TPAC 2008 slide. Photo Credit: Kaz Ashimura
More photos of TPAC

W3C TPAC 2008

22 October 2008

See also: Agenda, IRC log

SteveB: Welcome ... this an exciting time in W3C and the Web ... we have lots of great activities, competition
... we are motivated to find interoperable solutions ... there are now 75 groups at W3C ... the most ever
... over 1500 people in WGs ... not including 'public IEs' ... 35000 people subscribed to mailing lists ... 40 groups here this week
... next year TPAC will be held in Bay area
... now welcome Tim Berners-Lee

<glazou> hey, 350 laptops that's too much for Pullman Hotels :-)

Opening Keynote: "Cleaning up the Web"

TimBL keynote. Photo Credit: Kaz Ashimura

Tim Berners-Lee:Now we have the Web, and we realise there are more mantras. Internet was built on top of telephony, and the Web on top of Internet. Every layer should be a layer on top of another. What we learned in "Web kindergarten" is being added by the TAG.

<ChrisL> http://www.peace.ca/kindergarten.htm : "ALL I REALLY NEED TO KNOW I LEARNED IN KINDERGARTEN" by Robert Fulghum

<ChrisL> http://en.wikipedia.org/wiki/All_I_Really_Need_to_Know_I_Learned_in_Kindergarten

Tim Berners-Lee:And mantras and aphorisms are being added. Most of them are SHOULDs. There are other things we learn ... to listen to people. The listening you are doing now is passive, and therefore not very good. Active listening would involve more interaction until you understood what I meant... and a group culture would allow us to work together, but would simultaneously be a barrier for people from the outside.

Tim Berners-Lee: One of the tensions has been the broken markup that is out there... broken servers. To oversimplify, some people worry that it isn't optimal for the browsers to fix things up. Meanwhile, the HTML WG points out there is a need to look at broken stuff ... and browsers compete in this area. The aphorism here is "specify what goes on the wire, and not what processors do" and "liberal in what you accept, and conservative in what you produce". The original 3 pages HTML spec says ignore unknown tags to allow for extensibility.

<anne> Postel's law afaict: http://en.wikipedia.org/wiki/Robustness_Principle

Tim Berners-Lee: Mail servers didn't get worse because errors got fixed, but that is different with browsers. People don't fix broken markup for stuff from long ago. It caused a race to the bottom in the browser world... the motivating factor for putting stuff on web... a browser rewards a designer for putting stuff up.

<ted> remote (_not_ local) attendees are welcome to tune into audio broadcast http://media.w3.org:8000/tpac-2008.ogg

<ted> audio broadcast available provided we consistent electricity...

Tim Berners-Lee: You don't get rewarded until you've got it exactly right in the non-fixup world. Validators work by liking you to have everything right. The validator will reject you until you get it all right. Using a validator, you are only motivated if you want correctness.

<anne> without a DOCTYPE you get quirks mode, not standards mode :p

Tim Berners-Lee: In general, using a validator is too hard. A browser is easier to keep happy ... can we fix that? Maybe not with a yes/no, but with a score, and ideas on how to raise the score ... For example, 46/100 is your score.

Tim Berners-Lee:People tell me that browser manufacturers don't want to give users errors. But suppose the browser knew which were my pages... and only told me of errors on those. "View source" ought to give advice 'this is the cleaned up version'... so that adopting others' code will always improve the quality

<ChrisL> users more likely to be impressed by better functionality from a clean page; an arbitrary score gives them no business benefit

<anne> solving the halting problem might actually be more difficult than rocket science, dunno :)

Tim Berners-Lee: "Save as" should correct the code, too. And servers could correct it as well. Apache goes out improperly configured to deliver unknown filetypes as text/plain. This may be my fault from the original CERN server, but server manufacturers, please don't assume a content type.

<gsnedders> Is an extension being told?

Tim Berners-Lee: Have all content types for standard filetypes that we already know about, in W3C recs. Adding new content types is a major issue for adding new applications on the Web. Just give a 500 if you don't know the content type, so you spot it and can fix it.

<gsnedders> Can't they just send no content-type header?

<hendry> mime-support: /etc/mime.types # file a bug on that package in Debian for new mimes and all apps should use them

Tim Berners-Lee: Some people suggest that Apache could tidy what gets served.

<marcos> Roy wants to dump HTTP in the new version of apache, iirc

<hsivonen> gsnedders, Apache could but they don't ATM by default

<arun> marcos, what do you mean 'dump HTTP'? And who's the Roy that you refer to?

Tim Berners-Lee: So why do this? While the browsers will go on accepting invalid markup... Well, we are wading waist deep in these errors, and as things are going in 10 years we will be shoulder deep ... so this is not for us but for the future

<ChrisL> "think of the children"

Tim Berners-Lee: We built the Web on top of the internet. The interface to sockets hasn't changed ... it is a clean interface.

<marcos> Arun, roy fielding. He came up with a better alternative to http)

<arun> Marcos, aaah, Roy Fielding of REST fame.

<marcos> Arun, of Apache fame :)

Tim Berners-Lee: Http is a layer on that. The Web is a function that takes a URI and delivers content/meaning.

<arun> Marcos, well, and HTTP too for that matter

<marcos> Waka, yes

<arun> http://en.wikipedia.org/wiki/Waka_(protocol)

TPAC setting. Photo Credit: Kaz Ashimura

Tim Berners-Lee: If we can work towards a future where people produce clean pages, it will make the future web easier to use. It is about being simple, realizing that the future is longer than the past.

<marcos> This is silly, there is little correlation between markup and DOM. Create a clean DOM, forget about clean markup.

Tim Berners-Lee: 15 years of legacy pages will be as nothing soon

<gsnedders> 15 years ago? I was only just one year old then! I don't remember 15 years ago!

<arun> LOL gsnedders

Tim Berners-Lee: We are celebrating 20 years of the first memo about the Web in March 09 ... that doesn't seem long ago, but some of you weren't even born then. People say "I understand that there is incredible talent in producing the HTML spec more modular" and several have some up in the last few days to volunteer.

Tim Berners-Lee: Thanks for listening ! So, question: should we, can we, clean up the Web?

Steve Bratt: We'll go with the questions, then panel ... please be brief.

Ian Hickson, Google: A lot of what Tim asks for, such as "save as" is already there ... CSS errors, some HTML errors

Tim Berners-Lee: By default?

Ian Hickson, Google: Yes

Tim Berners-Lee: I want to see that.

Ann Bassetti, Boeing: Where are the slides?

Tim Berners-Lee: The notes I used I will post

<Rotan> Search for "score at the end of validation" at http://www.w3.org/2004/03/plenary-minutes

<Rotan> The idea of giving feedback for the quality of Web pages, as a byproduct of validation has been thought of many times before. I remembered suggesting this 4 years ago.

Web Architecture, Blueprint or Recipe

TAG Panel. Photo Credit: Kaz Ashimura

Steve Bratt: Thanks to Chris Lilley for putting together the program

Chris Lilley, W3C: Why this panel? There are a lot of discussions around Web architecture. It doesn't impose many many things, it encourage to do things. Where are we today? What needs to be done today?

<IanJ> Architecture of the World Wide Web, Volume One

Chris Lilley, W3C: I invite each speakers to introduce themselves and speak about the topic. The goal is to have a discussion more than a presentation.

Anne Van Kesteren, Opera: Working for Opera Software, involved in HTML 5 ... there are some clashes with Web architecture.

Norm Walsh, Henri Sivonen, Anne van Kesteren. Photo Credit: Kaz Ashimura

Henri Sivonen: I'm consulting for Mozilla. I see the Web as the public Web that people can access. The resources you can navigate publicly. I define Web as the information space accessible to the public via a browser.
If a mobile operator operates behind walls, this is not part of the Web. If you use Web service on http but not using the Web browser. I don't see it as part of the web. Same for Semantic Web ... Web Architecture are not principles for the browsable Web

Norman Walsh, Mark Logic: I gave a second look at the Web architecture document. I still stand behind the document. Many of the principles are SHOULD ... because there might be good reasons to do in other ways

Larry Masinter, Adobe: I'm not part of the TAG, but was on the AB when the TAG was being chartered.
Part of my observations: what's different between the Web of Today, and the Web Circa 1996; ... the Web has really changed
... and what people think about the Web has changed. The W3C has not been leading but going off in another direction. Some corrections are necessary.
In 1996 it was about documents, but now there are plenty of applications with ajax, video, etc. In 1996 it was mostly done by hand, but most content produced now is through cms, wiki, blogs, etc.... in database
The Web 96 is about publishing information and the Web today is about interaction, social network. There are hundreds of organizations developing standards, specifications now. The focus of W3C on XML, semantic Web, Web services, and narrow industries... It doesn't seem to match what the Web is. Here is a quote from B. Disraeli that comes to mind: "I'm their leader, so I must follow them." The focus of Working groups are not really about leading the web ... Instant messaging, video conference, etc. because people put these things on the web too. ... If you call for more resources, you would have to drop some of the activities. ... would it be very damaging if some of the things were dropped?

<olivier> [ for those not in mandelieu, note there is an audio broadcast available: http://www.w3.org/QA/2008/10/tpac2008-listen_discuss.html ] [for those in mandelieu, please listen to the live voice and don't torture our bandwidth any further, thank you]

Noah Mendelsohn, IBM: I'm working on the TAG. 1. What is the importance of doing such work on architecture?... 2. Is the tag effective doing that? It is important for the World Wide Web to have long term architecture. I have worked for Lotus, and there are a lot of constraints when you deal with millions of users and deployment. I have a lot of sympathy for HTML 5 people. It is not trivial. The Web has a lot of importance in a long term. Few of us would fly in a plane built by a group with mantra "ship early and ship often." It's easy to say I know my users, get out of my way. But how does it work on a long term. We have some stress between the TAG and some groups. Kludges have a long life; people in HTML 5 WG could be doing more interesting things if architecture had not been violated by kludges years ago. The compromises affect very often the people who made the compromise. The role of the TAG is to make sure that all concerns of the community are represented. I don't share Henri Sivonen's view. We can't just make decisions for small community... we have to include everyone.

<IanJ> Noah on content sniffing: If I want to publish an XML fragment as plain text (e.g., as part of a bug database), if browsers do content sniffing (by default, automatically), then I can't do that anymore.

Noah Mendelsohn, IBM: [sharing a mistake GET/POST in a development]
... I was able to point to the TAG doc URIs, Addressability, and the use of HTTP GET and POST
... to give the right thing to do.

Chris Lilley, W3C: Questions: How many people have looked at the architecture document?... How many people have looked at the TAG findings?... How many people disagree with the TAG document? How many people have told the tag they disagree?

<Norm> From here it was about 50/50 on whether anyone had read the architecture document or a TAG finding.

Chris Lilley, W3C: How the Web content is edited with regards to you.

<DanC_lap> polls: how many have read the arch doc? (most? 2/3rds?) how many have read a finding (half?) how many have found something they disagree with? (20ish) how many told the TAG? (19ish)

<Rotan> [many hands raised for the above questions. was not in a position to assess the percentages.]

Henri Sivonen: there are 4 top browsers.
... at a given point in time, the interoperable platform... The platform is expanding. The more of the browsers that implement a feature, the more it becomes part of the platform.

Larry Masinter, Adobe: I object to characterize the Web by only the 4 top browsers. (e.g., browsers in internet cafes in Africa). In Africa, in a cafe, you will find different version of the products, lot older and it is part of the Web.

<Steven> [applause]

Tim Berners-Lee: We have discussions about the scope of the Web
... on the TAG
... the TAG was created because WGs asked for it
... It was a service community
... initially, a lot of questions were back logs questions
... the tag has been doing writing on the things which are in between the specs
... and were not really defined somewhere
... The TAG also tried to connect groups together when they had conflicting reqs

Noah Mendelsohn, IBM: The value of one large network is greater of the value of two small networks (cf. Metcalfe's Law). If everybody ignores the GET/POST, we would not be able to write crawlers anymore. We can call Web whatever we want but we should look at use cases.

<raman> TimBL Should be asked the question "Define the Web" he did not answer it

<IanJ> Web arch definition of Web: "The World Wide Web (WWW, or simply Web) is an information space in which the items of interest, referred to as resources, are identified by global identifiers called Uniform Resource Identifiers (URI)."

Arun Ranganathan, Mozilla: I am willing to accept the proposition that the Web is larger than what browsers render. I agree with Noah's point about GET/POST, but a lot of popular APIs use GET for state-modifying operations. I don't agree with Noah's example of crawlers (those URIs are cordoned off). ...That is not a good point. We can't change the system. The arch doc provides good counsel, but when people build, e.g., bad APIs, you still have to deal with it.. you have to build in a way where you adjust to the system. I think that some of the panelists have conceded certain incompatibilities; I'd like more reconciliation. Some of the arch documents can be mutable over time.

<Rotan> Many of the "mobile" people treat the URI as an identifier with which one can obtain a *representation* of an underlying resource. Many people gloss over this "representation" idea.

Noah Mendelsohn, IBM: One of the rules the arch doc. can serve is as a guide of issues to raise before the community over time. If you are using GET the normal go for it. If you have a good reason to modify something, then ask for a change.

Larry Masinter, Adobe: The architecture document is not an implementation guide. A standard is not a description of the real world. You have to deal with indeed broken implementations, but that does not mean you should do an architecture which is broken... Architecture guidelines depend on the things which evolve in time.
The standard describes what should be. I think it is a fallacy that the standard also include the implementation guidelines; implementation advice changes over time. The URI standard does not have a limit on URI length; but every implementation has a limit. The implementations change over time; the standard needs to be stable. There was a separation between content, protocol, and reference. HTTP not tied to HTML, or vice versa. So if HTML were broken, maybe we could have a different type; the protocol allows that.

<Rotan> Sniffing to work around bugs is just one reason for sniffing, and something we wish we didn't have to do. But still want to sniff the context, to adapt to its particular needs.

Anne Van Kesteren, Opera: That's still true

Larry Masinter, Adobe: But it seems like every feature we want needs to go into HTML 5, the only content type.

Chris Lilley, W3C: That's not entirely true. Lots of people read html in email clients, for example

Larry Masinter, Adobe: We are breaking orthogonality (and making an architectural mistake) if we focus on one type of user agent.

Anne Van Kesteren, Opera: I don't think this is happening. There are different "classes of products" (a term used in the QA Framework).

Larry Masinter, Adobe: Does anybody here work for or with a mail user agent that interprets HTML?

<IanJ> (A number of hands go up)

Anne Van Kesteren, Opera: I do.

Henri Sivonen: On the topic of violating architecture. it doesn't really matter if any groups claim authority
... this WG doesn't have the power to make people stop doing wrong things. Emailing to the TAG and discussing right/wrong doesn't help. People still do things. People out there are still doing mistakes and you have to deal with that.

Chris Lilley, W3C: Not a question of authority...web arch says "if you do X, there are consequences." The consequences don't go away if people don't follow the advice; they multiple. Mechanism to stop transmission of state also stops other uses (e.g., can't share URIs with session id info in them).

Henri Sivonen: The purpose of standards is to document what needs to be implemented,so there will be interoperability...based on what is already out there. If there are n browsers, the standards should give guidance so that the n+1 browser can implement real world use case. Little point in writing a science fiction document that, if implemented, creates a piece of software that doesn't work. No point writing Science Fiction document. Standards should lower cost of implementation.

Noah Mendelsohn, IBM: We should have discussions (like role of text/plain) within the entire community.

Tim Berners-Lee: A protocol _is_ science fiction. You define a protocol so that nodes on the net can talk to each other (e.g., give me articles in a newsgroup). When you define the protocol,you define the sets of states that the nodes can go through, and the messages, and you go through some math.

<IanJ> (TBL goes on about usenet protocol)

Tim Berners-Lee: The point I'm making about protocol design is that it gives a set of rules, and it says "If you implement the rules, you get some properties." That's what a protocol specification does. IF you do X, THEN you get Y properties. You put forward your spec and say "does anybody want to play in this?". People put out protocols, and by "joining the club," you helped the system achieve the desired goals. That works. Now on the web we have a lot of legacy data, and it's valuable to document that. But we need to be careful about not losing the vision of protocol specification. Someone who shall remain nameless repeats [a statement from Wittgenstein's Philosophical Investigations] "Meaning [just] is use." But that's not the case with protocol....the protocol makes it fact. We should acknowledge that we should look back, but we should NOT define spec writing as descriptive. Spec-writing is prescriptive. If people don't follow the protocol in the future, we have the right to go after them and say "Excuse me, you joined the club and are expected to follow the rules."

TV Raman, Google: There was a point in this conversation when we were talking about the definition of the Web. Henri's characterization explains his reasoning. Larry said the web was more than what top browsers do. Arun gave some slack about that it might be more than what browsers do. Tim, can you say clearly what you think the Web is?

Tim Berners-Lee: I did both. We have to acknowledge the past, but we are, very importantly, designing for the future. We are designing, not documenting. Web is "humanity connected by technology" (the broader definition). Narrow definition: The function map, of URIs to meaning.

Larry Masinter, Adobe: I am more interested in what W3C should do operationally than a principled definition of the Web; and I think W3C should do more than just what browsers do. I prefer "humanity connected by technology, but with a focus on what has traditionally been the Web."

David Baron. Photo Credit: Kaz Ashimura

David Baron, Mozilla: I wanted to make a few points about how much the specs need to talk about what's actually out there. Browser vendors need this to keep cost of entry to enter and stay in the browser market as feasible.

Henri Sivonen: We can't work by ourselves ...there is so much reverse engineering to do, we can't do it alone. We need to pool our efforts in a standards body. We need to join and share our resources.

David Baron, Mozilla: I hope W3C is a forum for that type of standardization.

Norm Walsh, Mark Logic:I think that it's a healthy tension between real world/ world as we'd like it to be.

Ann Bassetti. Photo Credit: Kaz Ashimura

Ann Bassetti, Boeing: Old browsers are not just used in Africa; they are used in big companies, too. We are stuck with old browsers since apps were written that depend on those browsers. Three pieces of the pie: those who are defining architecture, those who are using the architecture. Interop is much more important than the next "whizz-bang" thing; we won't get there for a while.

Noah Mendelsohn, IBM: Some of my best successes in using webarch is in discussions with people in my company. In many cases, I've found these findings, etc. of great value in talking with implementers. (and not just spec writers)

Ann Bassetti, Boeing: Has TAG written a "top 10 things to do"? That might be helpful.

David Baron, Mozilla: For what it's worth, a lot of the reverse engineering I referred to is *about* improving interoperability.

Murray Maloney (unaffiliated): The Web is a gift. Among the things I learned in kindergarten is "polity." The Web is a gift. It's a paradigm-shifting change. The world has changed as a result in 20 years. We should be happy about contributing to that and making the world a better place. I'm a survivor of an airplane crash. 50 people died in the crash, in 1982. Sometimes, we don't need to rationalize too the browser vendors, or to the people in the higher echelons of the HTML WG why we need something....so if Boeing says "we need something" you should listen to them. ...it's really important that their systems work. It's really important that I can get to my bank account using my windows box...but the Web is broken. There's one Web, and _I_ am a Web browser. I've been reading markup since the 80s, and I can see in my head how it should look with a style sheet. Your browsers make it convenient for me, but it's not YOUR Web. It's everyone's Web. It's the Web of the people in Africa, or South America, or any place with a slow network that can't deal with loads of javascript. Please just make the damn thing work.

Anne Van Kesteren, Opera: I'd say that we are in violent agreement here. Part of the HTML 5 effort started in 2004 due to huge lack of interop among browsers. ...we want to improve interop by defining a common way of interpreting what's out there....this would solve some of the problems Boeing, and you, are facing.

Murray Maloney: My impression is that most of the browsers are doing lots of things wrong. I may have this wrong, but what HTML 5 seems to be doing, is codifying errors.

Henri Sivonen: If you are restricted to running IE6, then shouldn't we be doing just that in order to interoperate with your browsers?

<IanJ> [some applause]

Larry Masinter, Adobe: If you are going to document the past, write down how to write HTML that works ok in IE6. Help the millions of people who write HTML.

Tim Berners-Lee: When you define a protocol, you define the rules for each side: how to make a server work, and how to make a browser work. What you do is write the rules (perhaps in separate docs) and you demonstrate that stuff will work. I like specs where you can define things mathematically. The HTML 5 is particularly targeted at browser vendors, and I know that there has been a cry for aversion (e.g., generated) targeting authors. ...good point that there are a lot more of those [authors]....and I've heard movement in that direction to make HTML 5 two specs. ...let's ensure that when you put the two specs together, the system works.

Steve Bratt: Thank you, panel. Let's continue the dialog at W3C, this week in hallways.

<DanC_lap> +1 Karl. amazing to watch Ian decode TimBL in real-time.

<amy> we'd like to ask people to fill out a questionnaire on TP2009 - whether they'd likely be able to attend, stay at the hotel, and if charging a fee would be workable http://www.w3.org/2002/09/wbs/100/TPAC2009-attend/


Shawn Henry, W3C: Doing a demo session with a BOF session at lunch. ARIA stands for Accessible Rich Internet Applications. It works with AJAX and other technologies to provide information on name, role and state to assistive technologies. Assistive technologies help people with disabilities use the Web.

Steve Faulker: |showing JAWS (a popular screen reader) with a slider control.]

[JAWS speaking]

With non-ARIA slider, the listener is told that they are interacting with a button by pressing it. You can't press a button, you have to move it horizontally with a mouse.

[demo of a ARIA Slider control with JAWS]

Steve Faulker: JAWS announces the name of the control, directions to use it, and the values as it is moved.
... Example of a Tab Pane

[demo of non-ARIA tab with JAWS]

Steve Faulker: Without ARIA, the assistive technology doesn't know that the tab has focus.

[demo with ARIA]

Steve Faulker: Example of a tree navigation control

[demo without and with ARIA]

Steve Faulker: It is using attributes mapped to the API which is picked up by the AT and JAWS understands that the widget has a role and communicates the state.

Shawn Henry, W3C: There will be a BOF table at lunch where you can ask questions. Work on WAI-ARIA is done by the Protocols and Formats Working Group.

Impact of CMS on the Web

Shawn Henry, W3C: CMS systems represent a difference in the authoring practices of the past
... most CMS systems spit out dirty code
... this session: how development of CMS systems impact what we do here, the future, authoring guidelines, browsers, etc.

Jose Manuel Alonso, W3C/CTIC: [slides] I work at CTIC in Spain, fellow at W3C for eGov work. Today, i wear my institute's hat
... i will show you brief statistics on practice
... as i work on egov, the examples come from there mainly
... i ran some tools we have at the institute checking and describing sites of European governments
... sample: common portals of Europa
... most of the tools are very spread, custom developments, general cms systems
... it is a spread market
... in terms of code, it is better than expected
... validity of code is low, but better than, say, opera study
... we found a lot of different attributes that we did not know yet :-0
... some rss streams getting into the html content:-)
... we ran automatic tests on accessibility
... these are only automatic tests, very low level of accessibility
... it is improving but if you go to double-A (standard in Europe), most of the pages fail
... in terms of forms, they are doing better
... there is an improvement in code, lots more to be done in design
... repeating title is a typical case - only 20% use headings correctly.
... mobileOK tests were run
... we issue requests with the w3c user agent, no web site responded the way we expected. This has lots to do with the issue on the previous session. We think we have to fix this.
... the Web site managers the earlier they are trained on cms tools the better code they can produce.

Kai Scheppe, Deutche Telecom: [slides] DT consists of a variety of companies (T mobile, T online is now product and innovation). An overview of 'our world' ... one thing a cms system gives you is to propagate changes quickly, simultaneous work, versioning, creation of Web contents for non professionals

<DanC_lap> "CI"?

Kai Scheppe, Deutche Telecom: We have a template based mechanism ... you can quickly create a large number of sites. CMS is not an HTML editor... it is a template mechanism for mass production. We integrate content from the outside of the system, which can be challenging. We trade flexibility... maintaining a large system is another challenge. Time to market often takes precedence... editors are not programmers... automation is fine but editorial process should not be interrupted.

Larry Masinter, Adobe: I was invited because i come from Adobe. The Web today is not authored by hand (mainly) but it goes through systems that generate content from content that comes from other places. It is a great point of leverage, though it is difficult to get the word into that community.

Josh O'Connor, Centre for Inclusive Technology: [slides] CMS are tools that allow to manage complexity for a site. What is wrong? they can be complex and badly designed ... often inaccessible back end, so they cannot be used by people using assistive technologies ... quality wrong is sloppy, usability and accessibility is often poor ... lack of interaction with users particularly with people with disabilities ... ordinary people are just not technical ... lack of knowledge all around. Can we expect the user to understand all this stuff? No and we should not. The promotion of ATAG compliance would be a half way to push out accessible systems. Accessibility and usability are part of good design. Involving users in the development would be very good. ATAG= authoring tool accessibility guidelines

Charles McCathieNevile, Opera: I think i agree with josh that making atag part of procurement stuff is important. Kai, could you elaborate on your experience these kind of decision?

Kai Scheppe, Deutche Telecom: the vendor decision was made on the way it would fit the business model and I am almost sure that the accessibility was not part of the decision
... i figured at some point this should be done and legislation it would come
... some people know that but companies as a whole I am not sure it is on their radar screen?

Charles McCathieNevile, Opera: How much time would it take?

Kai Scheppe, Deutche Telecom: 1-2 years...

Sally Cain, RNIB: Let us not forget about the people who create the content and most of the cms systems do not give tools for people with disabilities cannot use those tools

Shawn Henry, W3C: by the way atag also covers that aspect, but yes

Janina Sajka: In the us cms system based is the whitehouse.gov which is an accessible site, managed by administrative people, non technical people.

scribe: Kai, you spoke about your company but what about your customers

Kai Scheppe, Deutche Telecom: CMS is a dumb system that generates code quickly, the people who generate the templates are those where you have to start
... the system cannot generate code just because the system it better, it boils down to the users of the systems
... we incorporate tables of of other company, the company sells a full html page which is then embedded into the page
... i am running a validator through the full portal to find those
... but a lot of the stuff we do is automated, difficult to find

Josh O'Connor, Centre for Inclusive Technology: A system could adopt some constraint based system that could improve the case
... there are tools that are better, drupal and Joomla, for example,
... there are systems out there, technology exists out there

<dsr> janina adds that the white house CMS is available under the GPL

Jan Richards, University of Toronto: I agree with you, josh, there are many places when constraint can be put in templates, accessibility elsewhere, there is a huge strength in the cms way. They have a kind of power

<ori> I think the main issue is does the template designed with standards in mind

Kai Scheppe, Deutche Telecom: Let us not forget that it is a stupid software, the great majority of the systems out there do not have it
... it boils down to people using the templates

Jose Alonso, W3C/CTIC: I agree with Kai.

Josh O'Connor, Centre for Inclusive Technology: It is interesting what Kai said, but there are some systems that just out of the box shine, like drupal, with accessible back end. That is the kind of things are we are talking about. Humans are the weakest link in the process, that is true.

Larry Masinter, Adobe: i think if you look from the cms's point of view, there are a large number of requirements, accessibility is one
... accessibility is one area where content management is looking at carefully
... other areas are also talked about like mobile where more attention would improve
... we have to take a broader view of what the web is, and this is an area where w3c can look at. Expanding view of Web into the entire ecosystem of creating content would be appropriate for W3C. We talked about accessibility, i wonder what kind of cms systems could be added that would add microformat, rdfa, etc? There is a real opportunity there.

Josh O'Connor, Centre for Inclusive Technology: that is great, the core stuff should be done before the higher abstractions happen

Al Gilman: In the cms systems it is getting the things personalized with rules, there is a leverage opportunity these systems. It gives the possibility to adapt the use the rules at the right place and it help in cleaning up the Web.

Jeremy Caroll, TopQuadrant: One of the problems Kai seems to have is that the content that is generated is not accessible or not even good html but browsers show it.

Kai Scheppe, Deutche Telecom: the cms should adapt stricter rules on what they are doing
... they should send error message to the browsers instead of doing it and they should send error messages instead. But we should proactively clean it up things

Jeremy Caroll, TopQuadrant: While cleaning up the crap we should also stop producing

Kai Scheppe, Deutche Telecom: The system managers would have to be convinced to be able to it...

Josh O'Connor, Centre for Inclusive Technology: There has to be a greater awareness.

<ori> CMS does not produce code by itself, it takes the template and the content created by the content editor. Most content editors know nothing about HTML and actually write the code using an editor provided by the CMS

Larry Masinter, Adobe: A lot of times the info you need to provide, say, a caption is loss in the process
... there is an opportunity to reexamine the work flow from the photographer up to the video and the page to preserve the information

Kai Scheppe, Deutche Telecom: One of the thing that we use inside is to use the validators W3C provide
... i have to explain a lot of things to my managers
... but because of the complexity, but the speed is such that we need more automated tools
... i combined a crawler with validators to find problems.

Jose Alonso, W3C/CTIC: What W3C could do? we should have more deployment of ATAG
... but what can we learn from CMS vendors on what they need?
... we may be missing something there
... i would love to learn more there

Josh O'Connor, Centre for Inclusive Technology: I would like to see larger vendors use and support accessibility more... the open source community has produced lots of good tools there.

Lightning Talks

Rotan Hanrahan, Mobileaware: [introducing lightning talks]
... questions should be lightning questions as well - at most 1 sentence

Social Networks Interoperability (Renato Iannella, NICTA)

Renato Iannella, NICTA: Social networks interoperability has been a common theme of discussion this week
... social networks are pretty much walled gardens
... you have to re-invite your friends when you get in a new network
... we're trying to see how we can build a community to reduce these walled gardens
... also, social networks are driving the needs for new technologies
... so we're talking about setting an XG to look at the big picture
... in particular on topics around data and policy portability
... we're discussing setting up an XG charter
... semantic web people, mobile and others in w3c, plus non-w3c contributors
... W3C is organizing a workshop on the topic in January
... another proposed workshop during WWW2009

@@@: is social network really a different application or a base for applications?

Renato Iannella, NICTA: there are different levels of social networks

Larry Masinter, Adobe: What about Open Social?

Renato Iannella, NICTA: They belong to the communities we would like to bring to our effort

Phil Archer, FOSI: Can you reassure me you'll look at privacy questions? e.g. being able to remove your content from several places at once

Renato Iannella, NICTA: I can't guarantee, but clearly a topic of interest to me

Social Networking: is W3C ready? (Harry Halpin, U of Edinburgh)

Harry Halpin, U. of Edinburgh: Lots of communities have been trying to address the problem of social networking and interoperability
... e.g. data portability
... lots of grass roots communities working on this area
... e.g. to build authentication/identity systems
... like OpenId, OAuth,
... outside of W3C
... I don't think W3C can afford to ignore them
... Also, W3C needs to look at how to address the problems of merging/querying social networks as social graphs
... Some of these questions are being addressed in the semantic web stack
... One of the big problems is about privacy
... which requires trust and proof
... But the SW ins't quite up to the task yet (?)
... Also, there are a lot of social networks around the work - many you may not know about
... we need to have them data driven and @@@-centered

Dan Brickley: I've been involved in these discussions with others - we have created a public mailing list to continue and further these discussions
... public-social-web-talk@w3.org

Art Barstow, Nokia: Looks like very interesting work. What is the relationship of what you're trying to do with OpenWeb Foundation?
... also, if you're creating an XG, please go with the RF option.

Harry Halpin, U. of Edinburgh: clearly we would go for the RF option

<danbri> see http://lists.w3.org/Archives/Public/public-social-web-talk/

Harry Halpin, U. of Edinburgh: For Open Web Foundation, I think they started it to play as a harbor for the IPR/patent questions

<danbri> (an incubator incubator....)

Harry Halpin, U. of Edinburgh: I think the W3C can play a more humble role in this space
... in looking at policies questions, not only targeted at software questions

Ageing issues and Accessibility (Andrew Arch, W3C)

Andrew Arch, W3C: 1/3rd of the population in Japan, and in many other countries, is more than 65 years old

<Ann Bassetti> we are very interested in social networking within big corporations as well

Andrew Arch, W3C: we've looking at over 150 studies of older users on-line, and they reported a wide diversity of requirements
... partly due to experience or inexperience
... cognitive issues are among the most reported issue
... most studies identify usability issues rather than technical issues in terms of accessibility
... We have a BoF table to explore this further today

Hop, an Everyware Development Kit (Manuel Serrano, INRIA)

Manuel Serrano, INRIA: what language to use to develop an ubiquitous web application?
... we have designed a HOP system that comes with a language and a runtime system
... the idea is to re-use the Web as a virtual machine
... HOP is fully compatible with the Web
... we have added a broker in the general client/server/db scheme
... typically the broker would be used to communicate between your device and the application
... the broker serves also as a full-fledged web server
... this could be used e.g. to create a diffuse music player, where one device can be used to raise the volume on another device that plays music from yet another device
... thanks to this broker that serves as central point
... HOP is available under GPL, and is based on pure web technologies

Rotan Hanrahan, Mobileaware: The presentation you just watched is a single HTML File - it is worth looking at ... HOP is quite impressive, I encourage you to look at it

XBRL and the Semantic Web (Dave Raggett, W3C/JustSystems)

TopBraid Ensemble semantic web browser (Jeremy Carroll, TopQuadrant)

Jeremy Carroll, TopQuadrant: TopQuadrant is a Semantic Web company

Ken Laskey, MITRE: One of the things to take into account: we closed the uncertainty reasoning XG in March - there might be more work coming in there. I think this is relevant for you in case where the information is incomplete, etc

Jeremy Carroll, TopQuadrant: [scribe missed answer]

Standards API and Debuggers (Charles McCathieNevile, Opera)

Charles McCathieNevile, Opera: We have developed DragonFly as a developer's helper - important for developers
... it is an open source tool, a widget, installed by default
... the architecture of Dragonfly allows to separate the watcher from the browser, e.g. debugging what's happening on a phone browser from a desktop browser
... We think this should be standardized for the greater benefits of web developers

Rotan Hanrahan, Mobileaware: We will have another session at the end the day; slides will linked from the minutes

MMI, VB and SCXML Demonstrations

[Rahul is doing his demo]

[Raj Tumuluri is now demonstrating Multimodal Interaction with SCXML, Voice, HTML, and Ink]

Deborah: no time for question unfortunately, feel free to catch the presenters later on

Future of XML Ecosystem in W3C Client-Side Work

TV Raman, Google: I view the web as a democratic thing, it is about everybody coming together to share. It takes more than 1, 2 or 4 browser vendors to do ensure democracy of the Web. It should be easy to create browser extensions for vocabularies. If these become successful, then the extensions can become part of every browser.

Henry Sivonen: browsers present a DOM tree. SVG, MathML, etc. are dom trees as far as the browser is concerned. Today the common way excludes svg and mathml (on text/html). On the client document.write doesn't work with XML. We should add features while keeping html working, and hence we should extend html to include svg and mathml. Browsers are forgiving of errors, and avoid the draconian error handling expected of xml.

Erik Dahlström, Opera: [ describes the benefits of SVG] exporting a fragment of SVG requires some care with current editors

Chris Lilley, W3C: How many people have used HTML and SVG in the same document? How many have found problems with SVG needing to be well-formed? Does any one have authoring tools that generate HTML? what about XHTML? What about authoring tools for SVG? What about tools that create a mix of HTML and SVG?

Charles Wiecha, IBM: We need different strategies for deploying XML vocabularies to the browser. Plugins are not very effective.Server-side rewriting is another possibility. If we had XBL support in all the clients, that would be helpful. In the meantime, we can use JavaScript libraries to interpret the XML. The Ubiquity XForms project is open source (Apache licensed). We are refactoring XForms into a series of bite sized specs. The ubiquity project makes good use of Ajax ...

(Charlie is talking at too high a bit rate for dsr to keep up)

TV Raman, Google: This is actually interesting. What new things do you need to do as a home based innovator? There's plenty of toolkits that are written in Java and which produce javascript. Dichotomy between markup and scripting. As an innovator you should be able to create new xml vocabularies without needing the browser vendors to do anything specifically for you. The only time that browsers need to change is when you change the browser platform. For many other things you can introduce new authoring abstractions without a need to change the browsers.

Håkon Lie, Opera: The people on the web didn't accept the authority of the standards people. They invented their own tags. It is easy to add new tags e.g. within the canvas element. (or was that the class attribute?)

<Steeeven> That's what Raman was saying

TV Raman, Google: There shouldn't be a huge division between XML and HTML. You shouldn't need to wait for a change to html specs.

[Raman and Håkon agree that you can do this today via scripts that interpret the dom.]

<Steeeven> XBL++

Håkon Lie, Opera: strong support for W3C working in HTML5, it would be a sad day if this had to be done elsewhere.

Norm Walsh, Mark Logic: I have being using XML since we spelled it as SGML

<MikeSmith> as Håkon says, there are many ways that developers can create HTML extensions and processing behavior for them without needing to ask browser vendors to support them natively

Norm Walsh, Mark Logic: We would be doing the community a big service if we could follow Murray's idea of one Web. Namespace declarations could be made more convenient. Merging SVG and MathML into HTML doesn't scale when we want to consider other vocabs.

Henri Sivonen: Over past 10 years only a few vocabs that browsers really need to support.

<MikeSmith> What other vocabularies similar to SVG and MathML do we have?

Chris Lilley, W3C: You can see that in 2 ways, perhaps it is too hard to add new vocabs.

TV Raman, Google: What is the cause and what is the effect?

Charles Wiecha, IBM: We have done such a good job on selling the benefits of XML, and we now need to allow these to be realized in the browser. We will see an N squared effect if we succeed.

David Baron, Mozilla: Depending on what kinds of extensibility you are talking about there is a danger there. Using Javascript is fine as it is already available. I use linux on a 64 bit cpu and don't have access t flash which is a problem for many websites.

Alex Mogilevsky (?): For HTML the web works quite well, but for other vocabs e.g. mathml it isn't so good. The consumers are suffering. We are still arguing for over 10 years without a good solution

Chris Lilley, W3C: If all you want to do there are solutions, but not that nice. Student's find it hard to install plugins. The content producers are no better than the students. People tend to give up and use formats like PDF

Charles Wiecha, IBM: Checkout the ubiquity project as an example of how to implement xml vocabs in today's browsers.

Henri Sivonen: You end up shipping an imperative program along with your markup. The browser scripting APIs are what needs to be standardized.

Charles Wiecha, IBM: We are way ahead of just shipping imperative code, we are relying on a well defined XML vocabulary.

<rahul> See Ubiquity XForms examples (slow loading off SVN :-)

Henri Sivonen: What about web crawlers? How do they see the markup that the script acts on?

TV Raman, Google: The crawlers would see the declarative markup. (the interpreter is supplied via a script element)

[Raman talks about the meme of spotting repeated idioms and providing declarative markup for them]

Henri Sivonen: How does a crawler know that a script is present that can interpret the markup?

TV Raman, Google: Having the crawler interpret the script is very expensive ... web hackers exploit hacks with document .write and this means that crawlers have a tough time.

Henri Sivonen: The script API is more important than the markup when it comes to standardization.

Chris Lilley, W3C: The crawler should need to worry about the script.

Henri Sivonen: For infrequently used vocabs there isn't a standardization issue for the markup

<marcos> the DOM can already recognize arbitrary elements.

<marcos> http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0A...%3CtheTimblTag%20hello%3E

<marcos> for example

Charles Wiecha, IBM: there may be intermediaries in the delivery channel that care about the markup.

[Raman talks about the role of proxies ...]

Noah Mendelsohn, IBM: Another way to think about flash, which has a very high penetration (98% in USA). The idea of being able to use flash to interpret new tags

(dsr has implemented SVG and an SVG editor in Flash using haxe)

Tim Berners-Lee: the connection between the script and the markup isn't clear. It would be nice is there was a binding from the XForms namespace to a script that can interpret it. Can we define a standard for such bindings? This would be more successful than plugins. How can a browser know where to find a script for a previously unseen namespace?

Henri Sivonen: doesn't this rely on a single normative binding? that sounds like a single point of failure.

Tim Berners-Lee: I don't see this as using a central point of control. We can use a distributed approach instead with multiple bindings as convenient.

Daniel Glazman, Disruptive Innovations: Maybe having better ways to plugin js libs and implement standards would be a good idr as as well

<timbl> glazou, can you summarize?

Tim Berners-Lee: You need to be able to experiment; and then some things get standardized.

Jonas Sicking, Mozilla: having better ways to plug-in JS libraries into browsers would be good. These libs have been badly integrated with HTML ... we as browser implementers were cautious, not sure it would be used. The cost is high because of the size of the spec.

Jonas Sicking, Mozilla: Another thing is that a number of W3C technologies are difficult to integrate into browsers. It would be good to ... as well having specs which integrate with HTML

<marcos> summary, XBL2 solves this problem

TV Raman, Google: Yes on large specs are a pain. ... people who say they are implementing HTML5 are in fact doing various diff bits. HTML5 is good example of a case where a smaller set of specs would have ended up serving us better. If we has smaller specs, we would know what they actually implemented.

Henri Sivonen: I didn't include Flash as it isn't XML. If you are on 64 bit linux there is no Flash. Flash is part of the platform in the sense that if you don't have it, there is a lot of content you can't access. On my phone I don't have flash

<Liam> [I have flash on my 64-bit Linux system at home]

Henri Sivonen: I agree with David Baron that extensibility is a problem when you [have examples like Flash where you have a single vendor deploying their own extensions to the core Web platform]. What needs to be standard are the APIs.... What I'm hearing is that extending based on the class attribute is bad, but it would somehow be good if we moved it allowing [arbitrary extensions based on element names rather than class values:

TV Raman, Google: The class attribute does not scale. Microformats suffers from that today....works for small things. "That small camel will break very quickly and you'll have many humps on the road." The problem [with microformats] is the same as with HTML5 in that control over extending the vocabulary is centralized.

<dsr> Example of div for video as something that wouldn't work, video relies on some kind of native implementation.

Henri Sivonen: The execution environment is what needs to be standardized.

Robin Berjon: A disturbing confusion is between the operational behavior between what what browser does and the semantics of markup.

Larry Masinter, Adobe: Declarative markup has other uses besides being interpreted by a browser. The ability to use markup in multiple ways is very valuable.

TV Raman, Google: Agrees strongly with Larry.

Robin Berjon: Crawlers not being able to understand new vocabs, that is a red herring. Crawlers are more interested in the text content than the tags. A second thing - a lot of people would like to be able to support new vocabs using a bit of scripting, where is XBL?

Erik Dahlström, Opera: XBL is work in progress for Opera. Likewise for other browsers ...

<MikeSmith> XBL or XBL2?

<MikeSmith> dino, please implement XBL2

<MikeSmith> WebKit has a Google Summer of Code student working on XBL2

Liam Quin, W3C: A few people have mentioned combining client side XSLT with javascript. This seems to work across most current browsers. This means the scripting link can be added unobtrusively via XSLT

<darobin> MikeSmith: I meant XBL2, XBL is more or less DITW I believe

Charles Wiecha, IBM: I prefer the XBL approach to XSLT.

TV Raman, Google: Client-side XSLT is a success story, despite having been declared unimplementable at one time.

Ian Hickson, Google: I am confused as to why XML isn't solving this problem ...

Chris Lilley, W3C: Please stay at the microphone. ... people want some kind of declarative markup to hang the imperative code off.

Ian Hickson, Google: Why aren't xml and namespaces the solution? This almost works ...

Tim Berners-Lee: Right now, people aren't inventing markup vocabs but are rather burying stuff in script

Chris: if your content hard codes the link to the script, where is the competition on interpreters?

Tim Berners-Lee: With well defined markup semantics ... (missed as mic fades). If the number of users increases, then the browser could provide support via extensions bound via namespaces or even native implementations. we need a smooth path.

Steve Bratt: thanks panel and wraps up before the break. I will stay in the room for anyone who wants to ask about the Web Foundation.

Semantic Web Activity Demonstrations

Ivan Herman, W3C: The demos are online, but please wait to do this...don't bring down the network

Eric Prud'hommeaux, W3C: Pushing queries to databases...hospital database of diabetic patients ...I knew schema, end points, knew how to speak to database. If you want to share it, I supplied a URL to id the end point, and mechanically produce a large number of triples. Now use conventions, use data structures, relationship names, and identifiers that others know. Transfer from one to another, do with a simple SPARQL construct. Pretty much all you need to do. Now have three copies of my database ...good for redundancy but not how to run things [laughs]. So let's instead do query transformation and push the query this way....work on one data structure; works on another data structure. Do this back to the SQL, go back to the SQL query; we know how fast they are. ... instead of copies of database I have queries. Configuration is trivial; that's the point. ...SDTM...a data model for clinical data for pharmas and drug studies ...pipeline said use query from one database (HL7) to another ...tell about that construct; gives me one query. ...pipe to another query and another construct ...goes back to another query ...gives same results. Lots of people doing this. ...I am using SPARQL constructs, Virtuoso does DDL. I find SPARQL constructs.
Thanks Lilly and Lincoln Labs for funding this work.

Scott Marshall, CWI: That was Eric Prud'hommeaux team contact of HCLS
...I have a great use case [Scott reads mission of HCLS]: "to develop, advocate, and support the use of SW tech for biological science, translational medicine and health care."
Get people back to being knowledge workers instead of hackers. ..So scientific questions and sources. In HCLS we found interest in neuro degenerative diseases. We found a question created with Alzheimer's Forum to find genes in signal transduction. You can see appear a number of data sources integrated into a knowledge base. When you look at bio-med data sources, there are over a thousand. Huge amount of literature.
See our question here, we used these four data sources. It is very important is that we used linked data principles. Used URIs to name things, HTTP URIs. ...when someone clicks from browser you get useful RDF ..include links to other RDF then you start to build Semantic Web...as it looks in SPARQL RDF query language...can see four different data sources are being integrated. We have done preprocessing for reasoning ...cannot reason across 300million triples ...it gives you an idea, is the query I just showed you. ...I showed you the faster one.

<Karen> [Web not working]

Scott Marshall, CWI: ...distributed query use case ....pushing queries to data ...the knowledge base; we did with commodity hardware ...a couple months of hard work ...big on wish list was to query sources where they are ...see in diagram; before you have ability ...go out to every data source ...and use diff. access methods to get data. ...story after is push out to diff sources and push out that way, instead of aggregating in one place [multiple distributed sparql queries]. That's it.

Ivan Herman, W3C: While we switch, any questions?

Jeremy Caroll, TopQuadrant: Is there a question about query vs. aggregating data; what about performance? ...when you do joins across the databases?

Eric Prud'hommeaux, W3C: Problem with that approach...requires someone to pave the road ahead of you to do the warehousing for you, but you cannot do because no one has paved the warehouse ...performance; all in C; nothing complicated ..query transformation is lost...

Ashok Malhotra, Oracle: We have one datapoint from OpenLink where aggregated data runs two to three times faster

Raphaël Troncy, CWI: This is a Dutch project (e-culture MultimediaN) using SW technologies showing this cloud before the demo. People who work in SW are used to seeing the data cloud ...you can recognize dbpedia

Raphaël Troncy, CWI: Museum is the old cloud ...use old controlled vocabularies. Getty is source; licensed. Project has converted all these vocabularies to RDF, so you get machine readable; and get cloud. Demo shows using all these controlled vocabularies. This is just an interface for annotating artworks used by professional annotators. They ID an artwork. What you see here, click on it, and look at image. It's a famous drawing from 17th century showing French Robespierre in public. You can edit the title and descriptions, that's free text. But you can fill in who, what, where, when property use control vocabularies behind the application, and have suggestions. See on right is various people, control vocabulary of the source from which it came. I have alignment, this was in two different resources. I want to say join artworks with this person. Select it; see the label; but it's a URI click and behind it, there are descriptions, properties, values, diverse agents and so on. I can see that with overfield, see an execution, type and I can get the values, suggestions from this controlled vocabulary. Again here I continue: artwork has been made in Den Haag (the Hague, in English). I will have suggestions, get the professional annotator pushes data into database, and go for next artwork. Nice end of story. Is they like how to use it ...the system is so close. Openness that they are bringing, so now they are upgrading the whole backend systems to integrate search

??, @@: Did you consider to provide a common ontology for your tool? ...so I align; your tool works out of the box

Raphaël Troncy, CWI: Use what others are using ...take controlled vocabularies ...and align them ...quality of data matters ...align before putting into the system

Diego Berrueta, CTIC: [Zaragoza eTourism demo]. City of Zaragoza was interested in promoting city; expecting millions of visitors, that's why we built this application. Then idea is to enable a customizable plan to visit the city. This is the home page, first step is to get a profile, here is a set of screen shots to show application, this is the plan/map. What to see, do in three days, if you click you get a detailed view... you see a lot of places to visit and things to do ...plan splits it up by morning and afternoon. Get a description, opening hours ...see in Google maps ...change tab and get more things like music, restaurant locations, shopping centers. So far, the plan is the same for everyone. If I change my profile, I have a form where I can specify specifics, such as traveling for a conference, rest; traveling alone, group, with children ...if have disabilities; many options and preferences ...go out at night ...things I like and dislike. What I want to visit. If I regenerate the route, I get a different plan, specially created for me, so how does it work? Data was already available somewhere, databases behind CMS of the city council - University of Zaragoza. ...collected in different information silos ...we wrote an ontology and populated it ...thousands of resources. Once you have integrated the data, you can do lots of interesting things. Two stages: semantic match making and rules based system, then we used a planning tool using these resources. Conclusion is that we used SemWeb, RDF to integrate data that was unconnected, then we used Rules to create a new service on top of this pool of data.

Eric Prud'hommeaux, W3C: Does the tourist board link to your site? ...this is part of the city council Web site; linked from the home page ...of the city council ...there are two links, one to application, another to an about of. They are so proud of it, they want to show people how it works. Only two links to an OWL ontology that describes the sources. Great to see this going public.

Steve Bratt: ...more and more applications for Semantic Web technologies ...something we need to pay attention to more and more. Useful stuff; increasing amount of data out there ...many conversations about specific technology areas. The next panel is should be of interest and applicable to anyone ...getting to Rec [Recommendation]

Getting to Rec

Some of the getting to Rec panel. Photo Credit: Kaz Ashimura

Olivier Théreaux, W3C: We have a number of seasoned/pickled W3C WG chairs, who are ready to share their experiences with you. Covering tools and guide for chairs dealing with communities, reducing impedance, etc. Raise your hand if you know how many tracking systems we have a W3C! ;-) More seriously, Rotan asked me to talk to you about tools. I will come to what tools do you think should be available at W3C. A lot of times someone comes and asks me "It would be great if we had this"
And I respond, well, we have that let me find you a link. A lot of tools and processing channels are already there. It's really difficult for even seasoned W3C chairs to find them. Let me point to some of them and show you how to find more. One way is the chairs list. It is used a lot to announce "I'm transitioning to this, I'm transitioning to that". Someone complained that there's too many announcements. I suggest finding a way to filter that... chairs@ is a great way to exchange tips. It's ok to ask for help when you don't know, there will be someone to answer your questions.

Olivier Théreaux, W3C: I hope everyone knows about the W3C Guide Book. There are hundreds of pages. You won't want to read them all. But you would want to read many sections.... It would be good to have a better indexing, so people can find the section they need. The guide is for people new to their role. We want to recreate the Guide, and this time as a faceted guide.

[Olivier scrolls through About the Guide page on w3.org]

[Olivier shows "Table of Contents; Facets"]

Olivier Théreaux, W3C: There are new facets... Timeline, Roles, Activities, Tools

<karl> olivier++, ianj++ for the faceted view of the guidebook

Olivier Théreaux, W3C: The title is "Collected Wisdom of the W3C Group Chairs and other collaborators". Please help us keep it up to date! Add information, remove information, etc.
Ill-known Tools. We have wikis as collaborative editing tools. They're still fairly tedious, but we are improving. There is a Last Call Comment Tracker. It is very useful to have this tool that lets you collect comments

<ted> s/deployed soon/deployed now/ for mediawiki wikis and we will be migrating the moinmoin ones

Olivier Théreaux, W3C: Discuss in the WG, get consensus, reply, track
... Mobile Test Harness. This is a semi-automated test harness. If you have a test suite this lets you crowd-source your testing needs. WCAG 2.0 evaluation database... If you have any idea of tools you think we don't have yet, and come talk to us. Maybe we have it. We have a lot of contributors and chairs that have helped us code tools. The tools are all open source
Michael Cooper will give a demo of the WCAG 2.0 tool.

Michael Cooper, W3C: The WCAG guidelines has a different problem to solve than other REC specifications. We're testing implementations in actual Web sites. A standard test harness wouldn't help us. We developed a framework for evaluating technical problems.

Rotan Hanrahan, Mobileaware: An awful lot of tools were contributed to W3C, and don't get enough airing. This is just one example.

<IanJ> IJ note to chairs: we welcome demos of tools at chairs meetings.

Bert Bos, W3C: Rotan thought that I have experience with interacting with different communities. That's true.... He also thought that would help people get to REC faster. I'm not so sure of that.
First is editors: editors edit in different ways. Some are programmers, some are not so technical. Accept that you need a number of tools as number of people edit in different ways. One tool we have in CSS is a post-processor. You can give it a complete document and it does nothing... Or a less complete document and it adds things
Then, the specs themselves. You can have one big spec for everything, or you can split specs into smaller modules. I think multiple smaller specs is better... You can get more people interested in given pieces of it, and people can work in parallel.... There are some disadvantages, too. More ambiguities, problems defining their interaction.

<olivier> [note from #tpac: the newly reborn guidebook will be made public, in order to help contributors in public WGs and IGs too. See the draft ]

Bert Bos, W3C: Then you need to get to REC. There are tools, but the best tracker is a human... Someone who keeps track of all the issues, and makes sure they get address. If you have someone in your group called fantasai, then you're in luck. Then you have a teleconference with the director and then you're done. But if your spec is very popular, then you'll always get issues. The public will not let you go to REC unless you are very strong.
Then the WG: the WG will keep returning the spec from CR to LC again. You have to get out of that cycle somehow. And I haven't found a way to do that yet. So, next implementers ... Implementers they are dangerous!

<glazou> BWAHAHAHAHA !!!!!

Bert Bos, W3C: They are mentioned in the charter. They think they are very important.
You'll have to get a lot of other people to contradict them before they'll accept a disagreement ...
Web designers: You can ask them what they want, but they will always ask for more. So don't ask them what they want. Suggest a feature and ask how they will use it, even then it's not so good. Very few people can design a good feature, but maybe 1/100 times you will find someone who has a good design. Make him immediately an invited expert. You might understand , but not everybody else does. If you don't get any comments on your spec, then it is not understandable.So go back, throw away half your spec, simplify the rest, and then ask for more comments
Yourself: Keep repeating to yourself: "More features doesn't make a better spec". Smaller specs are better. Even better is modular, that fits together with other specs. Even with specs other WGs write.
The best way to realize ... is to go out of this community. Talk to people outside the web community and realize that most people don't understand your spec. If you understand that, you can write a good spec.

Daniel Glazman, Disruptive Innovations: Daniel Glazman of Disruptive Innovations, co-chair CSSWG. I would like to point out the cool stuff in HTML 5 come from implementors. Secondly, Web designers are our customers. They know what they need better than us. We are geeks, we make things by geeks for geeks, and I don't trust us.

Doug Schepers, W3C: Transparency. This is something SVG learned well. We keep our tracker public so people can see what we're working on.
Responsiveness. Respond to people about their comments, issues. Accountability, and making sure you don't get blocked at LC. Keep track of everything. If your group is long-lived, then you're going to come back to the same disagreement again.
Make sure you understand why you made that decision, and document it.
New members of the group are going to do "spec archeology" on the past specs and try to figure out why things were done
Finally testing: it always takes longer than you think, even if you think it'll take a very very long time. Maybe not start with tests, but write tests in parallel. You don't want to lose momentum in the group and then need to start writing a test suite.
Getting a spec to REC status is a process that starts with "humm, maybe we should have a tech that does this" and your group's job is to get through the process to REC status.

Dan Connolly. Photo Credit: Kaz Ashimura

Dan Connolly, W3C: Someone once said you can't schedule consensus. It's worth it to try. Set expectations, have some slack, but set some kind of plans. When you get some hairy issue, you can cut that issue or renegotiate.

Steven Pemberton, W3C: First issue your spec. Make sure you have a system for recording comments. I think this is important and these half-dozen systems should be on one page.

Olivier Théreaux, W3C: We have that, let me get you a link

Steven Pemberton, W3C: Our WG works based on Jitterbug, a tool created by Shane McCarron. It's email-driven so that it retains a paper-trail, and it integrates with other W3C systems. Best is that it has a button that says "Give me my Disposition of Comments". That saves so much work. So you have to go collect your comments and sometimes you have to go solicit them. You don't want someone spotting a major problem at PR.
Then there are AC reps who don't understand the concept of Last Call
... We also have DoS (Denial of Specification) attacks
... people use process to produce hurdles for you
... So process is friend as well as enemy
... Triage
... In battlefield hospitals they have limited resources
... They split the wounded: those who will live without care, those who will die with care, and others
... W3C process is a battlefield, you have limited resources
... Before the meeting get someone to split the comments into three classes:

  1. those clearly right and need no discussion (typically editorial)
  2. those that are clearly out of scope
  3. the rest

Steven Pemberton, W3C: Spend your resources on 3
... Keep discussion short
... Discuss issues you think can solve quickly first
... Try to have potential solutions ready: avoid designing on the fly
... Designing on the fly is potential rathole
... Try for 5-10 minute rule
... IF you don't finish, put aside and try again later
... Reminder: you want to reach consensus
... Don't ask "does everyone agree with that"
... ask "any objections?"
... Listen for magic words "I could live with that" and stop as soon as you hear it
... You should be strict with comments.
Those that do not include constructive suggestions do not need to be taken as seriously as those that do.

<IanJ> http://www.w3.org/2005/10/Process-20051014/policies.html#FormalObjection

<IanJ> "An individual who registers a Formal Objection SHOULD cite technical arguments and propose changes that would remove the Formal Objection; these proposals MAY be vague or incomplete. Formal Objections that do not provide substantive arguments or rationale are unlikely to receive serious consideration by the Director."

Steven Pemberton, W3C: ... This is advice to those making comments
... Attain Consensus
... Decide on each issue. Try to get a response if you cant. If not, silence == consent
... Dissenters cannot stop a group's work, but try to avoid it
... Document the group's decision
... For non-accepts, document the link to the mail to the person
... Document objections,
... Formal objections are required by process to propose changes that would remove the objection
... Forms 1.0
... We had 250 issues. One email had 60 issues, extremely good review
... It's clear which one we accepted, which we didn't'

Dan Connolly, W3C: I have a postscript. Someone says "I would like an extra feature" sometimes you can skip, especially if you start with requirements
... If you keep a public resolutions list and document rationale, then if someone makes a comment
... You can point them to that group decision, say we already discussed this and concluded X because of Y

Dan Appelquist, Vodafone: I have a lot less experience than these other people, although we do have a doc that's in REC

Dan Connolly, W3C: It maybe simplistic to make this recommendation, because in our wg it's a best-practice document not a technical spec like xforms or css
... It is a fairly targeted effort
... We are working on a targeted set of reqs
... Piece of advice I want to impart
... I've had good experience keeping it short and being extremely stubborn with commenters and with also people in the group
... in a nice way
... I'm lucky in that we came into the group with set of reqs that were easy to fill. We could see the beginning middle and end here. We are going to devolve into an IG and then we're done
... I would like to examine whether there are other situations in w3c where that could be possible as well
... I see a lot of efforts that go on and on
... There's a certain pleasure to finishing something.

Rotan Hanrahan, Mobileaware: [thanks presenters] Questions?
... How many chairs here in this audience?
... How many expect a CR or REC within next 12 months
... How many expect to be on schedule?

Charles McCathieNevile, Opera: Initial schedule or revised schedule?

Rotan Hanrahan, Mobileaware: Those who realize they're not on schedule
... Would you benefit from some mentoring from others?
... Nobody wants that?

??, @@: I think that would be a good idea

??, @@: Blueberry on top

Eric Prud'hommeaux, W3C: Dan suggested that a lot of record-keeping gets you out of 11th hour trouble

??, @@: Frequently someone would say no, and then ask if you have any new information that would affect the discussion rather than rehashing it

Rigo Wenning, W3C: Staff counsel for w3c
... I'm also one of those hidden tools
... Engineers tend to negotiate a lot around legal issues, and normally they go around in circles
... The earlier you call me in, the sooner we can stop

Dan Appelquist, Vodafone: I'd like to talk about rat-holing. Is it a good idea for the group to decide it's a rathole, or the chair decides?

Frederick Hirsch, Nokia: When this happens in my group I start generating proposed resolutions

Dan Appelquist, Vodafone: And then I keep asking "are there any objections to this proposed resolutions"?...: You cannot let this keep going

Dan Connolly, W3C: ... don't' kill it ... if you're having a finishing discussion, let it finish. There were series of WGs I was in, I got better at scheduling over time. Also grddl was a smaller community, not everyone in the planet was interested in making comments

Bert Bos, W3C: Have a break when you discover a rathole. Usually after the break it's not so important

Robin Berjon: Schedule the ratholes for before lunch.

Bert Bos, W3C: or schedule ratholes right before a break

Matt May, Adobe: When you're sure that the issues are out of scope, ... consensus, closer to unanimity. It's critical to make sure that you consider these things fully and not making a decision by fiat... it'll come back to haunt you

Steven Pemberton, W3C: In some cases there are comments where it's very obviously clear that it's out of scope ... The group has seen it, but it's no use to discuss when you couldn't possibly do that ... Comments are entered and closed that will surely be re-raised during LC/CR exit process.

Dan Connolly, W3C: I use "that's out of scope" less and less.... I usually ask them to find it in our charter instead

Steven Pemberton, W3C: The precious resource is the discussion time

Robin Berjon: I wanted to bounce back on idea of mentoring. In my experience. I think W3C doesn't often enough make use of co-chairs... Following meeting 3-4 days straight and jetlag .. it is hard. Most participants fall asleep at some point

<chaals> [Keep *good* minutes of discussion. *Really* *good* minutes.]

Robin Berjon: Can't do that if chairing. Helps to have a co-chair. Second, in terms of timelines, I think W3C could be a lot scarier to WGs by threatening if group is not following timeline. With XBC, we had a difficult job, and we finished on time within 1 year. We told people "guys, this has to be finished din a year or we won't finish".

Dan Appelquist, Vodafone: It was during a telecon where I was falling asleep that I realized I needed to bring in a cochair. We have complimentary co-chairs, we have different skills that we bring together

Mohamed Zergaoui, Innovimax: I have 2 questions ... I want feedback, are groups with 2 co-chairs more on-time ? ... or any other trends
You mentioned the DoS. The fact is I think that there is only 2 point where the AC reps are really targeted by the spec: when it's moving to CR and when it's in P. Maybe the mailing list of the chairs could be shared with AC Reps interested in following the process.

Steven Pemberton, W3C: First question i have no idea. Certainly it keeps people more sane. I'm sure that it's not true to say that AC reps don't get to hear about LC announcements.

<dom> [the chairs@ list is open to any member rep, it is not restricted to WG Chairs]

Dan Connolly, W3C: You're welcome to subscribe to chairs list. Don't hear about LC as a matter of course

Tim Berners-Lee: There's been some comments, how to triage them, etc. How to point out how we've thought about it. Commenter obviously has another opinion ... save your spec from being interoperable with another technology... ...find the guy who's realized that you were wrong.

<darobin> MikeSmith: I did mean XBC, not XBL

Tim Berners-Lee: It's useful to point him to where you've made the argument ... Getting to REC is not really the object. The process is just another tool. We built it because we found without it we didn't do things properly. We didn't have CR, and then we found that people weren't implementing until things got to PR. So we changed the process. The objective is not getting to REC. The objective is getting an interoperable Web. Discussions about process change, there's nothing that IJ likes better than re-editing the process doc. There's no objection to discussion like these... [Tim jokes about people liking process editing]. But do, regard the process document as your friend. If it needs fixing, fix it.

Rotan Hanrahan, Mobileaware: One thing that works for us is "Respect" for other persons in your group and other points of view... And now our closing demonstration

Michael Cooper, W3C: I apologize for tech problems. My American computer didn't like your French projector.... This is one tool amongst many.
As I mentioned WCAG is a guidelines for what authors should do not a spec for implementors. So we developed a different approach for meeting our CR requirements. Maybe contains some useful thoughts for others.
Since WCAG is guidelines, an implementation is a Web site. We had to test that each success criteria could be implemented by an autho..Also treated guidelines as a black box. We need to make sure authors can understand the guidelines, not just that they can be followed. So we asked people to send us information about their implementation.
Surveyed about site size, type of size, which conformance level it was trying to meet, what technologies were being relied on -- need diversity of technologies.That was in one process.
After that, we asked them to come back and provide information. We asked the implementors to actually tell us what have you done with these success criteria. And they could say whether they match the criteria or not and to comment on their techniques, any problems etc. And then we needed to go back and evaluate them. Maybe author didn't understand the guideline.
We made a team of evaluators. They were presented with a page with info from implementer and then pass fail n/A options. This is guidance for you to provide substantiating data. You evaluate the Pass, then maybe add some comments.
We required at least 2 evaluators for each implementation. After getting evals, we would review results. Here 2 evaluators don't agree, so we had to go interpret results.

Michael Cooper, W3C: Finally we arrived at a set of WG results. Then we could track how we're doing. We need at least 2 for each criteria ... So again, a very different approach for testing for CR. Useful in some circumstances.

<marie> many many thanks again, fantasai

<fantasai> :) my pleasure

Lightning Talks - II

OMTP Bondi (Nick Allott, OMTP)

Nick Allott, OMTP: I'm CTO for OMTP, I'm going to try to introduce you to the Bondi Initiative. How mobile web applications can access devices... this is real. Many companies developing. Some brought to market.... risk of fragmentation. Fragmented security leads to security risks. Our weapon of choice is open source. Bondi is explicitly targeting W3C as the organization into which we can feed these.
... 11 APIs in scope; we accept the need to evolve and competition; all APIs mediated through adaptable security mechanism; in scope can access from @ and widget.
Re: W3C. starting point: of the 11 APIs we have in scope, 3 can map to W3C technology, the other 8 are up for discussion. Common sense is that security needs to be considered for all
... there's a need to do something quickly. open source important. security is essential. Bondi embracing elements to do work in W3C

Rotan Hanrahan, Mobileaware: No questions? ok

Pitch for the upcoming security workshop (Thomas Roessler, W3C)

Thomas Roessler, W3C: On these luggage tags we're saying the Web is moving. That says the web moves from desktop to mobile. It also means that the web is moving out of glass bowl. Case in point: widgets
... what we have are applications on the platform programmed in Java and HTML. They'd be much more boring if they didn't have specific tools like system() to control your machine
... we have people develop web apps and they have privileges on machines, many end up being unsafe. the functionality is coming. Matt will tell you how your browser will find you.
... the camera will take pics, the gps to find you, the phone to make it expensive
... we need to take the security very seriously. we have to protect the user from things going wrong and who will control what developers will do
... please give proposals by date; please come to London for event on 10-11 December, for the workshop. Position papers due 30 Oct. Please contact Dom or Nick, or check the workshop home page.

Rotan Hanrahan, Mobileaware: Questions? no? grab tlr (metaphorically) later

Beyond HTML5 video (media fragment URIs and video annotations) (Silvia Pfeiffer, Invited Expert)

Silvia Pfeiffer: I'm an invited expert for anything video, 15 year background, I run my own video start up. I'm here as a rep of Mozilla for video accessibility on the web. I'll talk about video on the web. We have video element in HTML 5 which is a step forward. (explains formats of example on the screen) ... we can go much further. we can make video more common.

<IanJ> wiki-style media annotations

Silvia Pfeiffer: mediawiki example. This is a wiki for video. What is important is to expose the structure and content to the user and server ...accessibility possible if we have info on the client side. We can do deep search (key word) time offset in video
... possible if we do time aligned or referenced by name to offset

<pdenning> ogg theora was mentioned

Silvia Pfeiffer: working groups looking at different aspect. we still lack exposure. We have some wgs; but we lack is the ability to expose video structure to the user agent. Timed text going in that direction, but not really yet

??, @@: Have you talked to WebApps?

Silvia Pfeiffer: Yes, on the what wg mailing list. I'm on WhatWG mailing list and we're discussing there
... we'll do this later, phase 2 stage
... difficult, so many formats and need agreement, will take a long time for a common API

Charles McCathieNevile, Opera: chair of WebApps. we have talked and we have an ongoing dialogue

Paul ??, @@: Have you considered the need for linking into the video?

Silvia Pfeiffer:This is what the media fragments working group is doing

Jim Allan: User Agent WG, would like to talk to you. Accessibility of video is more than just captions. I'm concerned if whether the player is going to be native to the browser or designers making their own

Silvia Pfeiffer: We need one standard way

A plea for members to provide Web inspiration (William Loughborough, Invited Expert)

Bill Loughborough, resident old geezer. Photo Credit: Kaz Ashimura

William Loughborough, resident old geezer (self described!): I'm William Loughborough. First I'd like to give personal history and shameless name dropping. entered MIT at 16 in 1942, working in the cafeteria encountered Norbert Wiener. He's the father of cybernetics. I cleared away his dishes as he was talking to a colleague and he said "did i eat yet?" now I find myself asking the same questions. I dropped out of MIT, entered the navy as a technician
... Dough Englebart was there, when he got out of the navy he got his doctorate, invented the mouse, etc
... I became an itinerate jazz musician and gambler. last 40 years I've been working on medical research and Web Accessibility
... Ward Cunningham the inventor of the wiki admitted he had thought of patenting it and we're lucky he didn't
... Tim said
... all of you are the smartest kids in your class, all of you have the capacity to bring the web to it's full potential. As Raman mentioned, the web is humanity interacting through technology
... As Dylan said, you don't need a weatherman to know the way the wind blows
... self advocacy groups w/ people first in their names, would benefit from an hour of your time
... others are called centers for independent living, for when they get out of institutions. also senior centers where you can change lives by giving an hour
... enabling the web is not a matter of 10 billion consumers of information but getting 10 billion contributors to the web
... not merely passive consumers. I know a lot of you do this kind of thing but we need to do more
... I'm working to make the web not a read only system but something where everyone participates.
... at ? they started telephones for pioneers. it now has more than AT&T and it numbers about six to seven hundred thousand. they do things like repairing braille machines. the companies are now a consortium, telecom pioneers.org. I urge you to get the places where you work to participate in these kind of things. I leave you with a mantra I use "everyone everywhere everything connected" I would also encourage you to grow old because a few years ago Tim was talking about RDF. then in a few months I'd read enough to write the Semantic Web primer. it indicates that you can still do mental aerobics even at my age.

XSPARQL (Alexandre Passant, DERI Galway)

Alexandre Passant, DERI: XSPARQL, project to bridge XQUERY and SPARQL .. translating data... XQuery function in SPARQL, can use concat ... test value in future expression. can't use it to get output value

<IanJ> Yay! xpath within rdf!

Alexandre Passant, DERI: export vCard to FOAF ... translating RDF foaf to xml ... you can check xsparql.deri.org. I'll be interested in talking w/ people.

<IanJ> http://axel.deri.ie/~axepol/xsparql/spec/

Geo-location (Matt Womer, W3C)

Matt Womer, W3C: I'm team contact for a bunch of different things, POWDER and others ... I'm trying to get people to join GeoLocation group. We defined mission. New WG, you can check it out www.w3.org/2008/geolocation . Currently we are moving along, we have 2 co-chairs
... we started work over the summer 08, API spec well on the way
... we're about to get started on primer and test suite. first WG meeting at Vodaphone in December
... we're going for first public working draft
... video based on firefox geod program. created by Mozilla labs
... (describes video)
... yelp data, google maps, shows nearby coffee and tea shops

<karl> I wonder if there will be guidelines for security and privacy done

Matt Womer, W3C: this code is easy, it's Javascript API. There's a get-current position, there's a watch position for tracking. the rest is up to you ... questions?

Rotan Hanrahan, Mobileaware: Geo Location is one group that will know where it is and where it's going

Karl Dubost. Photo Credit: Kaz Ashimura

Karl Dubost, W3C: Karl Dubost, w3C staff. Will there be guidelines about security and privacy?

Matt Womer, W3C: Yes, we'll have info in the primer section as well as the API doc

Thomas Roessler, W3C: That topic is very close to all the device API discussion and is in scope for the Workshop in December

Matt Womer, W3C: This is why we brought in two chairs. One from providers, one from browsers.

Larry Masinter, Adobe:There has been extensive work in the IETF about geolocation (and work is ongoing). I hope there's more coordination than is apparent (on privacy and security in particular...please be coordinate)

Matt Womer, W3C: We're just starting up but there will be coordination.

<IanJ> http://www.w3.org/2008/geolocation/charter/#coordination

<IanJ> "The IETF GeoPriv Working Group is working on the format, security/privacy implications, and protocols for exchanging geographic location information."

Dan Appelquist, Vodaphone: The reason we put forward someone to chair the group, this is one of the most important classes of API ... especially if you look at the things coming out on Google, App Store. There's clearly a need in the market for this and it's just as clear as we need to do it in a secure way.

Thomas Roessler, W3C: There is coordination between groups.

Why you should have a Web site (Steven Pemberton, W3C)

Steven Pemberton, W3C: Why you should have a website. we heard about Metcalfe's law: the value of the network square to the number of nodes ... if you do the math, if you cut network in half, halves the value ... important we have one Web. Flickr, etc. there's a danger. by putting a lot of work into a site you commit yourself to it. There's no way of moving your data from one Web 2.0 site to another
... how do you know which to use, what if a better one comes along. or re: social network, you're all probably like me getting bombarded w/ requests to be a friend or work associate

<raman> is there truly only one WWW? This morning I heard someone dismiss things like Intranets and other things as not part of the Web

Steven Pemberton, W3C:... or geneology sites. What if the site dies? if it shuts down you lose your date ... Google account closed, lost Orkut, lost 4 years of mail ... this partitions the web into sub webs and reduces the value of the web. This is why you should have your own site. aggregators could come and find info.

<raman> All google services let you download your data -- we call it "take your data with you when you want"

Steven Pemberton, W3C: What do we need? machine readable data. you need CSS for meaning. you need to add machine readable data
... when an aggregator comes to a web site it can see what the data represents
... you could ask to find a place on a map, add data by combing information
... rather than putting all your data on someone else's site, put it on yours w/ explicit semantics

Rotan Hanrahan, Mobileaware: Questions?

??, @@: What if your own site fails?

Rotan Hanrahan, Mobileaware: Google cache

Steven Pemberton, W3C: I'm assuming you have a back up of your data. If your website fails, all you need is domain name

<Ralph> Why you should have a website: it's the law! [Steven Pemberton]

<Kangchan> think that data interoperability is really important to the end users

??, @@: I mean focusing on your own website. So your data is always accessible. So of course I completely agree

Stefano Crosta: You've said several times moving your data to your own site, how does one do that

Steven Pemberton, W3C: I said that's one of the problems

Rotan Hanrahan, Mobileaware: One more question from Doug. All slides and presentations will be up as part of the minutes, presenters have included links, etc.

Doug Schepers, W3C: So the sites are not merely storage sites but authoring tools. certainly there's value in an authoring tool that lets you move info

<raman> The key bit here is data portability: http://www.google.com/search?q=data+portability&num=25

Steve Bratt: Let's give a hand to the panel

Mike Smith, W3C: one of the current co-chairs for HTML group. Harry Halpin suggested we have an open house to air concerns on work and for us to explain directly what we really are doing. Time tomorrow Thursday 2-2:30pm, Riviera. it seems that there are some groups that might like to make use of the time

<marie> many thanks, amy

Steve Bratt: thanks and thanks for the effort to address concerns of the community

<DanC_lap> (I hope upcoming TAG nominations are one of the announcements)

<IanJ> survey: http://www.w3.org/2002/09/wbs/35125/tpac2008-feedback/

Steve Bratt: please fill out the WBS feedback survey. Your feedback is really valuable. we use your feedback to plan next one
... next TPAC will be in California. we want to get a head-count, it would be useful for us to know how many to expect. Slides are great, I steal from them all the time. Please give us links to your slides. we'd like to use them as a resource

<Rotan> All the Lightning Talks are on my laptop, and I will be transferring the files to the W3C system very soon.

Steve Bratt: I really enjoyed the meeting today. I thought the moderators and panelists did a great job. My sense is that though we have feelings on where the web would go, we all can come together to agree that we want one web, not a semantic web, not an XML web - no walled gardens. I hope we're all a bit more motivated, there's a way we can find a roadmap, peaceful friendly co-existence
... thanks to William and Raman who reminded us of our person responsibilities to humanity, to do good things in terms of commerce, etc. whether we use the web to serve our local community or the world, I appreciate your efforts

Dan Connolly, W3C: The TAG nominations open 3 November

Ian Jacobs, W3C: this may be Steve's last opportunity to chair TPAC so many thanks to him (applause)

Steve Bratt: for those of you who don't know, I'm moving on to the W3F. Thanks to Nokia for the platinum sponsorship. the moderators and panelists did a wonderful job.

<mauro> [ applause to Nokia for their generous support as Platinum Sponsor]

<karl> clap clap clap

Steve Bratt: ChrisL chaired the program committee and I'd like to thank Chris and all the members of the committee.

Steve Bratt: scribing the meeting is difficult, lots of info. let's give a round of applause to scribes, systems, team in general

<marie> [applause for the scribes!]

Steve Bratt: Admin: Amy, Alex, Kana and esp to Coralie and Alex who put this together
... thank you all for coming here today and participating. I'll look forward to seeing you all at 7pm in the Iles room for a reception

<jallan> thanks to the spell checkers, name correctors, and text fixers

+1 re: spellcheckers, correctors etc. Scribes: Steven Pemberton, Karl Dubost, Ian Jacobs, Jeanne Spellman, Ivan Herman, Dominique Hazaël-Massieux, Mike Smith, Tim Berners-Lee, Philippe Le Hégaret, Karen Myers, Dave Raggett, Amy van der Hiel, Fantasai. Many thanks for Marie-Claire Forgue for helping to clean up the minutes and organize scribes!

Corrections? Contact Ian Jacobs, Head of W3C Communications
$Date: 2008/11/11 16:39:30 $