Agenda, IRC log
Noah: as Jeff said, the theme for
this session is integration
... one other thing. Many of you know I am chair of the TAG. It is the senior technical body responsible for the web. The reason I mention this is that 5 of our 9 members are chosen by election. This year 3 are up for election. Having you choose the best people is important to our success.
... it's always the case when I hear someone says "you should work on security," I think "do we have security experts on the TAG."
... skills which are needed are both technical ones and integration. we also need to solve real problems for real people. we need people who can work well with others (and who can also write)
... it's important we have a good slate of strong candidates. Current or former TAG are here, I encourage you to speak to them, and to look at the TAG pages. Raman asked me to say that he is one of the incumbents who is not running. You'd be welcome to speak to him
... our first speaker is Jeni Tennison. The Semantic Web has had increasing success in government, etc. The SW has not seemed to reach the same velocity as early web. We've invited Jeni as she's one of people that the smartest, most tasteful people look up to
Jeni: Thank you for that kind
introduction. One of my roles is that I am the linked data advisor
on Linked Data gov.uk
... I'm not an SW evangelist. Years ago I thought SW was a pipe dream. Now I view my role as making sure not just to used linked data "because Tim Berners-Lee says so". I want to make sure that RDF works.
... I don't particularly care whether RDF or SW reach escape velocity. I care whether data is useful.
... working with government sites is like working with the WWW but with more committees
... I'll explain what i see as the hurdles, and where the community and W3C should focus
... I've taken this as an excuse to have a bit of a rant and say things i wouldn't normally say to people's faces
... the world is very different from 10 years ago when I said SW wouldn't work. there is a lot of effort to use data in many ways
... many of these uses, visualization, they're like pre-web. no way to go deeper into the data
... I think we're heading toward a world where we'll want deeper interaction with data. "data you can click on" (a Stuart Williams phrase) will be useful
... However, just because you can get a hold of data doesn't mean you can use it
... (shows excel on crime data)
... What do these rows relate to? what is the definition of an acquaintance? what we often find with gov data is that the data is available but how you interpret it is hidden so there's no way of finding what it means
... also, to understand changes we need either links or implicit information. to get deeper interaction we need more sources
... RDF is a good approach to tackle these challenges. identifying things with URIs helps in several ways. you can create mashups which link to things. you can also get extra info at the link, extra context
... you get a value model that lets you combine extra value w/out terms, end user agreements, etc
... The other thing is that info with URIs can be referenced not copied. For e.g.: BBC uses wikipedia
... There are huge opportunities for the reference vs. copy info. Prevents out of date info too. Referencing using URIs is increasingly familiar
... Why is it that RDF is not being used? Let's see what developers say
... I was delighted that the Guardian data blog "it's not great fun learning to develop with the semantic web today"
... i'm going to try to put myself in dev shoes and see why not
... (shows RDF stack) it even looks like it's going to fall over on you
... they get to see the ugliest syntax first. They think they have to use OWL
... none for this is really true but it's what people assume when they think they have to start here
... If you're consuming this data, is it fun and attractive? if you compare to JSON, there's an easy map. there are parsers on every platform. even on XML there are standard APIs and simple path languages like CSS
... The RDF situation is different. there are multiple syntaxes (lists names)
<shepazu> great slides. and great point.
Jeni: It's really hard to know as a
publisher what to target and as a developer to know what to expect.
There's no standard API to use in code, there's no path language.
SPARQL is as close as you get but this is just another language for
people to learn. As a consumer of data it's not fun to use
... Another question. Is it painless, is it likely to go wrong?
... Generally if you're generating RDF from another structure, you're using technologies like trying to use string concatenation
... there are a vast array of formats, difficult to use
... Microdata, new proposal is quite verbose. look at RDFa, using a single vocabulary, having to fudge between languages
... I've covered the issues, now I'll look at what we should do
... if Developers are going to be persuaded to work with RDF, RDF needs to meet them
... we need easy to follow recipes
... we need to focus on the benefits that can be brought to end users
... we need to be able to reach behind a visualization for the source of data
... we need to make sure that using RDF increases the value of data, that it gives the ability to control context. a good argument for gov
... the more people who use it, the more powerful it becomes
... we now have some core hubs - dbpedia, musicBrainz, BBC
... and we have vocabularies which reduces the need for invention which ups the rate of use
... also we have the "everyone else is using it, I should too"
... Linked Data focuses on the useful core technologies. they've created guides and best practices
... we need the publication of data easier for consumers. then people can make educated choices about how to publish data
... On the consumption side, there's a generic RDF API w/in API. There's a potential for a path language
... What we need is something accessible and fun and easy to use, not XML's DOM
... it needs to be used in other contexts
... On the mapping of relational structures in RDF - my concern if they'll be used
... I've tried to focus on meeting the challenges on this gradually blossoming Web of Data.
... I think it's powerful but it's being avoided due to stigma and misperception. Focus on logic etc. which for publishing data is irrelevant
... What role can W3C play? I've thought long and hard. It's hard to answer. The biggest gaps are in implementations. W3C blessing can help with what implementations to use
... Supporting the profile of RDF - what is currently considered relevant and primers on this
... W3C could also provide a home for community standards. it's hard to trust a vocabulary run by a single person, especially from a gov perspective. W3C could also help by sidelining on 'Semantic web' term and focus on "Web of Data"
... like every community, Linked data is insular and in some ways deaf to outside concerns. We only grow if we listen and change
... W3C has a wide range of members, it's you, not the SW community, to listen to. Look again at SW, help the web of data blossom into something beautiful
<fjh> effectively repositioning from technology focus to application focus?
Doug: Doug Schepers, W3C. I've tried SW with SVG. I keep bumping into not finding the right ontology to solve the problem I'm trying to solve. I've seen some efforts to consolidate ontologies. I want a reference - publishing, animals, etc., - to tell me where related ontologies exist. It's impossible to dig into all for them. Do you think there's value in W3C making an effort into unifying ontologies
Jeni: I do think that there's legs to
that. There's something in the Gov data IG to help recommend to
govs which ontologies to use. And how to assess whether one is good
to use. It's hard to understand which to use, it's about community
and acceptance. What tools are available, is it maintained.
... I think this means you need to have curated vs. searchable lists
... I think this is role W3C can play
Jeff: Great talk. Great graphics.
Thank you. You talked about insularity of W3C and SW technologies.
This is a fair comment. A lot of problems are implementation.
... Can you recommend one specific tangible thing we could do tomorrow to break insularity? Getting the topic on the table is one step. But how do we change this?
Jeni: Good question. I did on a
previous set of slides that developers won't get involved with
something they don't care about. The only time people would get
involved is when you start reaching out and affect other things.
Eg: people looked at RDFa when it was encroaching on HTML
... a subversive thing would be to go into other areas, to listen to any objections
... I've found the best way to work out what is wrong with a technology is to try to teach it. creating primers helps you understand where people might struggle
Noah: You've made clear that ease of
use is a challenge and implementations. two threads in how to make
complex technologies easier to use is to simplify. the other is to
keep underlying technology complex but to have tools
... does W3C need to make simpler technologies or do people need to make tools?
Jeni: I think if you need tools it's too complex
Roger: Roger Cutler, Chevron. You said that the W3C doesn't do things like implementation. Suppose that w3C might do something because there's no place else it's going to get done. If you took that as a starting point, what then could the W3C do along these lines to aid?
Jeni: I think that creating a target language a really good parser and API that was well packaged, easy to install. That would be my suggestion. Something actually finished and completed with professional quality
Noah: our next talk is by Tim to discuss an effort by him and the TAG on aligning XML and HTML
TimBL: This is a topic that's been
discussed on lists and meetings like here at length, difficult and
... it tends to be a rat hole because there are lots of related intertwined issues both technically and historically
... what happened recently is that Raman sent a message saying we should fix it. "form a TF to look at the problem"
... at the end of this presentation there are links including one to this message
... there's a serious need for this, a whole community using XML tools to produce their websites. Sometimes they produce XHTML, sometimes XML, often serve as HTML mime type.
... they've benefited from the stuff being in XML, they can scrape site in XML. There's a community used to using these tools. Meanwhile the definition of HTML has divorced itself from XML
... you can serialize this stuff in XML but most webpages can't be straightforwardly parsed
... one way to define a goal for this work Raman asked for is: 1. to see how much power we can give to the community that uses XML tools 2. to look at the future, to see whether divergence can be decreased, have convergence later on
... what I'm not going to do is go into details of the technical arguments. The TAG has gone into a number of them. There have been lots of suggestions to tweak HTML or XML. This is just an announcement that we'll make this task force - we'll have people on TAG and people not on TAG
... I'm not yet announcing who is on the TF. but for those who are new, there are various areas where HMTL and XML have diverged
... HTML will cope with improperly nested tags (e.g., <b> <i> </b></i>)
<glazou> +1 !!!!
TimBL: there are tricky things where you can send page to XML and XHTML parser, but the famous one is <tbody> inserted by XML parser. but if you write a script that goes through the DOM it will miss the first <tr>
<inserted> [slide 3: Divergence Dimensions]
TimBL: tricky little things like
that. The center of this is in difference in irC
... it's a complicated issue with syntax w/out namespaces but a namespace gets generated
... One trick is for mime types which use XHTML, could have parts set up by default. you know you'll get namespaces every time by default
... doesn't have to be sent across net every time
... I put in completely separate bullet is that XML requires attributes. don't ask me why. it's down on the list as it's a silly problem. easy to see how to map
... The last one is a big issue - distributed extensibility
<Ralph> [slide 4: Polyglot]
TimBL: what does this mean anyway?
The need is that if one WG makes a spec, that a sub-community using
the spec, e.g.: HTML. they can use HTML and put in tags
... Let me mention that the polyglot document is important
... this document describes how to generate an XML document and put on web as though both XML and HTML there are traps to avoid. this polyglot spec tells you what they are
... It constrains the syntax (eg: lower case here). Don't put in noscript. If you're trying to do this sometimes there are choices eg: <tbody>
<dom> [slide5: Polyglot]
TimBL: Polyglot document very useful.
I find the stuff I write is polyglot. I forget and tidy them as
HTML but keep them as XML
... Polyglot is important.
TimBL: Distributed extensibility
isn't just a quirk. It's a difference in values and what people
felt were important things to optimize for
... On one hand, you say to have interoperability use one syntax. on the other hand, use different syntax, new functionality creeps into system
... it's how Internet protocols and HTML has arisen in the past. I'd like to look at different examples
... one example of extensibility is adding SVG to HTML. SVG has always been something which should be in HTML
... now SVG has been adopted by HTML5. the HTML5 spec now tells you how to put in SVG (instead of pointing at SVG spec). now you can put an SVG tag in by itself
... SVG is being incorporated and a discussion of how to make this happen is happening in the HTML Working Group
<Steven> XHTML+SVG+MathML (2002)
TimBL: Other issue is RDFa, a way to
put explicit semantics in a web page. There's a different WG.
Should be an extension to HTML
... Design Authority for this extension is a different group. There's a framework for coordination
... Now you have Facebook ML, we found out about it. they use namespaces, they didn't ask us, they didn't check with HTML
... lots of such examples. If an editor saves information about it's current state or version info, there are many things people put in HTML. groups don't want to come and check with W3C
... Distributed extensibility means there are a lot of people out there who need to use namespaces.
... if you have a domain name, you can have a namespace
... Important to note there is a scale - distributed to extensible
... video tags in HTML is obviously happening.
... The philosophical question is how to value the needs of the people at the bottom of the list vs people at the top of the list
<dom> [slide 7 Scope/Timescale]
TimBL: This is not a short term
thing. won't be expected to produce comments for last call (maybe
will, who knows)
... Look for a future path to minimize damage
... I respect Raman for raising the temperature here. He said: this is important. Everything which has two stacks which is divergent, there's pain for engineering.
... We should look at this, we have large number of people in groups working on low level. It's good to step back and ask whether this is right
<Ralph> [slide 8: Links]
Daniel: Daniel Glazman, Disruptive
Innovations. HTML had issues in the start. i think that what you
call divergence is XML 2.0 where namespaces are less
... I think this is the future. It's possibly the right time to start thinking about the future of XML. XML as the way it is now is maybe not the best thing for the web on a daily basis
Henri: Henri Sivonen. I've
implemented an HTML 5 parser that exposes the API and coerces the
doc tree into an @@. you can connect them. There's also an HTML
... I think this is a solved problem with parsers and tools and serializers
Noah: Next speaker Janina Sajka and expert on accessibility. She'll give us an overview in integrating accessibility features
Janina: thank you for the invitation
to be here. I'm not going to try to talk about everything going on
... I'm going to try to talk about things I want the wider W3C community to be aware of. HTML is giving us opportunities we haven't previously had with specifications which do their best to support accessibility
... We've heard the word accessibility. When I say it I'm talking about people who don't see or hear. There's a newer group of those who don't see or hear as well as most in this room. This group is coming and the implications are exciting but let's get there
<dom> [slide 1: Introduction]
Janina: I'm Janina Sajka. I chair
PFWG and co-facilitate HTML Accessibility TF
... also at the Linux foundation and editing a document for ISO
... a lot of background in advocacy and in systems deployment
... that the professional side but i come first as a consumer
... Accessibility really does change lives, it doesn't make them see gives them info they might have seen. Doesn't make people see, but gives them words. We want to make sure you don't need a mouse. We're looking to the same things as everyone else on the web. regardless of sensory or motor ability
<dom> [slide 2: WAI-ARIA - Our 1.0 Story]
Janina: In PF we have ARIA. We're
closing in on 1.0. It gives us a language independent way to
... it gives a series of standard kinds of tool, controls, interfaces. It's going to be very important. it's our consistent way of dealing with similar things regardless of host language
<dom> [slide 4 : Intentional Event - Our Newest Story]
Janina: Something brand new to us,
not to W3C, we've got the opportunity to do "intentional
... (undo, zoom, etc). how to map keyboard for undo or zoom, or intentional events common in application interfaces but very different in languages
... idea is to get to work on developing technology. We're being driven here because it's not just mouse, etc., but voice interfaces, etc. It's a whole range of different ways we don't need authors to stumble across
... It looks like this will happen in Web Events. this may change but it's our newest story. so stay tuned. we think the implications are far wider than just accessibility
<dom> [slide 5: The HTML Accessibility Task Force Story]
Janina: Moving on the HTML 5 TF. We
formed this together with PF + HTML. to work on accessibility
recommendations for HTML spec.
... many of our recommendations have been picked up. A few to highlight: Aria. you can not incorporate so have to map ARIA to HTML
... we hope at the end of the day we'll have essentially the same technology for Aria 1.0. We're not quite there yet. this is key expectation especially for applications new in HTML5. should make browsing far more compelling
... Canvas - bitmap, no concept of structure, no caret. We've got some solutions there
<Ralph> Janina: without a caret to focus on, how can one programmatically know where to focus a screen magnifier for partially sighted people?
Janina: We want to make sure that
canvas technologies are taken to an extent that HTML5 might rather
they weren't. Eg: write [something] text in canvas
... we are pleased to say this is much further along. Lastly, media. This is one of the most transformative changes in HTML5
... You can see this happening. Go shopping for a TV, you can see some come with an RJ45 connector. Is this just google or another interface?
... a convergence of technologies. We think this will be transformative not just for web but for education
<dom> [slide 6: Meeting User Needs on Accessible Media]
Janina: It's an enormous challenge to get things right on behalf of accessibility. Fortunately, there's a history here even before the web.
<Ralph> [slide 6: Meeting User Needs on Accessible Media]
Janina: Analog broadcasting, we call
that video description. During the time when no one is talking
another, very characteristic voice, describes what is happening so
you can grok what is happening or captions on screens.
... Captions really took off ~25 years ago as result of a law which said that TVs must be sold with capacity for decoders
... Eg: in a noisy bar, you can turn on caption thing and follow along. This whole area of media accessibility was being called 'captioning'. This is just part of the story. it's improved media capacity
<shawn> use inclusive language (not just captions) so that we realize the breadth of issues to be addressed
Janina: We also have things such as
clear audio, now with digital broadcasting you can create a sub-mix
to take out most of other sound effects to focus mostly on key
... It's a production level thing and being used in UK.
... slides to show how people with disabilities are being accommodated in Media
<dom> [slide 7: Tying it all Together with Media]
Janina: successful industry:
... not so important for short YouTube clips, very important for longer lectures, for eg
... What do these technologies mean in the spec It's not quite rocket science, it's more like building a bike. Synchronizing media streams
... Streamable text, Secondary Audio, Secondary Video
... We think this is transformative, we think we can deliver in HTML5. Laws have begun to show up, esp. in the US, to make sure that devices support access. We'd like to be the solution for the environment instead of having them invent this for us
<dom> [slide 8: Thank you]
Janina: If any of this interests you the Accessibly TF is an open environment. We want the community to be involved. You are welcome and thank you.
Noah: we'll take questions
Sarah: Sarah Hart. I know there are many different sign language dialects. how do you prevent getting the wrong one
Janina: we hope that there is an option to choose which one to use. There are many which have been defined
<Steven> Janina mentioned that W3C had earlier tried to define abstract events. Here is an earlier attempt from 2003.
Thomas: Alexey Melnikov is having travel problems
Noah: Thomas will be speaking for Alexey on the theme of integration and how we work with other standards organizations. IETF is one of the most important connections we have. We invited him to speak from the IETF perspective on what it is like to partner with us and how to make this more effective
Thomas: These slides were prepared by
Alexey Melnikov, peter Saint-Andre and Mark Nottingham
... I'll give a brief intro to the IETF, about how W3C and IETF work together and then Conclusions
... You probably all know that the Internet Society is not just helping us but is giving the general framework to IAB, IESG,
... IAB - appeals function, architectural guidance.
... IETF, we mostly work with Applications. There are directors who manage WG. Collectively they act as IESG
... they also fulfill function of reviewing and approving documents. Area directors are reaching out to community for help
... there are areas of overlap. they're looking for volunteers for review team
... Standardization is like and non-alike W3C. You can see cultural difference in slides. You always pick default of slides. There are no formal membership
... You can go to meeting (paid), participate on lists, etc. Decisions are made on the mailing lists never in the meetings (often in the bar). Of course the physical idea of the meetings are critical however decisions will be ratified in email. similar to how some W3C WG work (thinking Web Apps).
... They don't have a royalty free patent policy. They rely on disclosure. There are stringent rules and happen early
... as you look at implementation experience, if there are IPR issues, it simply won't get traction. This is different than our work but has a bearing on joint work
... they have an active tools team which are held in high regard
... There are hard topics where several people meet to hash out issues. the name for this is "design team"
... RFCs - it's not a standard. different types. it may be published by an individual, may be published by a group. not all IETF products. even of IETF products may be historical documents 30 years ago, wide variety
... Some notes on why it functions. They focus on interoperability format. What bits go over the wire. what bits need to go over wire to work
... Socket API is a notable exception
... Traditionally IETF does not do local APIs
... Attracts a broad range, they're fairly proud of review they get when the process works
... Another issue, vast scale of work. In terms of sheer volume there's more going on at IETF. Sometimes competing proposals. They try not to have competing protocols so it's not always possible
... (gives examples). Alexey notes that the IETF usually regrets this
... Decision making: there is no voting at IETF. A famous ritual when an issue comes up is the "hum"
... a way to get consensus
... Why does the IETF not work so well? See the previous 2 slides.
... W3C is a consumer of IETF standards. MIME, Language tags, IDNA, etc
... (skips language text)
... URI important. Now seeing a formal specification for about:
... IDNA, allows new characters. Permits a few exotic/vital alphabets. That is recent work now seeing deployment
... Other recent publications. including: well-known URIs. there is now a standard for putting metadata in URI
... Activities: internationalization, URN, OAuth
... also joint work, broad definition. XML Signature. Started as a joint group. Spec currently revising in W3C group that we keep IETF advised on. One issue why we don't do more joint work is our patent policy. On IETF you have disclosure, here we have Patent Policy.
... I'll ask Yves Lafon and Julian Reschke to stand. they are doing work on HTTPBis spec
... Joint projects: Cookies are being specified, WebSec, HYBI known as Web Sockets.
<Ralph> Alexey/Thomas: note that there is finally a formal specification of 'about:'.
Thomas: these are areas where for
core architectural issues we work together
... Ongoing projects: Workshop on Internet Privacy coming up in December
... differences between email and HTTP in MIME type
... contact API, several groups working on this
... Conclusions: we have a well functioning relationship. it is valued on both sides. some on management/strategic level. some on technical issues. both very important
... any questions for the IETF side, I'm happy to introduce Alexey. If questions on W3C please see me
Noah: thanks to all speakers
... Tim said he's got some specific questions
TimBL: I got two comments which I
appreciate. One was about architecture in which separate XML and
HTML parsers and one on @@
... when I spoke in 2007 I spoke about this about greater length. Over three and half years, the situation has not gotten simpler.
... There I discussed moving from both directions, looking for common ground, looking at XML 2.0. There's a significant community who do work on XML but you don't see their work. You don't see this on the web because it's enterprise. We have to make sure that even if we don't see it on the web that we listen to them.
... E.g., parser that didn't work. You can't just move community to XML 2.0. Personally, I'd do it like a shot. I can see it's a mistake to @something about quotes@
... In retrospect we should have done XML way back when DanC asked me is HTML and SGML application or not? And I said yes. He held my feet to the fire but I should have said no.
... There are a huge amount of systems and users we can't forget
... It's hugely expensive if it's not compatible.
... On the other hand there are people in the HTML community who assume XML should come to them because "hello, we're HTML!".
... same way, XML community "hello, we're XML!"
... It's hard to see the point of view of the other side. It's very rarely that people who have gone into all the intricacies of the spec (many in this room have). That is why, I'm afraid, I'm sorry, I took the comments but didn't make a response. I hope this counts as one
Noah: announcements: we're about to
take a break. I'd strongly encourage you to talk to friends and
think of good TAG nominees or speak to current/past TAG
... this room will likely have W3C staff here, we don't know whose machines are whose but there will be people in the room if that helps you make your decision
... the break is right outside
... be back by 10:45am please
BOF (Birds of a Feather) sign up sheets are in the reception area. please feel free to sign up topics and for lunches
[Philippe Le Hégaret introduces the session]
[Panelists: Frank Olivier (FO), Håkon Wium Lie (HL), Jonas Sicking (JS), David Baron (DB)]
FO: we have been releasing new versions of IE9 platform preview at regular intervals with new standards-based features being added
[Frank demos a HTML5/CSS3-based site]
FO: we run everything through
graphics acceleration, through GPU
... Performance is one feature that everyone seems to like.
[FO shows some SVG love]
FO: critically important: developers
want to create once, run anywhere
... interoperability is important
... We need more and more test suites.
[FO demos how to zoom around the world over the Web]
FO: this site uses HTML5 canvas
... we have been updating our preview every 8 weeks
... We released IE9 beta on Sep 15
FO: added 2D transforms
HL: by adding two lines of code in
your CSS stylesheets, you can have downloadable fonts in your
... text shadow was in CSS in 1998
... CSS3 border radius
... this was a request in the CSS1
... in 1995, I told Bert, nobody is going to want rounded borders
... that is _so_ 1970s
... we need to communicate through aesthetics
... this is stuff that people used to spend hours and hours doing in Photoshop
... and it can now be done in one line of CSS
... CSS lets you avoid the need to use images for cases where you don't really need them
... Media Queries let you say stuff like, when the screen is narrow, style the content differently
... forms are easier now with HTML5
<koalie> håkon++ for scary photo in extra narrow window
<Yves> håkon++ indeed
HL: you can add "autofocus" and
... to form elements
... you can specify ranges
... the browser can help you
... how many hours have been wasted in the western world creating date pickers?
... now the user agent can generate the date picker for you
... this page has text with logical flow
... this next page shows SVG
... video is the next battle
... we need to find a common format
... Google spent a lot of money to acquire WebM, and I think this is it
... this is the Wikipedia Lexicon
... CSS has a paged media module
... two years from now, half the books on the NY Time bestseller list will be formatted using HTML and CSS
... the printing press
... we would know how to use a Gutenberg bible if we had one here today
... we need to create standards that will last for 500 years
JS: we are going to show some fancy
demos containing lots of fluff
... we need to get video on the Web working
... we are using CSS transitions in this demo
... this demo shows video interacting with the rest of the Web page
... we using SVG clipping in here too
... here is another example that shows a drive through Dubai
... because it is an HTML video instead of a plugin video, we can interact with it much more
... this uses WebGL
... to give a 3D view
[JS turns of CSS to show that the video is displaying 3D at all times, and CSS is used to transform it to an interactive view]
... real time
... and ray tracing
... fully real time
... we are using the graphics cards to do the heavy lifting, all the calculations
DB: demo of OpenType font
... fonts have these features already
... but you couldn't use them in Web pages before
... this font in this demo has proportional numerals and automatic fractions
... and for tabular data
... font has better numerals for use in tables, to make it line up nicely
... also can access fun features from fonts
... can combine WebGL and Audio API
... audio API allows you to get data out of audio elements in the browser
JS: this demo is all WebGL
... integrates with Flickr and Twitter
[audio plays while WebGL stuff renders]
JS: We are using the Audio API to do
... and then using WebGL to do visualization
[images pulse in time to audio track]
PLH: I take questions now
[Doug Schepers, W3C, comes to mic]
shepazu: we are starting an Audio
... not just an Incubator group
Roger: Roger Cutler, Chevron. Håkon, what are you using for the music notation
HL: we are using ABC notation
<timeless_mbp> [ABC - http://abcnotation.com ]
Claudio: Claudio Venezia, Telecom Italia. Do you plan to address Augmented Reality use-cases?
JS: big missing piece is pulling data
from a video camera and/or audio input and getting it into the
... and how to do streaming
... work is being done on this
PLH: HTML.next session will talk about the proposed <device> element
Roger: there is no standards-based music notation that exports/imports well
<timeless_mbp> [Sibelius - http://www.sibelius.com/]
<timeless_mbp> [Finale - http://www.finalemusic.com/]
HL: point is, with a little bit of markup, you can generate music
Roger: would be nice if some standards body provided a standard for music notation
shepazu: MusicXML guy is in the Audio XG
<koalie> Audio XG
<timeless_mbp> [MusicXML - http://www.recordare.com/musicxml ]
[round of applause for the panelists]
[Paul cotton chairs the session]
PC: PLH, you owe me 12 minutes
... We stand between you all and lunch
<rigo> background http://www.w3.org/2010/11/TPAC/HTMLnext.pdf
PC: I am one of the co-chairs of the
... this session is about answering some of the questions from the previous session
... each speaker gets 10 minutes
... Patrick Dengler (PD) from MSFT will be first
... Larry Masinter (LM) has a series of questions to ask you later
... and this may involve humming, IETF-style
PD: what an amazing amount of
technology is in HTML5 today!
... so fantastic
PD: tectonic shift
... I believe there is an opportunity for us to finish HTML5
... developers see a lot of technologies coming together in HTML
... we have a chance to bring these together, establish a plateau
... ask, are we done with SVG 1.1?
... can we agree, these are the modules we can all agree to depend on?
... communication is not yet where it needs to be
... can we find that plateau, just as we did with HTML 3.2 and HTML4
... what are the features that I as a developer can depend upon?
... we owe more testing to the developers
... why don't apply the engineering practices that we have, and take those to the standards committees that we participate in?
... I am not talking about stifling innovation
... some things are vendor-prefixed and still in development
... we don't want to stifle that
... let's do a fast, forced pushed with specific timelines
... we want "write once" to the Ubiquitous Web.
... this is sorta what's in my paper, see that
[time for questions]
[Jeff Jaffe (JJ) steps up]
JJ: after a release with an incredible amount of functions, the next release is typically a bug-fix release
PD: yeah, there could be a dot
... but we do not want to spend years on that
... let's lead with test-driven specs so that we don't have to spend as much time in development next time
[Daniel Glazman (DG) steps up]
DG: we have 20,000 test we have been working on for CSS
<darobin> Election results in SVG, as Patrick requested: http://election2010.talkingpointsmemo.com/all
DG: for HTML5, you are going to have
... you are not going to finish this until 5 years
PD: could we do it on a scenario basis?
DG: do test-driven development on
your specs, and you do not run into these problems
... I don't think so.
... developers are going to do things that you do not expect
... you cannot ahead of time extract a scenario
PC: I was chair of the XML Query
... that test suite has 20,000 tests
... I agree that testing is important
... and you not going to be able to do it with just the current resources
... we should be having a dialog about *how* we are going to do it
DG: I don't disagree.
... but writing tests requires a lot of knowledge, and a lot of time
... the workforce required is huge
... but that is not what my comment was about
... users are always going to go deeper than what you have in your scenarios
<rigo> [Real World HTML: the <device> element by Adam Bergkvist (Ericsson AB)]
AB: I have been asked to talk about
the <device> element
... the device element represents a device selector
... allows the developer to access devices that aren't normally exposed to Web applications
... device picker
... when you install Skype, you know ahead of time that it is going to access your microphone and camera, and that is why you are installing it
... but when you access a Web app, you don't necessarily know what it might be going to access
... input element was judged to be not sufficient for the use cases we needed to address
... hence, the device element
... the device element itself is just a first step
... you then need APIs that go along with the element
... so far we have only one such API
... and that is a "stream" API
... event triggers on the device element
... set the stream to the video source
... can record while I am showing it
... brings a lot of new possibilities
... upload videos to a video blog
... adding communication to a game
... but in addition you need a way to share streams
... there is a proposal/applications of some kind for that
... we did some prototyping this spring and summer at Ericsson Research
... we used the device element and stream API to develop a prototype
... along with some media element extensions
... we invented something we called "media stream transceivers"
... using HTML5 WebSocket
... there is a lot of work to be done before this can be made real
... if you once allow a Web app to access your camera, how do you revoke that permission later?
[Adam concludes, floor is open to questions]
PC: where is this in the W3C today?
AB: this work started [in discussions on the WHATWG mailing list]
Dom: Ian Hickson has proposed this <device> element, and there have been discussions about it in the DAP WG
[Robin Berjon (RB) arrives at the mic]
RB: DAP is quite willing to work on this, we can get discussion going
[Tim Berners-Lee at the mic]
TimBL: you mentioned security, should the security be part of the html mark up thereby phishable or part of a drop down from the browser
AB: this could be done something like
we have now with the file-upload feature
... it can be done as a [replaced element], rendered natively by the browser
[Dan Appelquist (DKA) at the mic]
DKA: I think this belongs in the DAP
... because they are already working on the related APIs for device interaction
... and that also have privacy and security issues
PC: Why not do it in the HTML WG?
DKA: Why *would* it be done in the
... I don't think the HTML WG is where the expertise is. (Is this where the blood starts?)
[Thomas Roessler (TLR) at the mic]
TLR: This is a bit like the discussions about where the work on RDFa should take place
PC: rechartering HTML we have to put all those things on the agenda
PC: queue is closed
[David Rogers (DR) from WAC]
<timeless_mbp> [WAC - http://www.wholesaleappcommunity.com ]
DR: I think it is folly to put
everything in one working group
... we are looking at related use cases for Widgets
PC: my point was about the
... not about the APIs
DR: maybe we just need to work more closely
[Jeremy Orlow (JO) from Google at the mic]
JO: I think this clearly does not belong in the DAP WG
<chaals> [DAP and Web Apps (and a couple of special-purpose groups) share responsibility for the APIs]
JO: the APIs that you need in Web browsers are necessarily very different from the ones you need for Widgets
<chaals> [/me wonders what Jeremy thinks browsers are doing...]
<jorlow> right, I don't think I explained very well. I'm not saying that DAP is fluff or anything like that. My thesis was intended to be that camera access is a good example of why we need 2 APIs, not just one because the security requirements are pretty fundamentally different
<jorlow> security/usability that is
[the word "bullshit" is heard]
<DKA> +1 to Robin
<DKA> -1 to Google's attempt to sow FUD on the work of DAP.
<rigo> [XG Audio: Advanced audio functionality for Web browsers, by Chris Lowis (BBC)]
CL: the demos I'm showing were created by others
[Chris shows a demo]
CL: this XG is chartered in part to explore synthesizing audio/speech
<Ralph> [scrolling numbers]
<Ralph> [dynamic audio waveform]
CL: the API gives direct access to raw samples from the audio stream
<Ralph> [different dynamic audio waveform]
CL: so you can create
... and you can create filters
... visualization of FFT
<Ralph> [other visualizations]
... filter out the frequencies, do beat detection
<Ralph> [flashing text sync'd with music beats]
CL: can do visualizations using 3D
... Chris Rogers at Google has come up with a proposed API
... Chris Rogers proposed API is based on the conceptual model of audio "nodes"
... that can be passed around, chained together
... e.g, using a connect() method to connect a low-pass filter and gain node together
... this code is in the WebKit code [and you need to build/compile it in yourself or download and run a special build to enable it]
<rigo> music: An die ferne Geliebte
[demo of music playing along with score]
[demo of drum machine]
CL: can apply audio effects in real time
<Ralph> [Chris manipulates sliders in the browser to adjust the effects]
CL: those are demos from both camps
in the Audio XG
... we would like to talk to people in the [DAP and WebApps WG] and also [WAI]
... Al McDonald is involved with plans for setting up a full WG for this
[Roger Cutler (RC) steps up to the mic]
RC: it's more in the spirit of the Web to go after the notation
CL: the markup in the demo I showed is using MusicXML
RC: MusicXML [doesn't go far enough]
<rigo> [HTML6: Core Web Standards: Renewed Perspectives, by Larry Masinter (W3C TAG, Adobe)]
LM: program has me talking about
future directions for HTML
... but hearing the desire to see some "blood on the floor", I prepared a different talk.
... I would like to try to categorize some of the issues that have embroiled the standards community, and the HTML WG in particular
... plan is to hum for the statement that you hate most
... first, should standards "match reality"?
LM: point is, people arrive with different assumptions
<hhalpin> there is also the disturbing possibility that standards neither match reality or have any principles :)
LM: next, reverse engineering
<hhalpin> but no-one wants that luckily!
LM: Precision- how precisely should standards specify [UA] behavior?
<timeless_mbp> "Standards should be [as] loose as possible"
<noah> Mr. Session Chair: you're not distinguishing the case where two different people hum for the two choices, vs. the case where the >same< people like (or dislike) both
[rousing hum for standards precisely defining UA behavior]
LM: Leading- should standards lead the community, or follow innovation
[hums show more support for standards being driven by innovation]
LM: Extensibility: are non-standard
... Modularity- is modularity valuable, or disruptive?
... Timely- do standards take too long to develop?
... Authors ignore standards- can we expect authors to follow standards that include "authoring conformance" requirements?
... Always-On Committee
<noah> I feel that Larry is doing a very useful service by pointing out different perspectives that people have...I feel that 'voting' on these is confusing and counterproductive. In many cases, these are two sides of the same coin, and it's important to trade them off case by case.
LM: are open-source implementations
... Browsers and the Web- how important are browsers? Are other application classes equally important?
... Forking- is forking harmful, or does it enable further innovation?
<Ian> [Forking harmful won out]
<chaals> [Hmm. I am finding a lot of these questions are setting up false dichotomies...]
<noah> Why ask whether >browser< can depend on patented technology? Geez, they all run on patented chips. I think you >mean< to ask, should the specifications >require< the use of patented technology in order to produce a viable implementation.
LM: Accessibility- how critical is it
to address accessibility needs?
... Architecture- how important are architectural concerns?
... We are going to get HTML5 out without solving these issues
... so we need to address them in HTML.next
... maybe we will need two specs
... you can't solve these problems unless you first acknowledge them
[Jonas Sicking (JS) at the mic]
JS: there was not a single one of
these questions where there is only one answer
... it all [depends on the context]
... there is, for example, no clear answer to the question about "Should standards lead, or should the follow?"
... there was a clear undertone in these slides, about HTML5
... and HTML5 approached all of them in a variety of ways
<Steven_> +1 to Noah
[Noah Mendelsohn (NM) at the mic]
NM: you have given us some valuable things to think about it, but I think it was a mistake to [do the voting thing]
[Harry Halpin (HH) at the mic]
HH: I think there is a false
dichotomy behind [a lot of the questions in your slides]
... maybe there is a 3rd option
... which is, drive things based on what actual users really want
LM: I think the issues of modularity
need to be addressed
... unfinished old business comes up and bites you in uncomfortable places
[someone asks, "Is the room secure?" (for lunch)]
[Paul Cotton replies, Is the Web secure?]
[We break for lunch]
[jeff gives summary of the morning messages]
Jeff: integration is hard
... HTML5 is cool
... There is life after HTML5
[Jeff thanks organizers, program committee, and others involved in making the meeting]
[Jeff thanks various people by name]
Marc: "emotions are the social glue in human experience"
<koalie> [Emotion Markup Language was a W3C Incubator Group from Nov. 2007 to Nov. 2008]
[screen shots of various Web sites]
Marc: e.g. social web annotations; automatically sensing emotional behavior; voice/avatar generation with emotion
Marc: research labs
Marc: we've done 2 incubator groups, MMI WG now going towards Last Call
Marc: manual annotation of data /
automatic recognition / generation
... EmotionML is very diverse, broad variety of contexts
<jeff> Jeff: What would he have said if he had another minute?
Harry: Knowledge is (potential
... "things are looking grim"
Harry: Facebook is taking over
... but SN is larger than FB
... central to how we use the Web
... but most social nets are very insecure, using cookies
[slide, too many solutions]
[slide final report]
Harry: 30 pages report to summarize
the social web space...
... need work on Identity in the Browser
<koalie> W3C Social Web Incubator Group
[Social Web XG report at http://www.w3.org/2005/Incubator/socialweb/wiki/Main_Page ]
Harry: recommending an over-arching privacy activity at W3C
DKA: Dan Applequist. As co-chair [of
the Social Web XG]... please do ask us about the content of the
... it has some important ideas about what W3C should be doing about the space
... and risks about the way people identify themselves on the Web...
... I've just tried firesheep and now I can impersonate Anne Bassetti.
<AnnB> luckily my named is spelled "Ann"
Harry: if these issues are not addressed it could be the i-9/11 forcing people to closed platforms
<fantasai> very coherent presentation by harry IMO
Dan: hands up if you're using mobile devices that let you input data with speech
[lots of hands]
Dan: hands up only if they use standards
Dan: HTML Speech Incubator Group
<koalie> W3C HTML Speech Incubator Group
[slide 3, So Far...]
Dan: collected use cases, etc
[slide 4, Next steps]
Dan: no results yet
... just getting started
... want to propose our spec to HTML WG
Dan: How is this related to VoiceXML?
- it's a separate group and work, intended to be light-weight and
intended to go into HTML, VoiceML is for longer-term spoken
... we're still working out the boundaries.
[slide 1, Why is SW Different?]
Roger: scalable... and, open world assumption
[slide, open & closed worlds]
[slide - Oil & Gas closed or open?]
[slide, data silos]
Roger: "bad, evil"
[slide, tower of babel]
[slide, SW, connect the silos]
Roger: silos are effective, so SW recognizes this and connects them with ontologies
[slide, HCLS linked data cloud]
[slide, chevron SW Activity]
[slide CiSoft proposal]
Ian: Ian Jacobs. Roger, can you describe what it's like being inside Chevron communicating the beauty of SW to your colleagues?
Roger: The idea of SW is high visible
to our executives
... 2 reasons... we emphasized what you can do right now
... e.g., metadata managements
... and strategically in the centre of some of our big projects
... talking about workflows that span domains
... so it resonates, the big picture resonates very strongly
Ashok: RDB2RDF WG is going to write
... you can do the mapping and store it, or a virtual mapping and translate from SPARQL to SQL on the fly
... there's a huge amount of relational databases
Ashok: 15 Billion Dollar Business...
Ashok: Relational data wants this
'cos it wants to have semantics
... connect the silos
Ashok: slide History
[slide, Work in Progress]
Ashok: we have requirements and
... just published Custom Mapping - lets you use the Full Power of SQL to translate from relational data to RDF
... and in next month or two, FPWD of default mapping
... fairly simple db, press a button, and RDF comes out
... and maybe we'll write a formal semantics for the mapping
[slide, details, not covered in talk]
Artur: Artur Ortega, Yahoo. When
you're moving info from your silos up to RDF to the SW
... in the silo it's easy to find a unique ID
... e.g. a person's name
... but in the SQ they may not be unique, how do you deal with that?
Ashok: that was on my last slide
which I skipped!
... for default mapping we have a scheme to generate URIs
... but it's an open question we're working on
<DKA> On Firesheep: I was scanning actively for about 5 minutes and I picked up credentials enabling me to masquerade as Ann Bassetti, Phil Archer, Richard Ishida and 3 others (all now cleared). TO PREVENT THIS FROM HAPPENING TO YOU: use https: whenever possible (Twitter and Facebook both have https versions) or go through a VPN whenever you are on an unsecured (by WPA) wireless LAN.
Henry: Henry Story. On Social Web XG, trying to get an Open ID XG going
Harry: an identity workshop is under consideration.
Roger: Roger Cutler. Hoping that kind
of question can be a push and pull
... I don't view it as a technical question of how you do it automatically
person2: webID is a very low-level specific technology
Harry: we're planning on looking at
Henry's work with OpenID foundation
... and all sots of groups
... it would be good to talk with people who are here
Phil: Phil Archer. I think today was the first redacted slide I've seen, thank you Roger!
Phil: I have worked on lots of
things, including radio, so they thought I would be good at
chairing this panel. Go figure...
... Large panel, reflecting the fact that there are a lot of different threads in the relationship between TV and Web.
... Speaker: Kazuyuki Ashimura W3C, Tristan Fern BBC, Libby Miller NoTube, Yosuke Funahashi, Dong Young LG, Mark Vickers, Comcast.
... Kaz, can you set the scene and talk about the workshop in Tokyo and possible repeat in Berlin?
Kaz: Held workshop in Tokyo, in September, on TV and Web and possibilities for smarter integration of the technologies. We had 150 people, and got a lot of interesting and somewhat complicated use-cases to motivate work. Conclusion included creating an Interest Group as a first step, and that is now being brought to W3M for charter review.
Tristan: Work for BBC Research and Development, in a prototyping team
looking at the intersection of audience and technology. Can be TV,
Radio, Web, ...
... BBC is interested in delivering TV via Web, and how to deliver web applications to TV devices.
... Personally interested in how we create new experiences to combine TV and Web things. How can we join passive TV audiences to active Web audiences.
... and how do I get BBC talent to make stuff for the Web. Can we join search and recommendation from the Web with TV programming, archives, schedules, ...?
... Right now working on "second screen" app, using another device to talk around what people are watching.
Libby: NoTube is a social web and TV program. We're
interested in what people are doing now - using second screen and
social media at the same time, commonly watching it with other
people in a room.
... Making it easier to find and share cool stuff.
... Looking at things like realtime voting, interest level for long-tail content, and so on. Things like the mix between broadcast services and URLs...
... Interested in persistent pointers to TV things, metadata, ...
Yosuke: Tomo-digi (our company) is about Web and TV
convergence. Established 1999, when internet services on TV started
in Japan. Now over 120 million DTV sets, and doing a vast range of
internet services with broadcasters, so we have market experience
about Web and TV.
... My interest in Web/TV IG - Web is spatial concept, but TV is a temporal concept, with a timeline. When Digital TV started, there were several kinds of web standards. Broadcasters developed based on Web stuff plus broadcast extensions...
... Broadcasters extended different things. Now is a time of convergence for broadcasters. HTML5 is coming, and TV broadcasters should think of "one TV" in the sense of "one Web". This will be a really new approach that will bring us really new technology
Phil: Product cycle for phones and laptops are measured in months - but TV's stay there for years.
DY: Dong Young. LG make phones, TVs, etc. We want some web technology as a
platform for smart TV. I want 3 things:
... unfragmented market. Not different specs in different markets
... realistic standards.
<Ian> (as global vendor need unfragmented standards)
DY: Royalty-free standards
... TV is different to PC or phone. Harder to do input, big screen, ...
... We want new standards from W3C that are free, unfragmented and realistic.
<Ian> "and are willing to do this work here"
Phil: Note that you might have a good screen on your TV, but the electronics inside are dedicated to the picture, not processing power.
Mark: Mark Vickers (Comcast). Got involved in TV/Web stuff in mid
90's. Produced stuff for IETF, and so on.
... pretty sure none of that is in HTML5 - that's what happens when you are too early.
... Comcast started doing stuff on set-top boxes but is now focused on delivering to any device, and we're looking to HTML5.
... We currently use plugins for the web, but would like to deliver through HTML5, and we want 'one web'. Will only happen if the standards and support needed for TV is done in W3C. Otherwise we have today's mess of different extensions in different groups.
... Looks like KeyCodes are going to be OK (was a challenge). Access to content on set-top boxes via LAN is a challenge.
... External challenges include dealing with codecs that support captioning (for regulation requirements) and protected content, ...
Phil: Recurring theme - multiple
screens with different devices. People are still passively watching
TV, but there are things people do on the side of that.
... Libby, can you talk about the use of hash-tagging and social networking in relation to TV?
Libby: This may be just UK
phenomenon, would like feedback. In UK, you can get 8 of twitter's
top ten trending topics about a single TV show.
... programs are picking up on it, giving hash tags in programs, having people talk at the TV through twitter.
... When digital economy bill went through parliament there was a boring misinformed debate in parliament, but a lot of chatter and further information around it from people watching and interacting.
... The further information can be used later...
Phil: Do other countries have this twitter phenomenon?
[scribe hears 'france']
Kaz: TBS in Japan also did some stuff with live feedback around elections.
<AnnB> I don't watch TV, but I thought that tweeting in realtime phenomena might have gone on with program "Lost"....??
YF: Purpose of the program is to bring people not currently using Web, e.g., older people not accustomed to the Web, and show them what happens with twitter.
<AnnB> that was the real me ... muhahaaa
<AnnB> I'll never tweet again, now that I know about firesheep
Phil: 2 devices, one showing program, one being used to chat on web about it. DongYoung, do you want to see multiple devices? What sort?
DY: We make these kinds of products,
so would like to see them develop. If you have loose linkage (e.g.,
share a URL) between 2 streams of information, we can't do tightly
... We can put real information on TV, put with twitter can go the other way. Incorporate what user does with mobile device and reflect it in TV content.
... Could want something more advanced than twitter. Our priority is standardizing extensions - APIs for TV. And other use cases.
<kaz> [kaz remembers the impressive demonstrations by Japanese broadcasters during the September workshop ]
Phil: So if the APIs are in place we could do something more interesting and exciting than just tweeting.
Mark: Have been recently showing an
app where you can use your tablet to read the program guide and
change the channel. Basic 2-screen interaction.
... When TV program producers integrate this into content will be interesting. Above a certain age people just watch TV, but younger people almost all have a device with them when they are watching TV.
... Good examples in video game world, but nothing like that in TV yet.
Phil: Phone as remote control...
Mark: Comcast is a cable company, but also have some program channels. We're in process of acquiring NBC and Universal studios. All this content is going to be available in some downloadable form on the web. Convergence in delivery is leading to convergence in how companies are formed.
Phil: Can you give us some blue-sky thinking of what you would do if the APIs are there...
Tristan: Normally think of how to use
exiting technology. Interesting to think of new technology.
... Lots of interactive game show stuff - kind of obvious. We are looking at what you would do with e.g. documentary content and second screen to make it more interesting.
... Radio is interesting - at particular times of day it has massive live audiences in UK.
... Definitely social stuff around that, and extra information you could add by web.
Phil: What's missing, or hard?
Tristan: Finding and recommending TV content. Is done for web, do we need anything special for TV? How do you synchronize web content and live TV - delivering a picture relevant to program synch is not so important. Delivering captions or subtitles have to be very tightly synchronized.
Phil: I'm share name with a character who has been in a radio program forever. Am in a group of twitter people who are tweeting about the radio program. Rule is not to tweet spoilers, because we are using time shifting.
chaals: Web on TV has a lot to do for
... minimum thing we have to do to be as good as TV already is. What can we do beyond that?
MV: There are basic things already -
captioning is required. Requirement in US is to have parental
control system built into TVs sold. Those will arrive in Web when
government wakes up.
... Also multiple audio, e.g., for second language, or for descriptive audio.
... Another requirement is emergency alert - you have to be able to take over the content and provide emergency information like "tornado coming".
... what does that mean on the web - how do we do that push, do we need to do them?
... Right now, things like captions and parental control are done on the TV device itself. Even a set-top box just passes that directly through.
... If a TV has a browser, does it use that to do captioning, or does it pass it into the codec?
... these are unanswered questions.
Phil: It is depressing to hear 'we do accessibility because government requires...' - is there anything more that people do?
MV: Now we are looking at things that
go beyond requirements, can be easily delivered for relatively low
cost. Using interactive programs to deliver annotations adapted to
individual users. These are exciting steps forward.
... We start with government stuff because we are not allowed to deliver signal until we have done them.
Jeff: TV industry is diverse - people
make devices, software, content, regulations and requirements. And
these often differ by country. How do I think about getting a
consensus of this industry?
... You talked about multiple configurations and models and use cases. Is there a succinct use case hat needs to be focused-on by a subset of the industry to bring TV to the web, or is this a complex multiple-stakeholder topic that requires us to get each group of stakeholders at the table....
DY: I think there is a consensus that we as the TV industry have urgent need for some minimal spec platform for smart TV. What we need to do here is unify the various existing specs and defragment them. Quickly, before it is too late.
MV: There is a way to liaise with
other groups, and think that is important: DLNA (TV manufacturers
and broadcasting delivery folks) is one. They don't write standards
they pick among existing ones.
... driving more unified requirements to those groups would be helpful.
YF: From content provider viewpoint
there is a huge value chain, and in different countries the shape
of the value chain is different. I think we can reach agreement
about "one TV, one Web" around HTML5...
... We worked on java-based system 5 years ago, and compared to BML (XHTML1.1-based, used in Japan).
... We conclude that both languages can meet use cases. So constraints are resources, and the viewing styles of audiences. There are many sentences, but the semantics are the same at the base.
Kaz: Please note that requirements
are different by country. We held workshop in japan, we are
planning another one in Europe in March.
... We have proposed IG with 4 co-chairs - Opera, NTT/ITU, Tomo-Digi and LGE.
... we want to work with broadcasters from many countries
TF: Taisuke Fukuno (Jig.jp). Mobile browser vendor. Many complex problems. How do we define TV streaming and broadcast protocol for HTML?
Phil: The answer is that we have an
Interest Group which will define working items, and hope for a full
Working Group, who will answer such questions.
... We need standard APIs to unify the experience, whether from the TV on the sofa, or on the move, or ...
<Ian> [Phil's phrase is quite a nice WG name: "Augmented TV Experience"]
Phil: The Interest Group we are
putting together to deal with this, and hope that if it is
interesting to you, you will be there.
... When we started Mobile Web work, NTT DoCoMo came and asked 'where were you 5 years ago when we needed you?'
... Timing is critical - and we have run out of it. So please, join the IG.
[Phil apologizes for misunderstanding the question which may mean it didn't get a clear answer]
[scribe is worried that he may have misunderstood it for minutes too..., but hopefully not :| ]
[Moderators: Marie-Claire Forgue (W3C) and Ian Jacobs (W3C)]
<rigo_> Slides: 3D on the Web
KS: so many good graphics cards, but
no 3D on the web, we have to adapt the web for 3D
... must be in HTML to get it into the DOM
... W3C should play a role here
... two project XML3D, introduce some stuff in HTML
... X3DOM (Johannes) X3D concept and add W3C technology to it
... specific things like 3D events
... implementations on chrome and firefox, will be a hot topic on the web
... will launch an XG and invite participation
JB: X3D/DOM integration,
W3DConsortium is a liaison since year
... already used over the last 12 month, showing demos
... simple append child and remove child to manipulate, chain queries on your data
... can also play that on top of mpeg4 movies
... DOM level 2 mouse events ...
... so far no W3C API to access the camera of laptop, so using flash here
... to get this video into the browser
DKA: Dan Appelquist. Noting: Siggraph 1991, seen demos on computer graphics, so many demos around 3D. Since 1991 have enormous industry building around 3D. We are at the same turning point, see an industry developing here
<Ian> "video took off when video was available on the web; waiting for the same thing for 3d"
KS: video took off when there was video on the web, we hope for the same in 3D
<DKA> ...e.g.: Pixar. Who will be the Pixar of the 3d Web?
MW: on behalf of Andy Braun chair of
the POI Group who sent me notes
... showing tubes, getting more and more detached
... new user interface paradigm, augmented reality, becoming hot topic
... W3C looking into it, how web and AR could work together
... AR experts got together, where standardization is need. Information about location and ?? is needed
... more uses to POI than augmented reality, Foursquare and others, navigation apps, searching, wikipedia...
... Group is doing format to represent information about places. Public mailing list, encourage everybody to join
... looking for future places where AR makes sense on the Web
... see how it works with web techno
... see http://www.w3.org/2010/POI
MK: HMTL WG and Accessibility
... people getting creative, space invaders in your browser in canvas
... need some fallback content for canvas
... for image here is easy, need a shadow DOM, will only be used if the browser doesn't support
... focus on bitmap for key readers
... caret position for editing text
... games appeared (infinite adventure 3D..)
... limited number of characters facing the player
... main challenge in game is reaction time, hinders disabled people to play, there are techniques, readers
... but things go directly to GPU and GPU gives no alternative access
... second life don't have alt text, so difficult to move with alternative
... W3C should get into contact with stakeholders as accessibility bolt on will be ugly
<rigo_> Slides: XML Performance
MZ: performance is a feature that
... who cares for XML, Web? -> many hands
... evaluate performance on the web
... could be an IG or a WG, a new way to express performance
... XML still perceived as slow, real problem
... standards and products are mature, tool chain available
... native XML databases very efficient (other examples)
... also on Size: EXI
... keep all the info-set and get size smaller
... missing tools to improve performance, missing communication
... encourage to organize a workshop on performance.
Frederick: Frederick Hirsch. XML 2.0 Signature performance
Roger: Roger Cutler. Test-suite associated on EXI.
MZ: not focused on performance and tools
[Chris Lilley (W3C) presents WOFF: Benefits beyond Beauty]
CL: how many people (fast
... fonts not really new
... CSS2 had web fonts, intelligent matching, all sorts of stuff
... format wars, no decision where to point to
... SVG WebFonts, mandated format, on some platforms (mobile) but not on desktop
... 5 years of massive confusion and we got WOFF
... font vendors coming in... creating WG, => REC
... expect LC and CR cool
... not a format format, compressed font tables, licensing information
... can do by itself nothing, but combination with CSS it can do interesting things
... showing all kinds of scripts
... instead of using pictures as fonts
<Steven> is there a list somewhere of available fonts?
Artur: Artur Ortega, Yahoo. If
wingbats fonts are used, if symbols are used?
... how is this addressed to the textual alternative
CL: people rely on hacks today, will have correct font in the future
AO: what if you have flying pigeon?
TV: file of Unicode that has a pigeon-character
CL: you can use a freefont and embed
AO: what if images?
CL: you're using text.
TLR: people have concerns
... privacy by design is on the agenda again
... evercookie is just one example of many
... canvas is also
... fix canvas?
... no, same origin
... evercookie forces tracking on people and do not allow them to use tools to get rid of it
... we need to talk about what the role of the vendors is
... we need to talk about role of W3C
... W3C has organized workshops in November 09, July 10 and October 10, Next Workshop on 8-9 December
... deadline is on 5 November. Questions?...
DKA: I did not say "let's fix canvas" but I'm not saying "as intended" either. Evercookie has shown that we have to consider that question
tlr: what canvas does is that it works within a given same origin context. But how we let users control what happens in this context
David: David Singer, Apple. How penumbra of privacy goes. We should not just blindly implement specifications without privacy considerations
tlr: precisely the aim of the workshop.
<DKA> +1 to Thomas's talk by the way - we need to ensure that the public perception of HTML5 is privacy++, and the current message from press is very much the opposite, eg: http://www.nytimes.com/2010/10/11/business/media/11privacy.html
<rigo_> Slides: Web and philosophy
AM: first web in philo event in the
world organized in Paris
... philoweb 2010, conducted interviews with experts on knowledge engineering
... had a mix of engineers and philosophers and computer scientists
... topics ontology ... (see slides)
... abstracts are online, slides on slide-share
... IAWP International Association for Web and Philosophy
... new installments under way, London and Mountain View
... see conclusions on slide. Questions?...
Sarah: Sarah-Jane Farmer. Lot of web
stuff seems to draw in AI, have you been using the connection that
AI was drawing a lot of discussion from philosophers
... yes, Harry knows more
tlr: privacy challenges are partly practical, but also philosophical, looking forward to hear back
IJ: I wanted to do this as MIME
... but I'm not the MIME/type
IJ: presenting obstacles in
participation of W3C (unfamiliar culture)
... new standards TF studied prob space
... primary use case, some people in W3C and outside W3C want to cooperate around a non-mature thing
... anybody can participate at zero cost
... balanced patent policy
... OWF found out a good compromise
... no time limitation
... smooth transition to standards track, but not forced to go REC track
... also bits for liaisons
... there will be a forum, with help we can make it shiny
... momentum to create community specifications, will look differently
... Community Group is zero revenue and low cost
... created other groups called business groups with some staff
... requirements WGs to have hi bandwidth with chairs
... constructive participation improves our work
... raises number of people coming to standards process and helps us find customers
MC: MSFT gives an official blessing
to this new proposal
... we want to see the next generation web standards happen at W3C
... to do that there have to be process innovations
... people in the community have to be responsible, drive architecture, brainstorm for new ideas
... use case we see: person or group has some nice idea, want a way to get a proposal uploaded and visibility to get comments from community
... with some confidence, this needs some straightforward IPR policy, OWF is good starting template
... community spec... get to REC may even go to International Standard via W3C PAS status in ISO
... we wanted to see it under W3C umbrella, but people don't want to sell their soul in order to participate
... IJ talked about throwing ideas
... 3D canvas, a lot of attention, great
... all browser have debugging API, do we need a standard for that, let's talk about it
... idea of profiling W3C RECs, e.g. for email, why not community form for profiling
... talk about it
... Jquery note lists is a powerful idea, it is worth integrating into the other APIs?
... getting the end user and web developers/publishers into discussion what the real interoperable subset should be
IJ: mentioned OWF licenses, current proposal adapts them, are in close contact with Larry Rosen and ???
Roger: What would prevent a community group to do harmful, illegal, unethical stuff?
IJ: That's the definition of a Business Group ;)
IJ: we will install mechanisms to
prevent that on a social level...
... currently also managing discussions, also talking to Tantek
Daniel: HTML email, had workshop,
main prob is change of renderer in MSFT mailers
... problem of rendering engine
JB: at german W3C days, asked about
developers fees and positive responses
... perhaps consider lower fees for developers and freelancers
IJ: community proposal is not from scratch, have experience from XG
Robin: DAP image prob
... some people think that DAP is working for widget API and that it could not work with a browser
RB: caring for privacy and security,
following the browser security model, looking into metaphors
... no metaphor for deleting content
... also other stuff that is not very browser centric, but does not mean that this will have priorities
... where comes the heat from?
... looking forward for solutions.
IJ: 2 communities have started communications, some barriers were overcome, some persist
RB: those people saying DAP is not following the web model should come talk to us
IJ: I heard Robin offer to buy a beer later
MCF: many thanks to all speakers and the audience
Jeff: Let's address whatever is on
... I'm inviting Tim on stage to be available to address comments.
MZ: Mohamed Zergaoui, Innovimax. We had an item come up recently around zip. It has a fuzzy license. I'd be interested in feedback as it's being used in widgets, as the ?? working group would like help resolving this.
Jeff: I don't know anything offhand. If anyone knows, please come to the mic.
Jim: Jim Melton, Oracle. It will be addressed next week at the JTC-1 Plenary.
Jeff: We'll be sending someone to the
plenary, so we can caucus there as well.
... I take it everyone is satisfied.
<koalie> [please, fill out the feedback survey: http://www.w3.org/2002/09/wbs/35125/tpac2010-feedback/ ]
Yesha: Yesha Sivan. What are you unhappy about?
Jeff: I started just a few months
ago. My tenure was still counted in days when there was the famous
article in Wired magazine that declared: "The Web is Dead"
... I was unhappy about that for a while. I thought TPAC would be a good personal test for me to answer the question: is the Web dead?
... One of my responsibilities as chair of TPAC is to summarize the day. So my summary is: the Web is not dead.
... Considering the function and scope covered. We had one day, 40 people presenting. All exciting ideas.
... I worry about how we'll get it all done, but it's exciting.
... The subtext here is you can't believe everything you read in the press.
Mohamed: I'm talking at a conference,
XMLPrague. If people know about it, please make some noise.
... It will be in March in Prague. People are welcome to have w3c meetings there too. Please go to xmlprague.cz
<koalie> XML Prague
Mohamed: We're pleased to have w3c participation.
<darobin> XML Prague 2011 CfP
Julian: Julian Reschke. I can't help but noticing that the IETF meeting in Prague starts right after that meeting too.
Jeff: We're considering having a community group on Web meetings.
AnnB: This is probably the quietest meeting I've ever been at. The room is crowded, but quiet? What is it?
<masinter> people hummed!
glazou: There was not a lot of time for questions.
timbl: I think it's the jet-lag not
working for people.
... ?? said something positive. Everyone is busy, they've got stuff to do, but it's not at the top of people's minds to complain. By continuing to talk here we're probably even preventing people from editing documents.
chaals: I think the atmosphere has been good. There's no reason to rip holes in people when they've been mostly saying sensible things.
Ileana: Ileana Leuca, ATT. I think it was an excellent session, many new ideas. A popping environment. At the same time though, at least 50% of this room are not native English speakers. You need some time to understand the new idea, and appreciate it, and by then the three minutes are done.
AnnB: It's an important point. In the AB we have talked about this before. Anyone in the room who wants to submit suggestions, that would be really valuable.
<DKA> Idea: simultaneous translation services provided remotely via amazon mechanical turk and VoIP.
AnnB: I once was told a comment by someone and encouraged them to step up and ask it, but they were concerned they couldn't understand the response.
PaulC: People asked me why the hum?
It was a good way to get participation regardless of your mother
... One of the rules in JTC-1 is that the meeting only goes as fast as the understanding of the slowest person.
... I'd like everyone to think about facilitating participation. Repeat questions slowly. Repeat answers slowly. It's simple politeness to your peers.
PhilA: In every sense, I'm a native
English speaker. I've had the privilege of touring the world and
meeting many friends around the world. I'm eternally grateful for
so many hours that have been put into learning our language.
... Maybe one reason for the quietness: the Web is a victim of its own success. Look at the early specs. The people who wrote them are too busy making a living out of it.
... All the RDF people are too busy making money off it to make RDF2.
<ivan> actually, not absolutely true, we will get some of the old RDF guys in the new group (if it is set up...)
timbl: I apologize for being one of
the fastest and mumbly speakers around.
... Often it is really hard for the scribes, but I wonder if one possibility would be to have the IRC channel with simultaneous scribing projected to the side?
<Ian> [lots of support for projection of IRC scribing]
<koalie> [+1 to project IRC scribing]
Liam: One reason we have fewer
questions: people are making more productive use of IRC and the 2nd
... We're doing very interesting things with the bots. What about experimenting with simultaneous translation?
timbl: To IRC?
<Ian> Liam: translating minutes into multiple languages
Jeff: I think it would be a great
idea to do that if only we had the skilled people to do that.
... It's not easy, if anyone would like to volunteer let's try it as an experiment.
<Ian> Jeff: Quiconque qui veut se présenter est le bienvenue.
<koalie> Ian :)
chaals: I've been to conferences and
... The trick is finding people to do it. Looking at the minutes, sometimes they're very detailed.
<ivan> I am, formally. bilingual, but I would never accept to do such a translation
chaals: The quality is patchy, but you can try.
Jeff: When I was thanking earlier, I left out the scribes.
<Steven> Stream the IRC through Google translate
Daniel: Soohong Daniel Park. As a non-native speaker, I appreciate Ann and others taking care to help non-native participation. We also have to take into account that there are culture differences. We need more time to speak up. I'll speak up more at next TPAC. I really appreciate W3Cs kind and helpful and good to non-native speakers. Thank you.
<glazou> Steven, is it fast enough ?
<Steven> I think so
Kaz: Keio provided simultaneous translation for Web on TV workshop. It was very useful.
hhalpin: From a broader perspective, with Ian's community group effort: I'm doing outreach from Scotland to SF and it's crap trying to communicate. I'm wondering how W3C plans to get in touch with emerging economies that might not know of W3C at all.
<chaals> [+1 to usefulness of translation enabling people to just speak their preferred language]
<ivan> W3C Brazil had a meeting in Sao Paulo a few weeks ago which was translated simultaneously into Portuguese; it was really efficient, but needed two professional translators to get it working
hhalpin: The second thing I'd like to
hear about is time based requirements for the REC process.
... I'd like to see it clarified and feedback on what it takes to make a WG.
... Some activities start a full scale working group based on three Universities, others need browser vendors. I'd like to hear what the right conditions would be.
timbl: For many of the things, the
questions had different attitudes. It's different in every working
group. When we started this thing we had activity statements. It
had to answer whether it was lead by industry or academics.
... I think it's really important that W3C doesn't decide that those answers should be the same for every WG.
<Barstow> +1 TBL!
timbl: We should keep it as a flexible tool. Never build something that constrains it more. Within that, each different area of activity tends to have different answers.
Jeff: Thanks Harry for broadening the language and cultural norms. One of the norms for people to learn is coming to the mic after the queue is closed.
Ian: The new community groups should have a shiny new landing page for how we get stuff done. We don't have any funding for that design.
<chaals> [controversy != doing useful work]
Ian: Tomorrow night we have a meet-up here at the conference center between 7 p.m. and 9 p.m.. Everyone is invited. 150 local developers and designers. Free food and wine.
Rigo: There was a lot of chat in the back channel where we searched for creative ideas for tempests for more discussion. I had an idea but you closed the queue, so I'll spare you it.
Jeff: In terms of the schedule: next thing for everyone to do is fill out the evaluation survey.
<koalie> Feedback survey
Jeff: Then at 7pm we have the
reception on level 2.
... Now we've learned that in 3 minute increments we can't get people to ask questions. So people should feel free to approach people and ask questions.
... After that, the week isn't over, 2 more days of meetings. As Ian noted we have a public meet-up.
... There are an unlimited number of future conferences related to Web technologies. The next Multilingual Workshop is in March in Pisa. WWW is also in March in Hyderabad. There's a Web TV related WS, date not finalized. WWW 2012 will be in this conference center.
... Our next AC meeting will be in May in Bilbao.
... It's limited to just the AC, but in this case there will also be a number of companion meetings at the same time.
Xabier: Xabier Uribe-Etxebarria. I
wanted to present the best conference, but Mohamed already did
that, so this is the second best conference.
... W3B conference and AC meeting. W3B Conference is the global event that brings together key experts and key organizations discussing the Present and Future of the Web.
... Awarded World City prize.
... Hotels, conference center, shopping, museum, all within half a mile.
... Lots of support, the Mayor and the Spanish government, everyone involved. Euskalduna Conference Centre and Concert Hall. Awarded ??. Want to fill 2,200 people
... It's the biggest stage in Spain.
... We'll be doing webcasting.
... Afterwards we'll be doing a gala at the Guggenheim. Free for AC members.
<Bert> (Xabier explained that "W3B" had the B of Bilbao, but also of Business.)
Xabier: starts on Sunday, May
... See you there!
Jeff: We do not yet have the venue for TPAC2011. We'll get that to you when it's available. See you in an hour for the reception. Thank you!
<glazou> For TPAC 2010, please, fill the WBS form and give us feedback <http://www.w3.org/2002/09/wbs/35125/tpac2010-feedback/>