "Suggesting that a technical solution will ensure persistence is at best misleading"
<timbl> 3) Building a new resolution system as a layer on top of HTTP is bad from net traffic standpoint and round trips, when DNS already has the community investment in distributed naming, and has for example security being added ?? ??
<ht> There is a parallel with .biz
<ht> Monopolies are not necessarily a bad thing, if they are public services
timbl: This is a fragmentation issue. Compare with the namespace of table tags & attributes - was chaotic before standardization
... Deliberate standardization was in users' interest. Market wasn't forcing convergence.
... XRI vs. http: is analogous - namespace fragmentation not in users interest
stuart: Let's compose a thoughtful piece on this
... What actions to come out of this discussion?
... let's take a snapshot of this to store in tag working document area, as a start
<Stuart> ACTION: stuart " take a snapshot of this to store in tag working document area, as a start" [recorded in http://www.w3.org/2008/06/05-tagmem-irc]
<trackbot> Created ACTION-161 - \" take a snapshot of this to store in tag working document area, as a start\" [on Stuart Williams - due 2008-06-12].
timbl: Not too many people who are prepared to do editing work in this area.
... Possible that W3C should be putting resources into this.
... Let's consider focusing the TAG on this (the ARIA fallout). ? Bringing HTML5 and XHTML together?
ht: It matters. We can't come late to HTML5.
stuart: You have to build credibility
norm: As issues go it's one of the largest.
raman: Lack of modularization is a result of lack of editing resources.
timbl: We may get help from some people who have withdrawn from HTML5 a bit. [scribe's paraphrase]
raman: There are lots of people who are happy with the spec. Stepping in would be disruptive
... Very complicated
stuart: Would need to join the WG. Would I be representing my company or the TAG? Would this go beyond the time I already allocate to the TAG?
... We've been given a list of issues to comment on
others: It was long.
norm: We'd be better off 6 months from now if we could say we got groups at the 2 poles to talk to one another
raman: Implementation of new spec is proceeding 'organically'
timbl: Can we get a small group into a productive discussion?
norm: Workshop?
raman: A modular spec can be produced, even if it doesn't spec modular language growth
<Stuart> +1 to what raman said.
raman: Won't help html5 cause, but ...
<Norm> I do plan to respond to the message