18:39:28 RRSAgent has joined #pfua 18:39:28 logging to http://www.w3.org/2005/03/01-pfua-irc 18:39:33 rrsagent, make log public 18:39:38 rrsagent, make log world 18:39:40 aaronlev has joined #pfua 18:39:41 Andi has joined #pfua 18:39:56 RichS has joined #pfua 18:40:02 Rotan has joined #pfua 18:40:09 Ryladog has joined #pfua 18:40:20 joeclark has joined #pfua 18:40:22 wendy has joined #pfua 18:40:27 Zakim has joined #pfua 18:40:36 Hello, PFUA. 18:42:43 Kesh has joined #pfua 18:42:50 Meeting: PF, UA, WCAG joint meeting 18:42:57 ircleuser has joined #pfua 18:43:04 Agenda: http://www.w3.org/WAI/PF/Group/meetings/f2fmar05.html#agn 18:43:10 Chair: Al 18:43:15 ircleuser has left #pfua 18:43:18 Scribe: Wendy 18:43:19 cklaws has joined #pfua 18:43:25 janina has joined #pfua 18:43:56 dpoehlman has joined #pfua 18:46:07 louie has joined #pfua 18:46:33 http://www.w3.org/WAI/PF/Group/roadmap/ 18:46:41 m3m: concept of the roadmap (rich will fill in details later) - ensure that using w3c technologies and elsewhere will have a coherent vision of how to be accessible when all put together. 18:46:53 m3m: the doc lays out how the techs will work together. 18:47:24 m3m: from UA, having reviewed XHTML 2.0 w/the perspective of getting it to work between UAs and asst techs, the roadmap is the next logical step. 18:47:59 m3m: interested in working with PF. PF works with spec writers and UA works with browser developers. 18:48:07 m3m: want to represent the view of the ppl who will be implementing. 18:48:21 m3m: want to ensure that years down the road, we are all working from the same playbook. 18:48:33 m3m: want to get in front of what will happen over the next 5 years. 18:49:15 mc (michael cooper) 18:49:33 mc: wcag WG is working on techniques. today the techniques task force is meeting. 18:49:50 mc: we interpret how to apply the guidelines to particular technologies. working on html, css, scripting. 18:50:02 mc: also creating general techniques that describe things common to all techs. 18:50:49 mc: have identified several issues. 1. how handle issue of scripting where could take 2 approaches to make it accessible. provide functional alt or ensure that the action produced by script are accessible. 18:50:58 mc: w/out debates about if that is possible, those are 2 approaches. 18:51:19 mc: one depends on the browser's support of scripts the other assumes scripts are not supported (or can't be rendered accessiblty and requires fallback) 18:51:48 mc: 2. baseline - as we write guidelines/techniques what are we assuming is supported by user agents? 18:52:09 mc: if we have to assume that the UA doesnt' support then we may need to provide fallbacks. 18:52:23 mc: this is not desirable when many technologies could be accessible (perhaps in the future). 18:53:27 mc: have concept of "fallback techniques" that say, "until UA supports...the author has to do x instead" 18:53:47 mc: we're faced with the issue of "can we assume that user agents support things correctly" or "do we deal with the real world" 18:54:17 mc: even though our techniques documents are not normative, the techniques docs will need to be updated as technologies change. 18:57:07 mc: summarizes - we have 3 primary points: 18:57:14 mc: 1. our "baseline" issue 18:57:39 mc: 2. dealing with lack of support in all technologies, provide fallback techniques? 18:57:53 mc: 3. assume "real world" or "ideal world" 18:59:14 RRSAgent, pointer? 18:59:14 See http://www.w3.org/2005/03/01-pfua-irc#T18-59-14 19:02:35 ag: my version of why UAWG is interested, is b/c it was an attempt to take the smallest change in how ppl prepare scripted pages that would allow the UA to meet the UAAG for the scripted widgets. 19:02:51 ag: w/out heavy duty processing on the script. 19:03:01 ag: could present through an API a model of the document and it's state. 19:04:05 ag: therefore, natural progression for uawg and pfwg to discuss together. 19:04:42 ag: wcag is working on techniques w/variety of scripts. one is sripts. roadmap changes how that's done, therefore should understand so that they can carry message to the authors. 19:04:51 s/sripts/scripts 19:06:22 rs: asst tech support apis, scripted web content doesn't do that. ATs need to know active roles of elements. 19:06:29 rs: (acting as active widgets) 19:06:38 rs: have to provide state info to the at. 19:06:52 rs: accessibility apis primarily see static html today. 19:06:59 rs: changes are not provided to teh at as we should. 19:07:22 rs: typically, when someone writes in javascript, will create custom widgets that don't exist in html. 19:07:54 rs: everyone needs to udnerstand the accessibility architecture. when an at talks to an application, they are looking at an object model using MVC architecture. 19:08:00 rs: the info should target the data. 19:08:17 rs: we used to "read tea leaves" by looking at what was drawn on the screen and reverse engineer. 19:08:26 rs: we want to look at the data, obj by obj. 19:08:36 rs: when an at talks to a component can get the info they need. 19:09:03 rs: this is the JFC object info: role, states, etc. (graphic outlines) 19:09:23 rs: doesn't mean "onmouseover" which doesn't tell me much. 19:09:52 rs: text needs to deal with selection, slider: value, when changes, name, parent/child relationships 19:10:02 rs: in the DOM, have a parent/child relationship. 19:10:07 rs: important to be notified of changes. 19:10:18 rs: e.g., tree - need to know when expanded. 19:11:25 rs: javascript manipulates the dom, content, and style and is event driven. 19:12:11 rs: after csun, happy to make these slides available. can make them available in member space for now 19:12:20 wac: wcag wg and uawg need it in public space. 19:12:45 rs: basic example - use a table for layout. 19:12:59 Michael has joined #pfua 19:13:01 rs: accessibility is "hardcoded" when dealing w/html 19:13:23 rs: divs/spans are flexible, but don't have any semantics to convey to an asst tech 19:13:36 rs: no data w/a div 19:13:44 rs: should be able to allow the asst tech to map info to an api 19:14:23 rs: how do we fix the problem? start w/xhtml 1.0 and use namespaces to extend. 19:14:30 rs: what the html wg has done: added a role attribute 19:14:43 rs: 2. starting role/state taxonomy in the PFWG 19:14:56 rs: it will create a class hierachy of the objects we want t ouse. 19:15:00 rs: not new stuff, but using rdf 19:15:08 rs: using moz firefox for implementation. 19:15:28 rs: could add a plug-in to IE 19:15:42 rs: info is available through the DOM. UAAG requires access to the DOM. 19:16:06 rs: need at vendor support. have doug on the call who's been working with the team for window eyes to read firefox. 19:16:21 rs: important when talk about techs for wcag. need to create reusable widgets. 19:17:03 rs: tooling - it's not like running a checker on a web page, need to test the components w/asst tech. auth tool play important role. widgets need to be inc into the tools. 19:17:33 rs: gap analysis - will focus on the key items and how they map to accessibility api. 19:18:23 rs: much of this done by ua by default. major missing piece: role/state. also, only active elements were form elements and anchors. 19:19:40 rs: IE and tabindex - (WebForms group will implement) according to chart: 19:19:55 not present, <0, 0, >0....(see rich's slides) 19:20:24 rs: make use of role attribute to borrow role attribute from xhtml 2 (in xhtml 1) 19:20:31 rs: becomes a link to the class in rdf 19:20:50 rs: w/in each widget, can describe how to map. becomes self-documenting. 19:21:04 rs: like design patterns. by using rdf you build knowledge into the content. 19:21:32 rs: if you don't havfe that mapping, can still get the info. 19:21:39 rs: role can also be applied to other types of content, like svg. 19:21:56 rs: may want to create taxonomies w/in diagrams. 19:22:02 rs: becomes platform independent 19:22:12 pk: a site can define custom roles? 19:22:30 rs: in taxonomy will say "x is similar to y" and the asst tech will have some idea of what to do. 19:22:50 scribe: kerstin 19:23:11 rs: we're going to make sure we have the platform acc api 19:23:21 scribe: kesh 19:23:51 rs: rish shows an example in his slides of namespaces -- see slides (URI will be provided post-meeting). 19:24:29 rs: here is an example in flash, stock information, complex charting 19:24:54 rs: role taxonomy, rdf mapping, building in semantics about all of this 19:25:15 rs: it has states of active, visible, etc.., these are all acc states 19:25:37 rs: this is just an example -- it could be a parent/child .. 19:26:00 peter korn -- activity map looks like it's the container of all those dials .... 19:26:53 al: after the break, we get to ask questions and talk about issues -- let's just lay the background 19:27:20 rs: the other part is, skip-to-main-content issue: what we need to do is set metadata in the doc to find out where this info is 19:27:45 xhtml 2: where is the main content, where is the nav bar, banner, note, footer, etc -- standard rols attributes 19:28:01 rs: why is this important? 19:28:27 questions from the crowd .... 19:28:59 rs: I can use the same keyboard command to get to the same content on any webpage 19:29:08 rs: that's device independence 19:29:23 pk: how to map taxonomy to the button on the handheld? 19:29:47 rs: that would be agood thing to have -- the other thing that could happen is the browser could pick that up 19:30:08 agenda+ state as related to navigation landmarks 19:30:33 agenda+ is the contains relationship parent/child 19:30:50 agenda+ extensibility of role 19:31:07 rs: the next part is: states are not part of html 1.x 19:31:33 slide #15: 19:32:08 rs: you can set the focus progrmatically, but you don't want to be responsible for it 19:32:53 js: is this msaa of the future, too? 19:32:57 rs: yest. 19:33:17 khs: is this a wai sanctioned acc api? 19:33:28 al: it's not unique 19:34:32 khs: what would the impact be of not having css? 19:34:44 aL: the info is flowing the other way ... 19:35:06 rs: the other thing we are working on now, we need to be able to know what actions people want to perform 19:35:45 scribe = matt 19:36:45 rs: XForms. We're putting that into Firefox, expect it to be fully interoperable with Firefox. XForms has a standard set of events (validate, recalculate, etc.), so JavaScript is not needed to do that. 19:36:57 rs: I can expose that information to the AT, and now have real interaction with the doc. 19:37:10 scribe: mcmay 19:37:26 rs: Also a model for whether an item is selected, etc. UA can do the mapping automatically. 19:37:41 rs: Built-in relations for group, label, help, hint. 19:37:51 rs: Uses DOMActivate for device independence. 19:38:12 rs: Can come from voice recognition, for example, so don't need to be tied to keyboard. 19:38:35 rs: Can also understand semantics of e.g. date, so AT can read it properly to the user. 19:39:16 rs: XHTML 2 is based on declarative markup. Role attribute is built in. Navigation lists can be used with role. 19:39:42 rs: Current accesskey attribute is broken. It's device-dependent. Doesn't have a description attached. We don't know whether we're giving focus, activating, etc. 19:40:37 rs: We have made this an element within the document. Can put a title on it as description. For backward compatibility, you can attach a key to it. But this can be stripped out by the UA. 19:40:41 Kesh has joined #pfua 19:41:16 rs: We can specify navigation based on role. 19:41:45 rs: XForms is built-in without a namespace. You can tie the elements directly to the model. Also introduced nextFocus, and can use it to skip navigation. 19:42:24 q+ 19:42:26 agenda+ nextfocus. also need prevfocus? (slide 17) 19:42:56 rs: Have to put metadata into the content. Moving forward, have to make things easier for the page author, and that means building things like XForms into the technology. 19:43:18 rs: We can borrow from XHTML 2's features today, even if it's not quickly adopted. 19:43:27 pk: We can build this into open-source dev tools. 19:45:24 Next/Prev, reminds me of this: http://www.w3.org/TR/2004/NOTE-di-atdi-20040218/#navGenNextPrevious 19:45:51 taking a break for 5-10 minutes 19:52:41 Al has joined #pfua 19:54:41 we are back from break 19:56:25 JRG has joined #pfua 19:56:39 Lisa: Adaptive content and where adapative content is going 19:56:46 RichS has joined #pfua 19:57:14 One level I want to give an overview and give specific examples 19:57:22 ken has joined #pfua 19:57:37 Reason behind adaptive content, and want to provide everyone access to adaptive content 19:57:48 sh1mmer has joined #pfua 19:57:59 One problem is barriers to understanding and people don's want to change their practices 19:58:05 This most difficult 19:58:26 Getting content for everyone, including cultural, cognitive and disability factors 19:58:40 Using sematics in RDF to allow for adaptove content 19:58:51 !st we need to understand whats on a page 19:59:10 Dream of self describing content 19:59:56 At a later time someone provides an interpreter to provide information to a different audience 20:00:17 Ideally building that accessibility is represented by rich's presentation 20:00:24 Three step: 20:00:29 Capture content 20:00:35 Create knowledge model 20:00:40 Use the knowledge 20:01:08 At the moment to capture the knowledge is to use existing models, html is weak 20:01:27 Ultimately we want knowledge in the authoring tools 20:01:41 FYI: we will have a hard stop, brief break for WCAG Techniques to exit at 3:30. 20:02:09 Every time you add a layer to a graphic you could title it 20:02:20 Second stage is encoding the knowledge 20:02:54 The semantic web is the next generation of XML, which will add information about content 20:03:38 Software can treat the knowledge intelligently and can use the information more intelligently 20:03:46 Example: 20:04:09 See slide, interactivy model 20:04:39 You have state, actions and conditions that can effect each other 20:05:06 Two types of terms: predicates, symbolized by lines in the diagram 20:05:41 Subjects and documents are points in an xml document or rdf descriptions 20:06:02 Predicates define relationships between content 20:06:06 Content: 20:06:40 Data fields and relationships 20:07:01 Roles can be defined in these kinds of models 20:07:31 Web designers can define their own roles or a community like bloggers can define roles like bloggers 20:07:36 q? 20:07:54 agenda+ ease of which ppl can add new roles 20:08:11 Peter: My conern is with the ease people can add new roles, roles have no meanings if they cannot be presented intelligently to th euser 20:08:55 We have mechanisms to customize and create new intellifgence, but we do not have "kind of like" features 20:09:27 I am not saying that we can't innovate, but AT and mainstream people must cooperate 20:09:43 AG: We will discuss this, but not right this minute 20:10:11 Lisa: In the technology of encoding we need to address this concern 20:10:18 We need self describing 20:10:38 We need it to be extenable and that it can evolve as web technologies change 20:11:02 How do roles effect different users 20:11:35 How do we use the knowledge (self describing content) 20:11:41 One example is accesskeys 20:11:56 You have a pice of text that says take me to the site map 20:12:12 We have a piece of knowledge that the link goes to a site map 20:12:22 We can do tons of stuff with this 20:12:49 All users now can use Alt-S to go to the site 20:13:02 What about a russian, they might want Alt-K 20:13:23 The content author doesn't need to change anything, just a different interpretation 20:14:04 What about someone with a learning disability, I want an icon of the site map and I want it on the top right hand corner, this can be dome with XSL. 20:14:17 What burden am I putting on the author. 20:14:37 A lot less, since the author doesn't have to put in the individual content 20:14:49 They also may have content coming from several sources 20:15:07 They can have one line of information can now be turned into a role 20:16:03 What about abbreviation like WCAG, you can now have that linked to information about WCAG 20:16:38 Abbreviations can have knowledge to generate links if the user wants them 20:16:53 Slide on inaccessible content. 20:17:42 Lisa shows a document map that could be generated by using role information from the full document, creating a new interface slected by the user, not the author 20:18:15 You can create common interfaces that authors can exploit 20:18:27 Another use of RDF... 20:19:05 AG: Where is the dynamic content going 20:19:55 Rhys: The device independent working group is working how do we deliver interfaces to different types of technologies 20:20:07 We also need to know about content and the device 20:20:12 agenda+ DI to the second level 20:20:40 We need to take what the author wrote and mess with it and render it on a wide variety of devices 20:20:55 A major isssue is how do we make this easy to author 20:21:40 As we move more into dynamic events, including device changes, low battery, in sun light 20:22:04 We have received resources from multi-modal group to deal with some of this information 20:22:38 There are big problems, but xhtml2 and xforms are a big step forward 20:22:47 New techniques 20:23:36 I am excited about the new taxonomies being developed here to be integrated into markup languages 20:24:09 Joe: IE on windows have a good implementation of tabindex=0 20:24:26 Within the spirit of specs 20:24:44 But the specs say you cannot have neg values 20:24:56 Rich: You can violate specs for accessibility 20:25:02 AG: Can we do demo 20:25:41 Peter: One of the things we played with is formatting menus on dynamic braille displays 20:25:56 An exampel of using an abstraction to create custom user interface 20:26:00 for braill 20:26:20 Rhys: We have dome some of that too 20:26:36 Screen turned off for demo 20:26:44 Screen turned back on 20:27:07 Rich: 20:27:23 We tried to do hard things for our demo 20:27:31 What about a spreadsheet 20:27:54 We have stored the information as a menu and this is translated to MSAA 20:28:00 We gave the menu focus 20:28:17 It is a heirarchical menu system 20:28:38 There is state information on the state of the menu items 20:28:46 It is all written in div and span 20:29:21 It is accessible as long as there is a mapping between web page and acessibility API 20:29:31 We don't want to tab to everything 20:29:39 Now we are in a spreadsheet 20:29:54 Peter: Can you annouce what key you are using 20:30:13 Rich: Moving between cells with arrow keys 20:30:23 It reads the cells 20:30:34 Johns: Can you read headers 20:30:42 Rich; We are working on their 20:30:55 We can also attech headers to style information 20:31:15 Joe: Some people would use lists, instead of ul and li 20:31:28 Joe: Technically can't have avalid page 20:31:37 Ric: I don't care 20:31:44 Joe: I assume it will still validate 20:32:07 Rich: DOM and HTML are not going to go back and change things 20:32:36 Rich: Opera and Firefox are going to support tabindex=0 20:33:07 Joe: headers can dfined using html table header markup 20:33:10 Rich: yes 20:33:24 Editing a cell 20:33:38 We can read the label, enter a new value 20:33:43 technically the validator probably won't flunk it, I meant. 20:34:00 Changes a coupld of cells 20:34:27 PeterK: What is the parent 20:34:54 Rich: We have an edit field inside a label, that is inside of a td 20:35:16 Rich: this is complex spreadsheet, it is a real menu 20:35:24 I think this is much more usable 20:35:58 There is no form element 20:36:23 Thank you doug for working with us on this 20:36:46 verified: valid file with tabindex="-1" on element passes the validator. 20:36:50 Demo of windows magnifer with spreadsheet 20:36:54 on request, I can post the file. 20:36:57 magnifer follows focus 20:37:45 Peterk: This all MSAA, right? 20:37:48 Rich: Yes 20:38:08 ben has left #pfua 20:38:36 joeclark has left #pfua 20:39:41 Break, as WCAG people leave 20:40:38 Rotan has left #pfua 20:50:17 joeclark has joined #pfua 20:50:42 Rotan has joined #pfua 20:53:27 Michael has joined #pfua 20:53:35 Michael has left #pfua 20:55:34 joeclark has left #pfua 20:56:00 Gottfried has joined #pfua 20:58:36 agenda? 21:08:24 RichS has joined #pfua 21:08:59 agenda+ tracking and comprehension of complex mutation events 21:13:19 agenda+ navigation mode models ; consistency problems with tabindex and accesskey 21:15:45 agenda+ targeting real-world vs. ideal-world legacy-how-friendly 21:17:34 patrick has joined #pfua 21:18:13 q+ aaron 21:22:38 ack aaron 21:25:50 louie has joined #pfua 21:33:43 RhysL has joined #pfua 21:34:44 agenda+ how to make it normative [from Gottfried] 21:38:24 q+ Kerstin 21:38:43 ack joeclark 21:44:19 agenda? 22:23:11 mcmay has left #pfua 22:29:13 meeting recessed -- PF and UA to reconvene separately on Thursday. 22:30:21 dpoehlman has left #pfua 22:32:20 Al has left #pfua 22:33:51 Rotan has left #pfua 23:06:24 Al has joined #pfua 23:06:47 Zakim, who is on the phone? 23:06:47 sorry, Al, I don't know what conference this is 23:06:48 On IRC I see Al, patrick, RichS, Gottfried, janina, Zakim, RRSAgent, LS 23:07:27 Al has left #pfua 23:10:53 jessie has joined #pfua 23:27:10 RichS has left #pfua 23:27:59 jessie has left #pfua