22 March 2005 TAG teleconference -- 22 Mar 2005

<DanC> ah... httpRange-14 is not on the agenda... I kinda expected it to continue from last week... ah... two weeks... 29 MAr

<DanC> "Next week's scribe will be Norm."

<Norm> Crap. Sorry. I'm on another call and can't get away for a few minutes

<noah> Norm: I will scribe for a few minutes until you show up.

Agenda approval

<noah> Agenda at http://lists.w3.org/Archives/Public/www-tag/2005Mar/0078.html is approved

scheduling next telcon

<noah> Telcon on 29 March

<noah> Possible regrets from Noah next 3 weeks

attendance

<DanC> 15 March 2005 Tag Teleconference

<noah> VQ: Here: everyone except Roy, Tim and Norm

<noah> scribe: Noah

<scribe> scribe: Noah Mendelsohn

<scribe> scribenick: noah

Henry will scribe on 29 March

approval of minutes from 15 March

RESOLUTION: minutes of 15 March at http://lists.w3.org/Archives/Public/www-tag/2005Mar/att-0056/March152005.html are approved

VQ: shall we move the minutes in date space?

DC: No, leave them where they are.

HT: That means they are in an attachment in an email archive, which makes searching hard. I needed that today. Is it not policy to have them in date space?

DC: They need to be linked from the tag home page

VQ: Right, I've been doing that.

HT: Well, it's easier to grep if you mirror date space, but I can write a better tool.

ED: I somewhat agree, I'd prefer to see them all in common place in date space, per year.

DC: You don't have to go through list archives, they're all one click away.
... In any case, in general, I'd like them to be in a final resting place before we approve them.

<scribe> ACTION: Henry with help from Ed to draft proposal on where in date space to put minutes [recorded in http://www.w3.org/2005/03/22-tagmem-minutes.html#action01]

<DanC> (re filing the minutes in CVS/datespace, all of us can send mail, but only some of us can do CVS, and when it goes bad, it tends to fall on the team contact, i.e. me)

NM: Scribe's question: should we unapprove the minutes of 15 March until they land in whatever is the best place?

Several: No, they're approved, leave them.

F2f 14-16 June

VQ: Amy van der Heil reports MIT can host

DC: 3 real days?

VQ: maybe last day will be short, but otherwise yes, full 3 days.

NM: Remember that TimBL will leave early on 15 June due to family birthday

<DanC> TAG action items

<DanC> (very handy so far, thanks, ht)

VQ: When you make flight plans, please let me know so we can schedule wrapup on last day

+1 Noah

Pending Action Review

See pending actions at http://www.w3.org/2001/tag/2005/03/action-summary.html

DO: Asked some questions about which are assigned to him.

NM: Yes, ACTION: Henry and David to draft initial finding on URNsAndRegistries-50 [recorded in http://www.w3.org/2005/03/15-tagmem-irc]

HT: Also, Dave has ACTION TP5-8:David Orchard to contextualize his scenarios, such as more on what is happening with SOAP and WSDL. [recorded in Minutes of the W3C XML Schema Working Group 4th (37th) F2F meeting]

DO: Yes working on it
... I also worked on terminology for extensibility and versioning.
... sent to Norm and Noah for early review

NM: Don't have it yet,

HT: I have finished ACTION: HT to review " Storing Data in Documents ..." [recorded in http://www.w3.org/2005/02/22-tagmem-minutes#action05]

DC: right, and followup email is largely supporive

<DanC> action list Date: 2005/03/21 11:50:27

VQ: will update action list later today or tomorrow

<DanC> (re how long done action stay... a week or two, please)

Oasis work on Extensible Resource Identifiers (XRI) -- raised by

Norm, Oasis announcement[11]

s /[11]//

VQ: Hmm, Norm's not here, let's skip it until he shows up.

How do you extend the HTTP?

We received a request at: http://lists.w3.org/Archives/Public/www-tag/2005Mar/0075.html

VQ: I would like to have some discussion on what to do with this issue.

Norm joins the call.

DO: A couple of comments: 1) This issue could use some authoritative comments

<Norm> DO: expresses concern that TAG is picking up a lot of issues but we aren't closing them very fast

<Vincent> no echo on my side

<Norm> DanC: I think you write an RFC and get consensus from the community

<Norm> NM: Some groups use a new header, some use a new method (WebDAV). These have different characteristics in the face of things like "must understand"

<Norm> NM: I think he's asking for good practice, clarity on who should do what and when

<DanC> (hmm... I still don't see an issue any smaller than "please predict the future for me")

<Norm> VQ: Agrees, it's a good practice request. Not clear who's supposed to do this, us or IETF, for example.

<Norm> VQ: Shouldn't we do something?

<Norm> DanC: No, we're not obliged to take on an issue or formally respond to every request

<Norm> DO: If the TAG is going to decline, we should at least say we decline.

<Norm> NM: +1

<Ed> +1

<Norm> DO: I'd prefer if we could provide a bit of rationale. I don't think we get an enormous number of requests such that we can't reply.

<DanC> (yes, it's polite to explicitly decline. but if you try to formalize that as a policy, you'll quickly get into denial-of-service, and "have we already declined that request?" stuff)

<Norm> NM: summarizes, asks if we're ready to decide

<Norm> NM: I'd be interested in the opinions of timbl and royf.

<Norm> NM: Two options? 1. reject or 2. pick up the issue and prioritize it later

<Norm> DanC: Putting it on the issues list is a commtiment to resolve it

<Norm> DO: Some issues that we took up were reduced in priority before the first webarch but those are being reexamined

<Norm> DO: Proposes that we defer talking about this issue until timbl and royf are present

<Norm> VQ: I'll draw their attention to the issue before next time

<Norm> VQ: Return to XRI.

<scribe> scribe: Norm Walsh

<DanC> "The public review starts today, 15 March 2005 and ends 14 April 2005."

<scribe> scribenick: norm

ht: Included it in new issue 50. Reinventions of URNs and registries.

NDW: That satisifies my expectations of what we would do with this

<Zakim> DanC, you wanted to express some patent concerns about reading it at all

DanC: XRIs have crossed my desk a couple of times, but the package seems to be labeled "patent incumbered" so I'm not inclined to read it at all
... their deadline is 14 Apr. HT, are you inclined to figure something out by 14 Apr?

ht: That seems unlikely

<DanC> (we had pretty close to a finding on .mobi; we had a web page with TAG endorsement)

ht: At the very least, should we say "uh, guys, would you like to talk to us about this before moving ahead?"

Ed: I'd be happy to review it and try to highlight some of the major issues

DanC suggests mailing comments to www-tag or tag or straight to them. Any of those is fine by me.

Ed agrees to read them and post some notes about it

ht suggests taking a quick glance at urnsAndRegistries-50

VQ: Does that address your concerns?

NDW: Marvelously.

<scribe> ACTION: Ed to review XRI and provide comments [recorded in http://www.w3.org/2005/03/22-tagmem-minutes.html#action02]

xlinkScope-23

DanC: I believe we closed it in Basel. There was some kickback but eventually it did stick.

<DanC> http://www.w3.org/2001/tag/2004/10/05-07-tag#htxl

DanC: I believe the issues list should be updated to record our decision to close the issue

VQ: I'll do that.
... Any other information about this issue?

ht: My memory is that the public believes that the TAG said you should use XLink, HTML WG pushed back, TAG said you should consider it, HTML WG went quiet.

My memory is that the HTML WG said even considering it was too strong, but we stuck our ground.

<DanC> (well, yes, mark it closed, but note some outstanding dissent)

VQ: Any other business?

ht: I would be happy if we brainstorm about on URNsAndRegistries-50

DanC: Countless discussions go like this: I'll find some URN scheme or the equivalent, e.g. doi: and urn:doi:
... They've gone so far as to deploy plugins for doi:.
... what the plugin does is lookup http://doi.org/...
... So they own and operate a mapping from DOI to HTTP
... Ask these folks why not just use http? Why a separate scheme? One part of it is financial incentive for being at the center of one of these namespaces
... The other is that they don't trust DNS and HTTP.
... Engineers can't predict the future. I can't predict that DNS and HTTP will last forever.
... So they really do want their stuff to be looked up and they can't be talked out of it.

NM: They've got a mapping, the insurance they're getting is that if someone steals their DNS name, they can redirect to another.

DanC: Clearly they're creating aliases here, which we've discouraged.
... The other folks don't want their stuff to be looked up.
... e.g., urn:googlesearch:, they don't do anything about grounding that in reality and they don't feel embarrased about it.
... But for some reason they don't want to promise that an address will persist for a long time.
... Consider urn:ietf:...
... How do you manage it? Well, we keep a website with all the names in it.
... Duh!
... So they have no mapping, but to actually manage the namespace...they use a webserver!
... I promised to renew that draft if someone would stand by me for the incoming barrage, but there have been no offers

ht: Two things I'd add: apparently the IETF are now running a server that will lookup those URNs.
... I haven't persued it, but someone asserted it exists.

<DanC> A Registry of Assignments using Ubiquitous Technologies and Careful Policies

ht: The other example, the ITU are looking at doing this (as is OASIS, i.e. XRI)
... Both of these guys say they'll be running servers, in the OASIS case it'll be a SAML server of some kind
... The part of the puzzle that I don't understand how to respond to is, the argument that "we need something abstract" something not as concrete as URLs
... We need something independent of specific locations.
... That sounds like broken record stuff to me, but I'm hoping to hear "oh, they don't understand such and such..."

DanC: I can replay a conversation where I convinced one person.
... The name of the XML spec was a subject of conversation.
... Do you feel bad that there's no URN for it? Answer: yes.
... Why? Because we want it to survive
... Redundancy is the key, putting something in a newspaper gets lots of copies.
... So the copy of the XML spec is all the web caches around the world provides that.
... So he says "gee, then maybe we shouldn never have done that URN stuff"
... The way you make things valuable is by getting agreement that things are shared. So you can use a link instead of sending a 15 page spec.
... The way the binding between the names and what they mean is established is through protocols of some sort. HTTP is one example.
... it makes sense to makup new URI schemes for totally new communication patterns, but if it looks like DNS and HTTP, *use* DNS and HTTP.

<DanC> (http://norman.walsh.name/2004/03/03/266NorthPleasant is unavailable at the moment, but records some relevant experience of Norm's)

ADJOURNED

What's the incantation to get rrsagent to make the log public?

<DanC> norm, do you want it to draft minutes?

Sure.

I'll take a look at cleaning those up as soon as I get a couple of other things off my plate

<DanC> not bad... noah knows how to drive it. ;-)

<DanC> hmm... it doesn't recognize Norm as scribe too...

<DanC> ScribeNick: Norm

- DRAFT -

22 March 2005 TAG teleconference

22 Mar 2005

Attendees

Contents