Informal TAG Teleconference of 15 January 2013

15 Jan 2013


Norm Walsh, John Kemp, Marcos Caceres, Yehuda Katz, Alex Russell, Yves Lafon, Jonathan Rees, Noah Mendelsohn, Ashok Malhotra, Larry Masinter, Peter Linss
Tim Berners-Lee, Anne van Kesteren, Jeni Tennison, Henry Thompson
Noah Mendelsohn
Norm Walsh


Getting acquainted and ideas for refocusing the TAG

Noah: a very warm welcome to our four new TAG members. Although you don't officially join us until Feb, you are strongly encouraged to participate in all calls and other activities (except our rare formal votes) immediately.
... We'll go around giving everyone a few minutes to introduce themselves and give ideas about the TAG. Suggest you consider the questions at http://lists.w3.org/Archives/Public/www-tag/2013Jan/0026.html (reads them)
... Let's alternate, starting with a new member. Yehuda?

Yehuda: Actually, I'd prefer to hear some other responses first.

Noah: Sure, Jonathan?

Jonathan: I've been scribbling some ideas. I'm leaving the TAG. It's been interesting if not as productive as we might have liked.
... I got into the TAG from the semantic web angle, though I've had some more hard-core experience wiht standards than that. I worked on the scheme standard and kicked around the computer biz for a while.
... I've been struggling to figure out how to better integrate the declarative/RDF perspective with the more operational approach that you associate with web standards. I don't know how successful that's been; seems like not so many people interested.

<wycats> I can

Jonathan: It's been really interesting and intellectually challenging. We've been dealing with interesting, hard questions
... If we never published a finding on how language/versioning should work, for example, I think there was a good excuse. But I don't think the time was wasted, it's a hard problem.
... We flip back and forth between things that are important and things that are more urgent and I think it has to be seen as a sort of portfolio. Those are different kinds of problems and I don't think we should neglect either track.

<Ashok> Jonathan, your formalism on versioning was really good -- I'm sorry we did not follow up on that

Jonathan: Maybe even trying to divide the meeting time up on those two tracks would be a good idea.
... Someone has to look at the broader, architectural issues. Even if we can't come to concrete conclusions, we should try to write more and summarize discussions. Come up with useful products of discussions on hard problems.
... It seems like we ought to be able to do something about IRIs/IETF coordination.
... It's the TAG's responsibility to reach out and find the gaps.

Noah: Generally figuring out how we can have impact is a problem. Alex, I think you have to go soon, let's move to you.

Alex: The questions seem pretty deep. I think I want to see the TAG focus on are areas where it can be a better steward of existing work.
... There are a lot of areas where having a broad view can create opportunities to alleviate or illuminate tension between specs or layers of the platform.
... I think those are a good fit for the TAGs mission.
... Specifically, I think the TAG could do a better job of coming up with concrete advice for developers of JavaScript APIs and declarative APIs. Help uncover layering in the platform we're shipping.
... To get developers if not deep into JavaScript, then ready to [scribe lost thread]
... I'd like to do outreach to developers to come up with ideas about where they think there are gaps that we might help fill.
... I think I was saying that the TAG should be responsive to web developers. That seems like a good fit and we should work concretely to do that, through some measurable goals. We can use that to help prioritize our goals.

John: I used to work for Nokia when I was on the TAG a few years ago. I has only one term, and I would say that one of the reasons I thought very hard about runnign for re-election was TAG effectiveness and my role in that.
... I work for ESPN now and some of the issues are still the same. I was particularly interested in security, which I still am. In particular, the developement of origin and cookies and so on. I brought these personal interests to the TAG, and I was encourage to work on them, and despite all the discussion we had, I feel like we had almost no impact on those areas.
... Areas where we did have success...at least during my tenure...are in things like commenting on the HTML5 specification, which weren't all accepted, but some did have impact, and I found that a useful exercise.
... I think that leads me to a general sense that finding areas where you can work together, as the whole TAG or in groups, and write something that influences a community, that's a very specific place where the TAG has value.
... Finding those areas of agreement is tough, because people have different backgrounds and interests.
... I found it very useful in my day job to use TAG findings an WebArch as a way to influence people in my company to do things one way rather than another. Areas where you can explain things like that to ordinary developers are very important.
... I worry a little bit that it is becoming the case that the constituencies that are developing the web are becoming fragmented and contentious. I don't know if it's really an issue. I do find the stuff that the TAG has published are useful. Just having public conversations about technical matters is a really excellent thing.
... It would be good to maintain a sense of history about the web, where we've been and where we're going.
... Part of that involves collaborating or influencing other global organizations involved in the network. Persistence of URIs and content, things like that that go beyond the web and head into the social or legal space can't be ignored. The opportunity to have impact over those is quite different than making technical comments on a spec or educating we developers.

Marcos: It's hard to capture in a sentence what we're going to over 2-4 years.
... I agree and sympathize about the different range of issues that the TAG could tackle: from the technical to the social.
... The scope is super broad. We should try to limit the scope, ideally try to make the TAG of real value to the developer communities. Particularly web developers working with JavaScript and so on.
... At the same time, I'm not losing sight of things that are architecturally fundamental.
... And not completely losing focus on communities like RDF that are playing significant roles in areas like government data.
... How do we know if we're successful? By having impact on the developer community could be measured
... Not restricting ourselves to publishing findings in the W3C, perhaps, but also writing articles that target specific audiences in other forums.

<noah> Great idea. We've talked a bit about targeting other venues/media, but Marcos is correctly pointing out that we haven't thought nearly creatively enough about it.

Marcos: Those things are useful in terms of having a presence online. We can have a 1:1 communication channel with developers could be really great.

<wycats> noah: having some new members with direct channels to developers should help with that

Marcos: Some of the new TAG members do have things that they want to work on. For example, updating WebArch or focusing a bit more on the layering between HTML/JavaScript/WebIDL and how these layers interact.
... How all these layers interact seems a bit broken in places and that's frustrating for developers.
... Expanding WebArch and avoiding mistakes that were made in the past. My personal opinion is that it's almost as if the TAG takes one view of the problem space when there are other views.
... It would be interesting to have that other world-view is expressed. Having multiple perspectives about what the web is would be valuable, not necessarily saying that one is right or better.
... It's not that they don't work together, they all do, but I don't think it's been articulated
... I think that WebArch helped a lot of organizations understand at least what one perspective of the web is.
... My background is that I've been doing web stuff since 1996. I did project managment at Opera developing their extensions platform. I've been involved in standards for about five or six years.

Noah: Ashok, could you go next?

Ashok: I work for Oracle; I've been on the TAG for five years. I've published one finding on application state.
... The other thing I was hoping to work on were offline apps and offline storage.
... I think what the TAG ought to be doing is update the architecture document particularly with regard to webaps. There are lots of threads there that have to be coordinated.
... I think that's the TAG's highest priority.

Noah: Yehuda?

<masinter> what I would say I'll just type in

Yehuda: I've been a web developer long enough now to feel like I have a good handle on web applications development
... I did a lot of work on Rails in tandem with my work on jQuery.
... Most of my work on jQuery has been improving the DOM APIs. More recently, ember.js is my day job, an effort to bring higher level abstractions to the web.
... I think jQuery has fullfiled the goal of being a better DOM API, but I think there's a lot of room to grow there.
... I understand why a lot of web developers are frustrated by the W3C.
... I don't agree with a lot of those arguments, but I do understand them.
... A lot of web developers don't have a good sense of the consensus process, but on the flip side of that, a lot of web developers have a lot of pain points. A lot of important use case collection and collecting pain points is important.
... My main mission is to get a sense of architecture for the web that is more layered. I think it's important to have a declarative API for the web.

<noah> I strongly, strongly agree with what Yehuda is saying about the declarative/imperative boundary and what gets thrown under the bus too soon.

Yehuda: But I think that it's bad that from moving documents to something with a little interaction interferes with a lot of useful abstractions like crawlers.
... Today, to build something nicer, you have to give up on good searchable, crawlable URLs and that's a problem.

<noah> What I heard Yehuda say is: very important to have pure declarative for things like documents. What's really important is that as soon as you need a bit of interactivity, you too quickly move away from declarative and so declarative "gets thrown under the bus too soon". That's what I'm agreeing with [Noah]

<wycats> noah: you heard me correctly

Yehuda: I think that trying to look and see what the web will look like in 10 years is pretty hard. But that architectural goal that we all share is under attack right now because the declarative side hasn't kept up. There should be a declarative part of what the web is doing, but make it so that just adding a little bit of interactivity or a lot of interactivity can do it without abandoning declarative markup.
... We don't end up with a nice tower of abstractions that we should. It's more about the high-level architecture and less about the specific technologies.

Noah: Norm?

Norm: I'm Norm Walsh. I work for MarkLogic, I was elected to the TAG for several terms early on. I continue to focus on areas of markup, XML mostly, which isn't as fashionable as it once was.

I think the TAG has done a lot of good work. Some of it not as obviously cohesive as one might have liked. Norm: Norm: I think the web architecture document does a reasonable job of describing a mostly static web. Finding some way to articulate a broad set of principles for a web-applications web at the same architectural level would be a valuable.

<noah> Note that Norm is a former TAG member, but was on the TAG for a long time, and is the only one on the call who was on the TAG when Web Arch was written. (I joined the last few weeks)

Norm: Yves?

Yves: I'm pretty sure most of you are aware of W3C processes; but if you're not, ask me.
... My background is more on protocols. Part of the http-bis effort.
... I'm more interested in protocols than high-level stuff.

<wycats> Sorry if I was cutting out. I will try to get a better microphone/setup for next meeting.

Yves: I don't know exactly what the TAG should do, it depends on what the members of the TAG want to work on. It's hard to make progress on things that you don't have interested people to work on.
... I don't have a clear thought on what we'll achieve in the next six months.

<masinter> and me

<masinter> if you want my background, see: http://larry.masinter.net

<masinter> If you want a quick summary: "I know where the web design bodies are buried"

Noah: Peter

Peter: I work for HP and have been on the TAG for a term. I'm also involved in CSS and other groups. I was involved in the original gecko layout engine.
... I agree mostly with Alex and Yehuda.

<noah> Do you have ideas for how to connect better with WGs?

Peter: I think the unspoken issue is that the TAG is disconnected from other W3C Working Groups.
... I want to see the TAG engage more with the working groups 1:1, give them high-level architectural advice that they're not getting right now.

Noah: Ok, my turn. I wear a dual hat. I'm the co-chair but also a member. Usually, I try to be very careful not to mix the too roles so as not to unfairly advantage my own positions. Just today, I'm not going to bother separating, so the following is a mix:

Noah: I've been doing software work since the 1960s. I worked in the industry officially mstly for IBM, but also part of that time on campus doing research at Stanford and MIT. I did go to Lotus in the early 90's. That's relevant, I think, because shipping spreadsheet software for huge numbers of people is not unlike what the browser venders experience now. I worked on XML at W3C and on JavaBeans. I'm now teaching computer science at Tufts part time.

Noah: One of my favorite comparisons is with the phone system. In 1923, they invented the phone number system that we're still using approximately 100 years later. I think, painful as it is, I see part of the TAGs role is to help the community build a Web that will be viable in growing in 50 years, and maybe 100.

<wycats> I would be shocked if URIs didn't exist in 20 years

Noah: But I feel the tension of needing to ship software early and often. The challenge is, how to be careful enough to keep things clean for the long term, while growing the system fast, staying connected to the issues arising in real implementations, etc.
... I think that's a really tough balance and we have to think hard about it. Marcos said it's hard to capture in a sentence what we're doing. I don't want to rush it, but I think it's useful in answering questions about scope and priority. I have an example of such a sentence. Probably not quite right, but something like:

"Ensure that the Web as architected and deployed will scale for global use, and over many decades."

<wycats> we should not

Noah: At worst it's a bit of motherhood; but statements like this really can help a group to decide what it's about.
... To a greater extent that is visible from the outside, the TAG is it's members. What we succeed at tends to be bounded by what our members are good at and willing to work hard on.
... We can and should draw on the rest of the community; but we haven't worked on security because no one's elected security experts.

<wycats> I doubt anyone thinks we should spend all our time working on JS. We should focus on doing whatever we can to ensure that URIs survive. Today, that means fending off the desire that people have to build "applications" and deploy them using web infrastructure.

Noah: I also think our connection with Tim is very important.
... One reason is that I think Tim did something remarkable in the design of the web. His goal was to help the physicists at CERN share physics papers, but he put in abstractions that have scaled pretty well so far.

<slightlyoff> what wycats said.

<slightlyoff> (sorry, on a bus)

<slightlyoff> to a great extent I don't see tension in shipping software vs. good architecture

Noah: The W3C as a whole is a resource and a challenge and we can leverage Tim in that area.

<slightlyoff> and if there is tension, it's not represtented by the folks in the call/room

<slightlyoff> I hope we build something worth having, something that matters, and use our priviledged perch to help ensure that it's something that'll last...and that means having concrete opinions about what's working now, on a short timescale

<wycats> fwiw: slightlyoff and I are both on TC39, which follows a similar model. We are both strongly in favor of this model, I believe.

Noah: From my experience as chair, the TAG and the W3C tend to work by consensus. I'd like to continue to do that. We can spend some time experimenting with projects that don't have consensus, but ultimately we need to get the whole group behind what we're doing.

<slightlyoff> TC39 does as well

Noah: We try to figure out when the TAG is ready to speak as the TAG.

<masinter> i have some things queued up to type in :)

Noah: The last thing I'd say is that I'd like to make a lot of room for discussion of new things, but we'll eventually have to figure out whether to continue or abandon each of our existing project and open issues.

<wycats> wycatd

<wycats> wycats

<wycats> it was me

Yehuda: Alex and I are members of TC39 and it works in the same consensus process; I'm happy to hear that we'll be taking time to achieve consensus.

<masinter> * more background: I think I was webmaster for the second commercial web site (Xerox) and of the site first web application (PARC map browser)

<masinter> * My recommendations on what the TAG should do: http://masinter.blogspot.com/2012/12/reinventing-w3c-tag.html please give Recommendation few moments consideration & discussion

<masinter> * I think the top problems facing developers aren't the ones they think they have: governance & security

<masinter> - governance: http://masinter.blogspot.com/2012/12/governance-and-web-standards.html

<masinter> - security: http://masinter.blogspot.com/2012/12/web-standards-and-security.html

<masinter> * My main concern: continue or wrap up Publishing & Linking

<masinter> * personally: still want to pursue link between situational semantics & security models, but don't think it's TAG work

Noah: I think we should be socializing our personal interests even more, we need to find out what we and the community have consensus to be working on.

<wycats> I see a lot of empty lines

<wycats> is that correct?

<noah> http://www.w3.org/2001/tag/products/

Noah: I don't want to talk about process too much, but let me say a few words about them.
... The products link is a "public dashboard" that I try to keep up-to-date.

<wycats> that sounds fine with me

Noah: We use the W3C "tracker", especially its action system.

<masinter> main summary: read my 're-inventing the TAG', and give each Recommendation 5 minutes of thought & discussion.

Noah: That's important because that's how I keep track of who's doing what and when it's due.
... Please do read at least the section on actions in the "how to participate in the TAG" document
... In the products page: work on fragment identifiers, work on publishing and link, and URI documentation and discovery. That's basically in the space of how do I get metadata about a URI. Is it ok if the URI is for something that isn't sort of documenty. It's partly a semweb thing.

<wycats> how much work is left to do on those?

Noah: Those have been our top priorities and I think the first two at least are ones we'll want to continue driving forward.

<wycats> clearly TAG would not create the spec

<wycats> the WebIDL spec

<wycats> is it in the purview of TAG to have opinions about WebIDL

Noah: I think the question I heard was: there are some issues around WebIDL, we wouldn't write the spec but is helping to resolve issues around it in scope?
... I don't think there's a closed-form answer because it depends on how you answer my earlier question.

<wycats> there are ways in which that is true

<wycats> that was a useful answer, thank you

Noah: If WebIDL is doing something that violates princples, then it's a high priority. We can help otherwise in a best effort basis.

Jonathan: If we wanted to deal more with applications, then WebIDL would seem like an important piece.

<wycats> I don't mean that TAG should care about specific interfaces

<wycats> I mean answering questions like "what is WebIDL for in terms of the broader web architecture"

<noah> One example TAG question would be: what are the pros/cons of strongly typed interfaces both on the wire and/or locally in say JavaScript interfaces

<wycats> ok awesome

<wycats> thanks so much

Noah: The TAG is us, you get to help us change our opinions on issues like that.
... Chair's hat on, we need to figure out how to answer questions like that. We may have to iterate some.
... Let me work on logistics a little bit. I don't want to make formal decisions about anything, but I would like to talk a little bit about the f2f.
... Usually we schedule f2f meetings at f2f meetings, we look about six months out and try to find a week when we can all be in the same place.

<wycats> jQuery Foundation might be able to help with that

<wycats> we already piggy-back Board meetings on jQueryConf

Scribe isn't planning to record casual discussion of scheduling in great detail

<wycats> Yves: sweet

<wycats> I can make it

<wycats> I'm 90-10

<wycats> a plane is taking off?

<wycats> I am in CA!

<wycats> I can live with going to Europe with enough notice

<wycats> fwiw: I will have to request additional funding from jQuery if most meetings are in Europe

<wycats> it looks likely that I can get the funding

Noah: I'm tempted to propose London or Edinburgh.

<wycats> wrapping up sounds good to me

Noah: I'll try to get that firmed up in the next two weeks.

<masinter> would like some feedback on P&L

Noah: Sounds like we're wrapping up.

<masinter> would like feedback on publishing and linking, as to what the TAG wnats to do with it

<masinter> ok

<masinter> thanks, that's good feedback

<wycats> :D

Noah: Thank you all for joining; warm welcome to the new members. We're adjourned

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.134 (CVS log)
$Date: 2013-01-17 23:27:05 $