Chair: Jutta Treviranus
Date: Tuesday 9 January 2001
Time: 2:30pm - 4:00pm Boston time (1930Z - 2100Z)
Phone number: Tobin Bridge, +1 (617) 252 7000
The Latest Draft is the Recommendation dated 3 February, available at http://www.w3.org/TR/2000/REC-ATAG10-20000203. The latest techniques draft is dated 18 September, available at http://www.w3.org/WAI/AU/WD-ATAG10-TECHS-20000918. The newest versions of techniques information are available from the Sources area: http://www.w3.org/WAI/AU/sources
JT Hopefully people have looked over this. There are outstanding issues inherited. Jan has put a draft out. That is linked from teh sources page.
CMN It is actually the previous version that was linked - I will update today.
JR It is me inserting AERT under 4.1. The issues are the way that reformatting is done, and the issues that were already outstanding.
CMN I have not had time to give a thorough review of the AERT bit.
JR It is an enormous multi-level list of lists of lists
LK It makes the document awfully long. Could it be formatted like a branch?
MMP or an appendix?
GR We could provide a single HTML file and then an archive file where it is chopped up.
CMN We have had an ongoing work item to get different views of the techniques availble. That is finally close to done, (the information is in the sources directory) and we can use that technique to easily add or remove things from particular views.
JT Is there a way to compress the document so we can get a simplified version of the techniques.
WL Like a table of contents beyond the index level.
JR For AERT or all?
JT For all techniques.
JR Isn't it the checkpoint text.
WC For WCAG there is a gateway techniques document. Under each checkpoint it has an index to a heading for longer more complex techniques.
CMN We are in a position to do that - we just need to decide what we want and set it up. But it getes easier.
JT So do we want to do that?
CMN I think we do.
WL Are you talking about putting an ID on each thing?
JT Something that gives enough information to know what is in the technique.
CMN Having Techniques that are fuller, including a heading and then more substantial amounts of content per checkpoint.
WL Who is the audience?
JT Comment on integration is that we want to get differing views.
CMN We want to identify the things incorporated wich are actually WCAG techniques that we can then give to them, and re-incorporate as we generate, as WCAG techniques.
LK I am wondering if the raw document can be used as a database of techniques that can be pointed into from teh description language.
CMN The answer I think is yes. It is set up to have names for everything.
LK Do we point to the source or the published version
WC I don't think so. Published version is better.
CMN The namse stay stable so you can point to a published version.
/* WC gets droppped off
CR There are maybe a dozen outstanding issues - some we may not be able to but I think we can resolve some
JT Are they WCAG or Authoring Tool Questions?
CR Both
JT How do we deal with them?
CMN Give WCAG issues tothem and just take what they give us
WC Part of them are areas where we don't have algorithms. There are a lot of them, and there is an existing list. Some of them in fact may be resolved. And there are things like the triggers for ADL.
/* SL leaves
Action WC: Check that the AERT issues list is up-to-date
Action CMN: Take the updated issues list and turn it into single agenda items (for appropriate groups).
WC Can we put some of these on the agenda for teh joint face to face?
LK How many of these are choices between a couple of techniques, how many are things we don't know how to do?
WC I think most of them are things we don't know how to do - for example how do we identify and deal with the use of alternative pages?
CMN First thing is please register - it helps for planning and the plenary session, and the Hotel rate expires.
CMN How many people are going to be there for more than the two days?
LK, JT, GR, CMN, BM, WC
LM Microsoft are not sure - at the moment we are reorganising.
JT One thought for the plenary is to determine whether there is a need to add to or expand ATAG. One thing might be to look at techniques and guidelines from teh perspective of non-HTML technologies to be prepared to talk to other groups.
CMN For the purposes of the meetings that are member-confidential, people in these groups who are not representatives of W3C members will have to sign a confidentiality agreement.
HB Have people requested phone-in?
CMN No
HB It would be good to have Sean Palmer available
CMN Do we want to meet some of our time with WCAG?
JT Can we do it for part of a day to work on AERT WCAG issues?
CMN Yep. There is also a preliminary discussion on the glossary scheduled for Tuesday - at this stage it was for the people who are going to edit it. We could make it a joint working group discussion.
HB I think that it should be up to each guidelines editor to justify why they are diverging
JT So we want the editors to bring issues back.
HB Yes.
JT So we want an editors' meeting on that.
CMN We may also want a joint meeting - we have t odo joint work on unifying the glossary at some stage. Another proposal is to have a joint list across working groups. How do people feel about that?
JR Sounds good
MK I think it is very important to be able to do that.
HB In that glossary there were some sources that look as if they are PDF files
WC No, they are definitions out of the techniques for PDF document.
CMN There are people in the WCAG group writing techniques for using PDF in an accessible way. The definitions in those techniques are being incorporated into the grand unified glossary.
CMN Did we want to meet with WCAG for some time, to resolve AERT issues?
JT I think we do - how much time? Can we ask them to our joint session, or the Friday session.
WC I think it would be better for WCAG to join thursday.
JT End of the day?
CMN How many AERT issues are WCAG?
WC 11
JT Two hours?
CMN At least.
JT I don't know if we will resolve them then, but we might get enough interest that they will then go resolve them.
Action WC: Invite WCAG to join us at the plenary on Thursday for between two hours and the afternoon.
HB I am concerned that UA will be busy working and its members will be in other meetings
CMN True. This is something unavoidable. But please take advantage of the other time available to talk to people.
LK Is everyone familiar with what this is? (Silence). It is a formal why to describe what it and isn' accessible in content. The question is how this relates to AU. Should there be ATAG requirements for this? How is it most useful to tools? What sort of interest is there among developers?
WL The language is currently vapourware, right?
LK Right. There are descriptions (CMN and partial implementations of proofs-of-concept)
JT The idea is that different tools can be used for evaluation and the repairs?
LK Right.
JT What kind of form are we looking at?
LK At this stage it looks like being RDF.
JT From an authoring tool perspective, it allows the tool to use a third-party evaluation and then do the repair, or allow the author to choose their own tools for evaluating, or for repairing, ...
LK There is also sending it to another person for evaluation, and getting back the results in this language for the tool to use. If AU adopted this, how would it fit in? Guideline? Technique?
JT There are a few steps before putting it in as a guideline or checkpoint. Not least we need to create it...
LK There are things taht the authoring tool can do to make it apply - generate content that can be more easily assessed. For example keeping IDs stable among elements makes it easier to generate information that works if the document is edited.
JT The most useful case seems to be in aggregate systems - that include authoring, client, serving, sort of mixed together.
LK Can you say more about WebCT?
JT It is an example - there are others. But rather than being an authoring tool it is a courseware environment. It isn't clear if content comes from the tool, or if it is crfeated or edited in other tools. There are guidelines, templates, assistants, to help you construct the courseware. At the end you have the student interface and the authoring interface, and in some cases the student is creating further content. In a hybrid situation like that this would be really useful.
CMN Lotus Notes is the same kind of system (but not pitched at education).
LK This would be a case of content management systems
JT It is a bit more - making authoring and collaborative work and browsing all part of the same mix. And may be several people using several tools to work on a changing document.
LK So you might start with a "normal" authoring tool and browser. Then you have conversion from different authoring software.
JT There is also the evaluation and repair utilities. There may be a better table fixer that you could send a file too, and then take it back cleaned up.
LK Web Content Management Systems can have workflow built into them. Person A creates content, person C edits, Person L approves and publishes. If someone rejects a document then this would give them a way of saying why they rejected it.
CMN Or of assessing whether they want to reject it or not.
JT We don't know how easy the language will be to use
WC So long as we do it right, you could provide multiple presentations of the things identified, according to different audience needs.
CMN How much interest is there in this kind of thing, among developers?
WC It would be useful to have expressions of interest. We can come up with all this, but it is a question of whether or not it will be used
JT Are InSight intersted?
LK SSB (Tim Springer) has expresed strong interest.
JT They have a schema that does something like this, right?
LK At the moment they have something that provides a list of errors - kind of raw material. Is there any work going on of which this would be a special case?
CMN I would look at the RDF interest group, and the QA activity who are working on test suites.
LK Are there people on the call who are part of authoring tool developers and have a reaction?
MK I am a product manager. I have to think about what to do with it really...
LM To be honest I couldn't hazard a guess. I would have to talk to them.
CMN Do you think it is worth taking your time to go talk to them?
LM I am not sure. I need to put it on the table and find out.
MMP Not much more.
LM I am technically a point of contact - I am not part of the group on a daily basis.
LK If people are doing workflow management in their systems, the language that has things like "you meet this checkpoint" can be generalised to meet any kind of checkpoints - as well as WCAG you could use standards for making sure that a product conforms to other kinds of checkpoints. Alternatively there might be stuff out there doing this and we could just adopt that.
JT I think some of those are at a higher level - this document follows the template.
WC We have talked about some of those things - sometimes we want to address things element by element, or summarised information on a page or site level. We got really abstract about making claims about Web Content, but can also talk about functions of an authoring tool and where they go.
JT What you are talking about is more something of a descriptive tool that gives information about a document rather than something that any tool canmake use of
LK No, it is the latter.
CMN One of the reasons that Authoring Tool assessments came up is that Dan Brickley and I had done some work on that already, and it turned out to be usful for SCAG assessment as well. It also was fairly simple - it isn't making it work that is hard - it is agreeing on the details that we want.
LK Is there an example implementation that would be useful?
LM Give me an example of an example
CMN There are a number of teools that make assessments about Web pages. There are differnt tools, that can be used to edit those pages. So imagine if the Wave used this language to produce its report about which images had bad alt text in a page. Then the page, and the evaluation information is opened in an authoring tool, and it asks the author to fix only thse images which have been identified as having bad alt text, since it has information that the other images have good alt text. Example 2. A page has been rated and claimed as meeting everything for WCAG double-A. The page is edited in tool X, and two images are changed. The tool knows what has been changed, so it relies on the fact that manual checking has already been done for the rest of the page, and doesn't need to worry about asking for all that to be done.
LK e.g. for Amaya, you arrange for Amaya to mark any image that is identified as having bad alt text, and asking the user to fix it.
WL Something that sounds good is a 508 checker.
CMN 508 is a big use case in the US. WCAG is a big use case in Europe. RDF allows us to have both, and say which things are the same - this is handy
LM There is a concern about making a claim that may or may not be valid.
LK You can use it to say what tool did the testing.
BM If you figure out how to auto-test the last 508 requirement, let us know.
CMN So manually test it, and record the result of the test in the Evaluation and Repair Language...
LK Need to wrap.
CMN This also crosses into Semantic Web Activity, and Quality Assurance - there are people interested there, and I am also interested in it as something for those areas as well.
Next meetings at usual times for each group.
Last Modified $Date: 2001/01/10 01:17:40 $