W3C

- DRAFT -

TAG telcon

18 Jun 2007

Agenda

See also: IRC log

Attendees

Present
Dan Connolly, Rhys Lewis, Noah Mendelsohn, Dave Orchard, T.V. Raman, Henry S. Thompson, Stuart Williams
Regrets
Tim Berners-Lee, Norm Walsh
Chair
Stuart Williams
Scribe
Henry S. Thompson

Contents


Admin

SW: ER will join us at 1 hour into the call

HT: Works for me, but we shouldn't wait for him if we run out of other business

SW: Will rearrange agenda to put PiC last
... Regrets from NW
... I have sent review of Cool URIs for the SemWeb to the tag list. . .

NM: I read it, I thought it was good -- ship it

SW: I will send it to the public list and we can discuss it at a later time

DC: Since RL is here, we could talk about httpRange-14 today and keep the momentum up. . .

SW: OK, we can try to do that

<DanC> (link from agenda to http://www.w3.org/2001/tag/11-minutes is hosed)

<DanC> ah... more like http://www.w3.org/2001/tag/2007/06/11-minutes

<DanC> I propose to approve http://www.w3.org/2001/tag/2007/06/11-minutes (version of 2007/06/11 23:46:49 ) as a true record

HT: I'm happy with those minutes

SW: Approved.

Naming and Virtual Worlds

<Noah> Reviewing email I sent this morning: http://lists.w3.org/Archives/Public/www-tag/2007Jun/0074.html

NM: I'm not a virtual-worlds expert, but with my TAG hat on, I am interested in the relationship between virtual worlds and the Web. We know the pitfalls of having multiple Webs, in which resources from one can't link to the others; so, I think we want one Web. So I'm trying to queue up the issues we as the TAG should look at relating to virtual worlds.

Talking about this, particularly about integration, tends to dive down to technical details too fast. I want to urge that we look at use cases first. Some questions:

  1. Should I be able to click on a Web link and end up in a virtual world? Understanding whether we even want to achieve that is important before we debate details of URIs, schemes and protocols.
  2. Relating to URNsAndRegistries-50, what things in v-ws should have URIs, and what kinds of URIs, using what URI schemes? What about consistent metadata _in_ the URIs (note that Second Life SLURLS encode latitude/longitude/altitude). We could have built a Web in which everything with a physical presence, such as the store down the street, always had a URI with geospatial coordinates. There were some good reasons we didn't. (Though such URIs are used for services like MapQuest, Google Maps, etc.)
  3. Shouldn't we be looking at this from the same perspective we established in the generic resources finding? Consider the following example:

Compare the following two approaches to a virtual store:
... a) "For good deals on washing machines, see http://example.com/cheapWashers (Second Life users, go to http://slurl.com/secondlife/Example/100,200,300)",
... b) "For good deals on washing machines, see http://example.com/cheapWashers (works in Second Life too!)"

NM: I much prefer (b), per our recommendation in http://www.w3.org/2001/tag/issues.html?type=1#genericResources-53
So if we want to push this approach, how would it work in detail? 30x to a slurl? content negotiation. . .

<DanC> (the "For good deals on washing machines... Second Life users, go to" example is good.)

<Zakim> DanC, you wanted to ask about the social structure behind slURLs and to relay regrets from timbl (http://www.w3.org/2007/eGov/eGov-policy-cfp.html )

DC: Regrets from TimBL
So there's a company which owns slurl.com, and they can in principal do whatever they want. There's nothing special about this, is there?

NM: Well, there is a set of rules about 'regions', the bit before the 3D coordinates. I presume that, just as we see DNS governance issues relating to the authority part of HTTP URIs, there would be analagous issues in how the owners of Second Life assign 'regions', which are the path segment ahead of the x-coordinate in a SLURL.

<Zakim> raman, you wanted to ask Given URI u_1 in the real world, does it take you to the same place in the virtual world? Conversely: given a URI in the virtual world V_1, does V_1 go to

TV: One of the key ideas behind the generic resources finding, is that there is a single web, so that you could e.g. take a URL you picked up from a mobile browser and use it on a desktop I'm not clear that this works the same way wrt V-Ws Some things, such as www.ibm.com, will make sense in both 'worlds'

TV: but lots of other things, which I might find in a V-W, will not have an analogue in the real world, and vice-versa So the conneg might fail

<Noah> Raman is questionning whether things that exist in virtual worlds >necessarily< need a manifestation in traditional Web space. I'd turn that around, I think there exist many examples of resources where you >do< want that generic approach. Requiring it in all cases is a different question, and not what I was pushing.

<DanC> (the ICANN-level world is different from slurl.com in that there is open competition between domain registrars.)

HT: So just like conneg for e.g. a French version -- there may well not be one

RV: [not well scribed:] We just don't know when changes in a V-W will result in changes in the (web-accessible) 'real' world, and when not

<Zakim> Noah, you wanted to talk about virtual worlds connecting to each other

NM: There is discussion of moving an avatar from one V-W to another. . . Just as we recognise that BitTorrent is a very different kind of protocol from HTTP, so when I'm 'walking' through a virtual world, I'm worried millisecond-by-msec about what I'm going to run into, which makes the protocol very different from HTTP. I'm just worried that the discussion isn't happening in the right order -- I don't hear people from the V-W side thinking quite as much as I'd like about the deeper Web integration issues before locking in various bits of naming and protocol design. (Or maybe they are and I'm just not as aware as I should be.)

DO: It's not just the 'dot.com's of the virtual worlds, such as Second Life, but also customised V-Ws being built for local purposes
... Sometimes even nested within one another
... Consider the relationship between Google and V-Ws -- we need the URIs to be built right so that they can be found too

NM: The Web3D Consortium are involved in this space, and they are W3C members - we should consider whether we want to ask them for some involvement. They may be able to give us good guidance on how to work constructively with the virtual worlds community ... if we don't decide that it's just too early for the TAG to be involved.

NM: Also, I know that some of these companies are small, and working hard just to meet customers' needs month-by-month. We'll have to show them that there's real business value in paying attention to these architectural issues. Again, I'm wondering if the Web3D folks can help us to engage appropriately.

<Noah> I'd like it recorded that I mentioned that Web3D is one of the consortia that's very active in this space, that they are W3C members, and that Don Brutzman is their AC representative. I suggest that we may want to solicit his advice on what, if anything, the TAG can do that would be constructive.

<Zakim> ht, you wanted to argue we should worry about this

HT:I think it is important that we get this right now, or we'll regret it later. Universality matters.

HT: Even if we only put a stake in the ground, saying "These are the trade-offs, you really should look at this for the medium term"

SW:Should we worry about this?

DC:Not sure whom we would engage.

SW: Not sure how we would go about doing this, I only engage with this space informally

DO: Precisely because they are at a formative stage, they might be influencable

DC: Who would we talk to?

DO: Well, I could talk to my friends at [xxx], who are doing a startup to build customised virtual worlds

NM: I agree with DC to some extent -- I think this is important, but I'm not sure I yet know exactly what to suggest we do that would make a difference.

<DanC> "The tech startup Doppleganger has designed a virtual lounge where fans can go online to chat with members of Interscope Records' Pussycat Dolls." -- http://publications.mediapost.com/index.cfm?fuseaction=Articles.showEdition&art_send_date=2006-5-15&art_type=13

NM: We could at least have something like a finding, so that when they do have time for this issue there will be something there. One can debate whether this V-W stuff is going to take off, but if it does I think it's important that the Web integration be really good.

DC: The next generation may not see web-pages or email messages at all, just V-W. . .

TV: Some V-Ws are better integrated than others -- MySpace is, SecondLife isn't

NM: Unless I'm misunderstanding, MySpace aims at a Web-grade user experience. Systems like Second Life are much more aggressive in trying to provide 3D, Animation, etc., and that necessitates the use of technology beyond what HTTP, HTML, etc. gives us today.

<DanC> (there's also the point that whether one could build SL on web tech or not, the existing SL does not use much web tech.)

NM: If my fingernail in SL has a URI, is there a (negative?) impact on what the actually technology is doing as my arm is 'moving'?

<Noah> Let me explain that bit about the fingernail: I was really saying that I think that the virtual worlds protocols are, for good reason, tuned for very high interactivity, animation, 3D collision detection, etc. Expecting those protocols to use a URI for, e.g. each little bit of my body (a fingernail) may be impractical. Offering a URI for purposes of gateway into the Web world may well be a good thing to do.

HT: Briefly, to come back the question of what we could do: Many of us seem to have contacts, we could use them in at least three ways: 1) Invite them to a telcon to discuss this with us; 2) Draft a finding and ask them to review and comment; 3) Host a workshop: "Come discuss naming and Virtual Worlds in a neutral context" and use it as an opportunity to promote integration.

SW: MySpace is very much about being on the Web and publishing there, but wrt SecondLife it's not clear to me how much the Web matters to them, it's primary focus is being its own self-contained game space.

SW: What would our first steps be if we were to take this up?

HT: First step -- a short summary of how naming works in V-Ws today

NM: I'm interested, but my queue has gotten long. . .
... I have vacation and other obligations. Probably wouldn't make substantial progress until August the earliest, but I'm willing.

SW: I'm prepared to schedule discussion at the end of the summer

NM: Should I contact Don Brutzman and point him to these minutes, and ask for advice on how to proceed?

HT: I will try to help, but like NM I have things queued up between now and holiday time.

NM: "We think there might be issues here, we're looking to see if we can find ways to make a difference" is an acceptable summary of our situation?

<scribe> ACTION: NM to contact Don Brutzman to query about possible contacts about naming in V-Ws and integration with the Web [recorded in http://www.w3.org/2007/06/18-tagmem-minutes.html#action01]

PasswordsInTheClear-52

TV: Mary Ellen Zurko offered to help move this forward, I think we should take her up on this

SW: Let's postpone PitC until next week, I'll invite MEZ and ER, trying again

Tracker

SW: Dominique Hazael-Massieux has done lots of good work to get us going with this. Got the issues moved across. Alas no short names for issues. The model of the life cycle of an issue is very simple -- only members of the group can raise an issue, actions and issues are either open or closed, it's forms based

SW: If we do go over to this wholesale it would change the URIs by which we identify our issues

<Noah> If we do go with tracker, what are the pros and cons of having the old issue URIs redirect to the corresponding tracker entries. I have >lots< of old content out there with links to the old URI issues.

<Zakim> ht, you wanted to worry about actions vs issues

HT:Looked at this today. Lots to like, but 3 concerns. Taken together, I'm unhappy. 1) shortnames is first issue. Could be added reasonably easily; 2) Actions are not attributed to issues in Tracker; 3) Third, can't understand why issues can only be attributed to group members. What about LC issues etc?

DC: You can connect actions to products, and products to issues. . .
... but maybe that's a category error

SW: It would make my life a lot easier as chair, replacing XML edits to a monotonically growing file with using forms backed by a DB is a big win. Dom has done a great job, but I don't want to ask for more effort until I know we are planning to commit to it

RL: The assumption is that actions go with WG members, not issues
... I've used both EXIT and Tracker, and it's much easier to use Tracker

RL: I think associating actions with issues is a generic requirement, and we could ask Dom about that

HT: I would be happy to talk informally to Dom about what the Systeam would be willing to consider adding. . .

SW: I will keep using EXIT until we decide to change

DC: "He who does the work makes the rules" -- you can take us there if you decide it's what you need, don't make it a WG decision

SW: I'd like to take as much history as we can with us, but could live with not getting full benefit except for new business

<DanC> (that's a great RFE for tracker: "show me *everything* on one page")

NM: I like the fact that I can see everything about an issue in the existing issues list, and I hope we don't lose that in the new system
... Not necessarily all the issues, but everything to do with one issue, should be all available in one place

NM: Worst case that's only true for the old history

<Rhys> I note that, for example, the history of Range 14 seems to be at http://www.w3.org/2001/tag/group/track/issues/14

<Zakim> DanC, you wanted to think out loud about porting history

DC: The existing (EXIT) approach is modelled on Ian Jacobs's approach to workflow, with edited references
... whereas Tracker is just an automatic collector of mentions of issue numbers

DC: We could at least just email each issue section from EXIT to www-tag with the right number in the subject line

<Noah> Well, I really liked having the >edited< history. Don't know whether that's practical from a workload point of view. I guess I didn't notice that tracker really was just grepping emails. Makes me nervous.

<Noah> We have a www-tag archive, and can sort it by subject already.

<Noah> So, what do we use as a URI for, e.g. URNSandRegistries-50?

<Noah> We had one until today, and I'm starting to think that "cool" URI is about to change.

SW: I'm inclined to make the move, but not quite yet, will continue to discuss with Dom and HT

DC: What's the latency for the current issues list?

SW: Weekly -- I try to bring it up-to-date on Wednesdays, always by Friday, in preparation for doing the agenda
... also depends on the availability of the previous minutes
... sometimes it takes most of the time I have set aside to prepare the agenda to get the issues list up to date

Future Directions

<DanC> http://lists.w3.org/Archives/Public/www-tag/2007Jun/0069.html

<DanC> # tag participation in W3C/OpenAjax Workshop on Mobile Ajax 28 Sep in Mountain View? Dan Connolly (Friday, 15 June)

DC: DC and NM are interested, but can't go. RL is going. Position papers due by 15 August -- RL writing one for TAG, or Volantis, or ???

RL: Good question, not sure

<DanC> "To participate in the Workshop, you must submit a position paper by 15 August 2007 explaining your interest in the Workshop."

RL: I'm involved via the Open Ajax Alliance side of things -- they're looking at the mobile dimension. I'm assuming the TAG would like feedback rather than to take a position

DC: Not sure about that

SW: Connects up with our discussion at the Mountain View f2f about actions and how they fit in. . .

DC: What do we think about Mobile -- one position is write once, convert for devices. That's the 20% case, I hope the 80% case should be that you just write a good web page, and it works on any device, but the reality is the opposite, for example I can almost never see the gate for my flight from my cellphone

<DanC> "aa2go considered sub-optimal; just make aa.com work on mobile too" is my position.

RL: So, would the TAG like to submit a position paper to the Mobile Ajax workshop?

SW: Quite possibly

httpRange-14

SW: http://lists.w3.org/Archives/Public/www-tag/2007Jun/0075.html is my review of the "Cool URIs for the SW" draft

RL: There is a reference to that document in the draft finding

SW: The question of what the relationship between the two documents is open. One view was that we should just adopt it. but there was pushback about target audiences

RL: There's also the fragment side of things, where the Cool URIs doc't is incomplete, per our discussion in Mountain View

SW: And on the other hand, they have some detail on 30x which the draft finding doesn't

RL: Yes, I took away a request to update the draft in the area of 30x

HT: I think the target we ended up identifying in MV for the finding is importantly distinct from the place Cool URIs is, and we need a finding in that place

Summary of Action Items

[NEW] ACTION: NM to contact Don Brutzman to query about possible contacts about naming in V-Ws and integration with the Web [recorded in http://www.w3.org/2007/06/18-tagmem-minutes.html#action01]

Minutes formatted by David Booth's scribe.perl version 1.128 (CVS log)
$Date: 2007/06/19 10:54:56 $
=