TAG telcon

17 Mar 2011


See also: IRC log


Dan Appelquist, Yves Lafon, Peter Linss (in part), Ashok Malhotra, Larry Masinter, Jonathan Rees, Jeni Tennison (in part), Henry S. Thompson
Tim Berners-Lee, Noah Mendelsohn
Jonathan Rees (pro tem)
Henry S. Thompson


<scribe> Agenda: http://www.w3.org/2001/tag/2011/03/17-agenda


Regrets for 24 March: tbl, hst

Scribe for 24 March: pl

<DKA> Minutes 10 March OK with me.

JR: RESOLVED: Minutes of 10 March 2011 are approved

IETF Meeting in Prague

<Larry> IETF agenda is https://datatracker.ietf.org/meeting/80/agenda.html

<Larry> /me quotes: 4. Technical Session:

<Larry> "The Future of Applications"
Panel session moderated by Jon Peterson
Jonathan Rosenberg (Skype)
Harald Alvestrand (Google)
Henry S. Thompson (W3C)
Possibly more


HST: Thanks to AM and NM for input
... I've included versions of the material they sent
... Plan to use a subset as appropriate

LM: You're supposed to be talking about the Future of APplications
... So change the title of the talk

HST: Will do

LM: Applications are going away, to be replaced by Web sites
... We may not like this, but it's happening

<DKA> +1 to LM.

LM: So Web Arch is application architecture
... Put this earlier
... to clarify why webarch is relevant to talk about the future of apps

YL: You can see the replacement happening both ways

<Larry> some sites might be replacing a web site of documents with a web site of one application, but it's still "web architecture"

YL: Website has only one URL, all content is computed
... Web Arch is not cast in stone -- Web evolves, TAG tries to keep up

<Ashok> +1 to Yves' comment re. evolution

HST: Yes

LM: Remove 5 & 6 because I don't like what they say. For example I think URIs don't have owners. Resources maybe

HST: Noted

HST: May be cut if time is short

LM: How relevant are they to the question before the panel?
... Slide 7 was true -- are those assertions true of Web Apps?

HST: Every single one needs to be re-examined

<DKA> Under slide 11, you might want to include a link to the joint IAB/W3C/ISOC workshop on privacy from last year: http://www.iab.org/about/workshops/privacy/

LM: Historically there is an Arch of the Web of Docs
... Now we have to migrate that to the Arch of the Web of Docs and Apps
... Make that clear earlier
... That gives us a context for 5, 6, 7, #!, etc.
... Side-effect free? View source less helpful if it's all JS?

HST: Thanks, that's valuable as both source of fixes and as guidance for rhetorical stance

<Yves> I would note that the issue about media types is a good example of possible cooperation

<Larry> Maybe we should work either now or by email on what the design issues are in moving from web of docs to web of apps... e.g., does "view source" still work? Do redirect, cache and proxy still work with web applications?

JR: Presentation is on 28 March

<Larry> slide 8: "how we see ourselves" "how we saw ourselves"

HST: I will not be on the call next week

AM: Wrt Privacy
... There's been a lot of discussion of this on the IETF privacy mailing list
... There will be people there who know a lot about this -- more than we do, pbly

HST: Happy to convey that we are the junior partners in this
... need IETF help

LM: The TAG is tracking more than leading
... W3C is running workshops

LM: We're asking for help in some cases, putative authority in others

LM: In contrast, slide 15 is our lead

<Larry> "a mess" isn't very informative

AM: What's the polite way of saying that?

LM: We have a work in progress, which tries to move this forward, so not "a mess"
... This is an example of an evolution point

<Larry> it might be a mess, but it is natural

LM: Lead with W3C priorities, put TAG's second on slide 10

HST: Not sure

LM: Top-level goal is improving IETF/W3C engagement
... so the W3C goals are the highest-level agenda-setters

JR: HST, are we done?

HST: Yes

LM: This is thought-provoking, which is just right
... We can use this to organise how we think about organising our Web App arch. work -- it gave me a new perspective -- anyone else?

JAR: Yes, that makes sense

IETF meeting on registries

LM: MNot noticed a change proposal from Mike Smith wrt content types for <canvas> in HTML5, which proposes a registry
... Is the W3C gearing up to run more registries?
... Is this a way we should go for extensible vocabularies?
... There was a reason IANA moved registration management from one person to a 'political' process
... Such tasks shouldn't be taken on lightly

LM: When we have web-based protocols that need an extensible vocabulary of parameters
... and looking at ISOC's sponsorship of W3C
... IETF, ICANN and IANA are independent organisations -- is there coordination needed here?
... Should W3C stumble in to running registries -- has the membership committed to resourcing the indefinite provisioning of this service?

JR: In the IETF case, I thought new registry entries were declared by RFCs. . .

LM: Not always. IANA has a contract to perform registry services, under the direction of IETF.
... So if the IETF publishes an RFC which creates a registry, it has to specify how registrations are managed
... It can be first-come, first-served, or managed by IETF, or devolved in part to other organizations
... In some cases there is an appointed expert reviewer or panel of reviewers, e.g. Graeme Kline for URI schemes
... But the RFC that covers URI scheme registration is being revised to accommodate IRIs

LM: Registry steward looks after fairness, safety, ...

HST: So life is complex, these things ramify.

HST: Just because the XPointer scheme registry is simple to operate doesn't mean it's always that way

LM: The ownership of the text/html media type semantics is an example of why the process matters
... Mostly it doesn't matter, but when it does, there has to be a clear story
... Sniffing isn't disconnected from this either
... Not sure W3C has taken on board all the potential complexity of running a registry

YL: Consider image/svg+xml took a long time to be defined, only officially registered a few months ago, but successfully in use for years

<DKA> I share your concerns, Larry.

YL: Move to have everything defined by RFC is not necessarily helpful

LM: There have been gaps in the processes, that needs to be resolved

[Jeni Tennison joins the call]

HST: Thinks TLR will be in Prague

LM: It would be good if someone from W3C staff who is up to speed on registry issues was at [some meeting]
... What to do about the now-rejected link relation registry
... Anyone from HTML WG at the IETF meeting?

YL: I will check

HST: LM should maybe brief TLR

ACTION Larry to liaise with Thomas Roessler about the registries issue background

<trackbot> Created ACTION-539 - Liaise with Thomas Roessler about the registries issue background [on Larry Masinter - due 2011-03-24].

ACTION Larry to try to arrange for Thomas Roessler to participate in the meeting about Registries at the IETF meeting in Prague

<trackbot> Created ACTION-540 - Try to arrange for Thomas Roessler to participate in the meeting about Registries at the IETF meeting in Prague [on Larry Masinter - due 2011-03-24].

Copyright and deep linking

[Peter Linss joins the call]

JR: Links in the agenda for the background
... including discussion with Thinh Nguyen in December 2010: http://www.w3.org/2001/tag/2010/12/02-minutes.html#item01
... DKA, what about ACTION-505?


<trackbot> ACTION-505 -- Daniel Appelquist to start a document wrt issue-25 -- due 2011-03-01 -- OPEN

<trackbot> http://www.w3.org/2001/tag/group/track/actions/505

DKA: Shell document exists, needs content
... What is the audience, what is the point?
... The recent legal issue has put some energy behind it
... Guidance for a court? Focussed on difference between link and transclusion?
... That needs action on people to contribute content
... I can do some, but not all

AM: Thanks DKA
... Typically the TAG writes on technical stuff
... this is not quite technical
... So what can we write, and for whom? We are not lawyers. . .
... Where is the TAG in this difficult controversial situation?

JT: I'm trying to draft something, as an aid to thinking this through
... We can contribute some terminology: how information moves, by fetching, caching, etc.
... And what happens with it: linked, transcluded, etc.
... That could then be used and referred to be the people involved in the legal discussion

JT: We could also give guidance/good practice to webmasters about putting acknowledgements in to pages etc.

<Larry> I want it to move through "recommendation" stage, and turn into a W3C (and IETF?) consensus document, to give more weight to it than just "TAG as another group of experts"

HST: Last autumn I finished teaching a new undergraduate course, and in the process of tidying up the public-facing version of the coursenotes I realised that I wasn't sure exactly what the right way to handle images, audio and video that I had shown the students. I went looking for guidance on precisely the transclusion vs. explicit linking issue in this regard, and found nothing. I even got into fairly obscure details, such as the fact that although you might think using hovertext to credit a source was a good idea, it doesn't work if you eventually publish via PDF, because the hovertext is lost. So I am very much in favour of pulling together some Good Practices guidelines
... My only quibble with JT's plan would be that in many cases, as in my example, there is no webmaster involved, the scale is too large for them to keep track, as many users self-publish within an instituational server framework.

<JeniT> Sorry, by webmaster/web developer I meant author

LM: To be useful legally, but w/o legal opinions -- maybe we should look at existing expert testimony
... to get some guidance as to what might be useful

<JeniT> Do we know where to find those?

LM: I feel pretty strongly that we need to take this through broader review, by putting it on the REC track
... so it gets community review

JR: That's what Thinh said
... The minutes of that meeting are very useful


<Yves> http://www.w3.org/2001/tag/2010/12/02-minutes.html

JR: How do we coordinate with the rest of W3C?

AM: Maybe speak to Danny Weitzner?

YL: May be too busy, in gov't these days

JR: I can talk to Hal Abelson. . .
... Maybe the first thing is to let DKA and JL get something written and that will let us
... get started

DKA: With respect to what can/should we be saying, I like JT's suggestion that we start with terminology
... aimed at informing the legal community
... After the conversation with Thinh, I thought we had consensus on a bit more than that
... In particular, something that is already out there in our Deep Linking finding:

Attempts at the public-policy level to limit the usage, transmission and publication of URIs at the policy level are inappropriate and based on a misunderstanding of the Web's architecture. Attempts to control access to the resources identified by URIs are entirely appropriate and well-supported by the Web technology.

DKA: Documenting the parts of WebArch that support that proposition are what JT is suggesting

<JeniT> yes :)

HST, AM, JAR: +1

<Larry> i would like to separate out the opinion part from the definition and architectural part, even in separate documents

ACTION Jeni helped by DKA to produce a first draft of terminology about (deep-)linking etc.

<trackbot> Created ACTION-541 - Helped by DKA to produce a first draft of terminology about (deep-)linking etc. [on Jeni Tennison - due 2011-03-24].

ACTION-541 due 2011-03-30

<trackbot> ACTION-541 Helped by DKA to produce a first draft of terminology about (deep-)linking etc. due date now 2011-03-30

HST: REC track gives us all the coordination we need

LM: We haven't done many REC-track documents -- we might want to work harder than W3C Process requires at the early stage
... to let people know what we're doing
... Part of that would be to solicit additional material

JR: Maybe see that as the doing the equivalent of chartering

<jar> well not exactly..

LM: We have to be careful about describing what we think we are doing

<JeniT> :)

LM: For the time being, that's a pointer to some requirements on the Introduction to the document being drafted

JR: No-one wants to give legal advice, which is one reason why there is no guidance wrt HST's problem
... THere are at least some non-legal issues, such as giving credit (as opposed to licensing)

JR: where some advice could be given w/o serious repercussions

<Larry> we want to give advice which is useful in a legal context, but doesn't itself make legal recommendations, since the technical issues are balanced against societal and financial ones to come to a conclusion about what is or should be legal or not legal

JT: So, aim to talk more about being a good web citizen/being responsible
... rather than making any legal claims

That fits with giving credit

<jar> +1 good citizen

<Zakim> Larry, you wanted to argue against 'good citizen'

<Ashok> But will the good practice protect you legally?

LM: Balancing the technical facts versus societal goals
... A lot of societal goals are mixed in here, and they are much harder to give advice about
... than getting the facts clear

LM: In particular there are access-control mechanisms, say passwords, by which material can be
... barred to some and allowed to others
... Then you say something about conventions for using such mechanisms
... Those are facts

HST: Asks a complicated question about what "give me credit" really means
... as a way of asking how we could safely give guidance on how to give credit

<jar> http://creativecommons.org/licenses/by-sa/3.0/legalcode

JR: I was thinking more along the lines of what the form of a credit notice should be, a la Chicago Manual of Style, in a social context such as academia
... Even if something is in the public domain, you can still credit someone

JT: I'll work with DKA and we'll get something out

various: Thank you JR for chairing

<JeniT> +1

<DKA> thx!

JR: Adjourned

<DKA> +1

<DKA> +1 to great chairing and organizing, JAR

Minutes formatted by David Booth's scribe.perl version 1.135 (CVS log)
$Date: 2011/03/23 15:12:39 $