- From: Charles McCathieNevile <charles@w3.org>
- Date: Thu, 13 Sep 2001 19:51:48 -0400 (EDT)
- To: WAI AU Guidelines <w3c-wai-au@w3.org>
JM == Jeff Murray, Testing accessibility with XP (left at the break)
GB == Girogio Brajnik
HS == Heather Swayne
CMN == Charles McCathieNevile == <chaals>
WC == Wendy Chisholm
KHS == Katie Haritos-Shea
Phone:
JT == Jutta Treviranus
JR == Jan Richards
PJ == Phill Jenkins
IRC
<oedipus> == Gregory Rosmaita
<liddy> == Liddy Nevile
log:
*** wendy (~wendy@tux.w3.org) joined #wai
<chaals> slide: quality bar
<chaals> KHS why are accessibility and usability seperated?
<chaals> JM If we are doing bug triage we can let something through that
isn't easily usable, but things must be accessible.
<chaals> .. it is just a case of which things we work on first.
<chaals> CMN similar between level-A and triple-A, where the closer you get
to triple-A the more you are working on whether things are usable, since the
are they available problems are solved earlier in the process.
<chaals> example: a colour picker that is a table, but is presented to
assistive technology as a list.
<chaals> So user isn't immediately aware that they can get more then the
first column.
<chaals> CMN Would look at a bunch of differnt requirements - documentation,
access by keyboard, presentation of information, etc...
<chaals> Probably fails a requirement, don't know off-hand if this is P1 or
P2.
<chaals> Really User Agent Guidelines issue.
<chaals> JM Web piece is rather small.
<wendy> HS: sharepoint - manages and creates pges.
*** Channel doesn't support modes
*** chaals has set the topic on #wai to AU meetin
<chaals> KHS are these things acessible via Web?
<chaals> HS can be - depends on setup for individual use case
<chaals> ** JT joins
chaals thinks bobby is the third brother, after greg and the
middle one.
<wendy> CMN what are the other add-ins to frontpage?
<wendy> HS all office products have that architecture.
<wendy> CMN take the accessibility add-in, put it into excel, put w/excel,
output html, run the checker, open in frontpage and do x, y, z?
<wendy> HS would have to change architecture of apps.
<wendy> s/apps/plug-ins
<wendy> .. specialized based on unique functionality.
<wendy> xx at basic level, everything is similar, but each so specialized.
<chaals> xx == JM
<wendy> CMN (resondping to issues w/checkpoint 6.3 - dev time and design
issues)
<wendy> HS Many dev teams turning off applets - dev argues it is the tool.
<wendy> .. how could i give a text representation?
<wendy> CMN when talking yesterday re: visual studio, put function on
server.
<wendy> .. e.g. form validation.
<wendy> .. issue with replicating functionality in diff languages. e.g.
perl on server (cgi) and javascript on client.
<chaals> JM Testing tools should be certified byWAI so people know they are
exposing things that are important
<wendy> CMN Up until now we have not done certification of anything.
self-claims that are unverified.
<wendy> .. w3c has started a QA activity. Lots of people making this
request.
<wendy> .. WC's group looking at examples to run tools over.
<wendy> .. WG hasn't gone through them yet.
<wendy> .. test suite that you can be pretty sure that WAI agrees with.
<chaals> JM our tool is getting into final stage
<chaals> CMN It could be good to run your tool over the test suite and get
results, so when the WG reaches decision
<chaals> ... you can compare your results
<chaals> JM sure
<chaals> action WC: give JM pointer to test files.
<chaals> JM would you be intersted in test tool results, how would you use
them?
<chaals> WC interesting to compare tools.
<chaals> .. checking of test files
<chaals> CMN Au uses evaluation results fo:
*** KatieHS (~kshea@192.239.92.63) joined #wai
<wendy> .. proving implementability.
<wendy> .. new part of W3C process.
<wendy> .. testing the spec and devleopment.
<wendy> .. being able to answer "what tool can I use?"
<wendy> .. we are trying to give people a clue about how to evaluate their
tool against ATAG.
<chaals> JM how we test for accessibility:
<chaals> There are 2 main contacts across office.
<chaals> have a represenetative from each application group.
<chaals> they might farm it further out to team members, but they are
responsible for reporting back - where are we up to.... etc.
<chaals> do general testing with the whole team, and also go down to detailed
tests of particular features.
*** Katie (~kshea@dialup-209.244.110.85.Dial1.Seattle1.Level3.net) joined
#wai
*** Katie has left #wai (Katie)
*** KatieHS has signed off (Ping timeout)
<chaals> HS tester's role in spec-writing process
*** KatieHS (~kshea@dialup-209.244.110.85.Dial1.Seattle1.Level3.net) joined
#wai
*** Phill (~pjenkins@pixpat.austin.ibm.com) joined #wai
<Phill> Phill is back, will join phone in a minute or two
<wendy> hi phill.
<wendy> jm when get the spec, we go through each aspect and ask "how would
we test this?"
<wendy> .. if not testable, then can't do.
<wendy> HS summary, user scenario, benefits, etc.
<wendy> GB How validate a spec? it's the right solution to the problem.
<wendy> HS Prototyping part of spec writing process.
<wendy> .. usability studies on them, incorporate feedback into the spec.
<wendy> GB Always do user testing?
<wendy> HS Not always.
<wendy> .. if can prove w/out usability testing, then may go forward. if
get pushback, then have to get the data.
<wendy> JM It has proved very valuable.
<wendy> .. aiming for 90%.
<wendy> .. user scenarios. "who the user, what's the problem, what are we
trying to solve."
<wendy> .. must support that user experience.
<wendy> HS Must be able to articulate the user experience.
<wendy> GB For a certain feature, what is the # of testers vs devs.
<wendy> JM We start looking at the spec, devsto testers, specsto testers.
<wendy> .. try to do 1 to 1.
<wendy> .. our best testers can keep 4 devs busy.
<wendy> .. depends on the feature.
<wendy> .. sometime implementing pen support or spell checker.
<wendy> .. looks like a huge feature, but taking code from another
feature.
<wendy> .. all we have to test in the integration.
<chaals> WC in WCAG 2 we are looking at testability - how can we improve that
process?
<chaals> Talk about use of scenarios and personas for spec design.
<chaals> stuff that is important for testing in office.net
<chaals> tables, applets, frames, shortcut keys, text alternatives for rich
media
<chaals> MSAA 1.3 / 2.0
<chaals> 508 compliance
<chaals> WAI single A compliance
<chaals> localization
<chaals> How to decide when to drop support for legacy situations...
<chaals> ... it's just hard.
<chaals> JT Persona - do you have personae who have disabilities?
<chaals> HS we get a request for a "disabled" persona to cover accessibility.
<chaals> .. there has been a successful one where there are a bunch of
personae, and accidents happen to them so they acquire a disability...
<chaals> JT IT would be intersting to get people to have these, rather than
just a 508 checklist.
<chaals> HS Yep. This s an ongoing discussion- how to do it effectively
without having people end up targetting one or two specific cases...
<chaals> JT do you use cases of how this persona would use the prpduct?
<chaals> JM we do that in user testing - put someone in a number of
situations and see what happens.
<chaals> .. it is not the priary method, because it is very inefficient, but
it is useful.
chaals wonders if phill is on the phone
<chaals> back to slides...
<chaals> ..Localisation.
<chaals> JM there isn't a lot of demand for accessibiltiy internationally.
<chaals> .. We expect there will be.
<chaals> ..we are geetting localisation under control, and now getting lots
of stuff put out in multiple languages
<chaals> ..when this is on the web, for servicees, etc., you need to be able
to customise it.
<chaals> ..how does that happen?
<chaals> JM there are no variables on accessibility in differnt languages
<wendy> CMN There are some issues.
<wendy> .. language usage. rules change.
<wendy> HS ATs around the world rely on MSAA just like here.
<wendy> JM Not our job to identify bugs in Jaws or other ATs.
<wendy> CMN You have a well-defined delivery platform, focusin on IE.
<wendy> .. MS could decide, this part of the work will be done in browser,
this part in content.
<wendy> .. in WAI content, might have IE, or Opera or Word or FrontPage or
DreamWeaver or ...
<wendy> .. there is overlap. until everyone gets it right, that overlap
will be bigger.
<wendy> .. sometimes the browser tools will solve, sometimes the authoring
tools.
<wendy> .. not a priori someone's responaibility.
<wendy> .. no real reason to say it must be this end or that one.
<wendy> .. not sure we resolve that.
<wendy> .. other than getting implementation further along.
<wendy> .. a general principle is the impact on requirement.
<wendy> .. requiring a UA to do something - they are specialists.
<wendy> .. authors are much more varied.
<wendy> .. tendency to desire on the UA side.
<wendy> .. need to also ask content developers, which really means
authoring tools.
<wendy> JT Do you take in user input or AT dev input re: bugs or problems.
where does that come in? eval team?
<wendy> HS Everywhere, all the time. Try to get AT dev involved earlly in
the process.
<wendy> .. get feedback from community groups. we coordinate that for
product groups.
<wendy> JT it goes into the next release?
<Phill> Is that Heather from MS talking?
<wendy> HS It depends on where we are in the product dev cycle.
<wendy> yep.
<wendy> HS and how serious the problem is.
<chaals> JM would recommend really strong requirement for documentation of
applications - it is really critical.
<wendy> JT I've noticed that the access pieces seem to be in windows
products faster than mac.
<wendy> HS Refering to office?
<wendy> JT Yet.
<wendy> HS mac and office product devs are separate.
<wendy> s/yet/yes
<wendy> HS PC team is larger, more market research, user scenarios, etc.
<wendy> .. tend to dev more features.
<wendy> .. mac then looks at. irrelevant of accessibility - features in
general.
<wendy> .. how they approach accessibility is diff.
<wendy> ** HS explains.
<wendy> .. native accessibility in each environment is very different.
<wendy> JM tablet PC is a new realm for us.
<wendy> .. there are intrinsic issues w/tablets for accessibility.
<wendy> JT Onscreen keyboard?
<wendy> JM Yes.
<chaals> HS Accessibility stuff has been re-organised on web.
<chaals> ======break
<chaals> phill, are you there?
<chaals> we are back online (and so is Jan)
*** Liddy (~Liddy@128.250.190.59) joined #wai
<chaals> agenda for rest of afternoon:
<chaals> Publishing evaluations, if Phill turns up.
<chaals> Where our documents are at and where they go now.
<chaals> Interaction with WCAG - where are they at?
<chaals> Next meeting / Plenary planning
<chaals> 1. where are we at:
<chaals> Wombat: ready for first public draft
<chaals> needs director's approval.
<Phill> Phill here - this chat doesn't doesn't wake me up like my internal
IBM version does...
<chaals> we don't havea requirements doc, but I think we wanted to clean up
a few bugs, get it ready for WCAG 2.
<chaals> JR pretty much it.
<chaals> HS Last wombat draft didn't map to checkpoints
<chaals> CMN we hadn't resolved that yet.
*** KatieHS has signed off (Ping timeout)
<chaals> .. it is resolved that we will identify those specifically for 3.1
and 3.2
<chaals> HS OK. I will probably look at this and send proposal to list so
that there is better wording match.
<chaals> PJ are we with WCAG 1.0 or 2.0?
<chaals> CMN we are with WCAG 1.0 until WCAG 2 gets to last call.
<chaals> CMN Techniques document. I will put out a new draft, so we can get
it to update of Note.
<chaals> .. then would like to map the ATAG 1.0 techniques to Wombat
chekpoints
<chaals> HS Why do that now and not later?
<chaals> CMN Think it makes it easier to read Wombat with techniques.
<chaals> PJ agree
<chaals> JR It would be good if WCAG 2.0 is years away to publish ATAG 1.1 as
something where we could replace WCAG 1 with WCAG 2 easily
<chaals> CMN yes, that is our development plan.
<chaals> PJ also wanted to talk about testing techniques for softweare
accessibility
*** KatieHS (~kshea@dialup-64.154.185.76.Dial1.Seattle1.Level3.net) joined
#wai
<chaals> **JT joins
<chaals> PJ Evaluation techniques. Seems like User Agent, us, and some other
groups have published doscuments on how to test compliance. Is there a
consistent process on how to do that yet?
<chaals> i.e. what does W3C recommedn for me to say that my browser conforms
to HTML 4
<chaals> I think that is something we have started to do.
<chaals> CMN there is no recommended process.
<chaals> JT We had discussed this about ayear ago - we had a draft protocol
for notifying a developer that an evaluation had beeen made, and giving them
a bit of time to respond. We never formalised that.
<chaals> PJ I thought the evaluation document was as a developer how I test
against ATAG
<chaals> CMN It isn't necessarily just a developer
<chaals> PJ Wjhy isn't that part of ATAG itself? In IBM stuff we have testing
techniques in our specifications.
<chaals> We have checkpoints in one document, techniques for meeting in a
second, and evaluation techniques in a third document.
<chaals> JT There is no one way of meeting checkpoints. When we fill this out
we wil have huge document.
<chaals> JT noone would read through the entire techiques doc still. It is
like a reference document so you can find specfic stuff.
<chaals> likewise, if you have to assess a tool you want a step by step
mannual - there are different audiences.
<chaals> PJ You expect people doing evaluations to be the audience
<chaals> JT Evaluation and implementation techniques are a bit intertwined.
You wouldn't want to not look at evaluation techniques.
<chaals> Q = HS, CMN
*** Liddy has signed off (EOF From client)
<chaals> CMN it is good to have testing techniques that aren't the same as
implementation techniques
<chaals> PJ I did evalaution techniques against 508 that were 20 pages.
<chaals> .. will send that to the list.PJ it is overwhelmeing to the tool
developers how much stuff is in there.
<chaals> JT so one of the big things is to structure it so there are
meaningful views of the stuff that is trelevant for a given user.
<chaals> e.g. a video tool developer only wants to have to read the stuff
that is relevant to video tools
<chaals> PJ OK.
<chaals> Action PJ, send IBM software testing techniques to list.
<chaals> ==
<chaals> evaluation techniques. Do we need to compare our method for doing
evaluations with how User Agent do it?
<chaals> Education and Outreach wants to publish gallery of sites - look at
the process to evalaute a site against WCAG.
<chaals> Maybe we should review each other's documentation.
<chaals> what if someone doesn't want a document published?
<chaals> JT reason we didn't pursue this further is that QA group began and
we think that something will be done across groups.
<chaals> PJ They seem to be focussed on testability off groups - scope seems
a little less broad than we thought it would be.
<chaals> JT Yes. We should review our draft protocol...
<chaals> CMN Yes to PJ
<chaals> CMN It is useful to look at the EO group.
<chaals> ..It is good to lok at UA evaluation method.
<chaals> ..And we should look at test suite stuff coming from ER
<chaals> JT yes, we should look at protocol of testing, asking developers to
look, etc.
<chaals> PJ and how do we deal with old evaluations.
<chaals> JT What do we do about tests taht the tool users
<chaals> PJ How we evaluate a tool is in the evaluation doc, the tests are in
techniques.
<chaals> we should review the protocol for testing?
*** oedipus (~oedipus@tc1dyc37.hicom.net) joined #wai
<chaals> JT Yes, should be reviewed.
<chaals> CMN review of protocol - on the one hand we need to be helpful to
developers and on the other hand we want to have a light enough process that
people talk to us and give us evlauaiotns
*** Liddy (~Liddy@128.250.190.59) joined #wai
<chaals> JT there is also the question of a database of evaluations for
people to read
<chaals> JR Is that our role?
<chaals> PJ the evalauations are useeful input.
<chaals> JT When we were talking about this is meant taking classification
and markup.
<chaals> CMN we can do this in EARL now and there are tools taht we could be
using.
<chaals> But I think that it is low on our priority list to make tool
comparisons for purchasing.
<chaals> PJ I can find comments I had about protocol and start thread on list
- should I?
<chaals> Action PJ Make it so
<chaals> JR What is the concern about our process?
<chaals> PJ I think ours is good, but I need to check what I was saying.
*** Katie (~kshea@192.239.92.63) joined #wai
<chaals> JR At ATRC we have a bit odf a situation where we have done a few of
these and are debating whether or not to post them, considering that sharing
the information with the companies has spurred them onto development.
<chaals> PJ To keep them working, hide their dirty laundry?
<chaals> JT strategy for best results...
<chaals> JR These tools cn be changed on the fly and outdate an evalaution
overnight.
<chaals> e.g. a tool gets documentation updated on the fly.
*** KatieHS has signed off (Ping timeout)
<chaals> CMN They should update their versions
<chaals> JT one thought is to give them a chance to update and fix their
things
<chaals> CMN Right. It is hard to make formal rules about when to hold off
and get a better thing made...
<chaals> PJ If the developer says "please don't publish until we have fixed
this" do we wait? How long?
<chaals> we should as a group try to get some general conseensus, but should
still publish evaluations so we know where the technology is at.
<chaals> JT we do have every intention of sending it to the list following
this next round.
<chaals> ===
<chaals> where is WCAG at?
<chaals> CMN We have a critical dependency on WCAG and they are in process of
developing WCAG 2.0
<chaals> Wendy.. tell us the answers...
<chaals> WC Wendy lives in Seattle.
<chaals> Action PJ, remember that Wendy lives in Seattle...
<chaals> WC New draft published. We have success criteria, benefits,
definitions, and examples.
<chaals> 4 guidelines, 21 checkpoints.
<chaals> We have 4 techniques documents in progress: HTML, CSS, ECMAscript,
PDF
<chaals> There are discussions for others.
<chaals> We have some major issues that we have been going around on the list
<chaals> (WCAG had face to face monday, Tuesday)
<chaals> been trying to get rid of the "elephants" - major issues.
<chaals> PJ Was that successful?
<chaals> WC Yep\
<chaals> KHS e.g. require techniques and success criteria for something to be
a checkpoint.
<chaals> Notes from WCAG have been sent to mailing list.
*** KatieHS (~kshea@dialup-209.245.163.8.Dial1.Seattle1.Level3.net) joined
#wai
<wendy> here is the address of Gregg's message - summary of consensus on
elephants.
*** Katie has signed off (Ping timeout)
<wendy>
http://lists.w3.org/Archives/Public/w3c-wai-gl/2001JulSep/0960.html
<chaals> CMN I was at meeting and they are going forwards again.
<chaals> I would be impresed ifg they got to last call in 6 months
<chaals> My wild guess is they are a year from last call.
<chaals> ======
<chaals> next meeting / plenary
<chaals> People who sai they would be at plenary:
<chaals> CMN, MK, GB, GR, KHS, WC
<chaals> Won't be there: WL, HS
<chaals> PJ planning to go there
<chaals> liddy said she might. Liddy?
<chaals> South of FFrance, 25 Feb - March 1
<chaals> like the all groups meeting we had in boston last feb.
<chaals> (but warmer)
<chaals> JT depends on who we want to meet with.
<chaals> JT how much do we get done in face to face vs teleconfernce / list
<chaals> JR last plenary wasn't most productive, but amsterdam we got a lot
done.
<chaals> CMN If we are going to lose people to another meeting it isn't
worhtwhile.
<chaals> JR end of the week gets to burnout.
<chaals> JT aim for beginning, no overlap with UA/WWCAG then I think it is
worhtwhile.
<Liddy> I am here but 'just' !
<chaals> Action CMN - talk to Daniel Dardailler. We want early in the week,
no overlap with other folks.
<Liddy> I will try to get to the F2F, yes
<chaals> WC WCAG is another option, but loads CSUN.
<chaals> Don't think ERT will be at plenary, and think it would be good for a
group to be there.
<chaals> JT There is still the invite in Vancouver
<chaals> HS who is that offering?
<chaals> JT WebCT
<chaals> liddy, are you available for a face to face meetin gin Nice feb
25-6?
<Liddy> I expect so - yes
<chaals> resolved: We think we will go to plenary (except we hold the right
to withdraw...)
<chaals> CMN Do we want a face to face between now and then.
<KatieHS> Liddy, Katie here, what was it that I volunteered to test on Word
XP and Tidy and/or xhtml this am?
<chaals> JT keep Vancouver option open instad of plenary or after plenary.
<chaals> CMN right.
<chaals> Anyone want to meet in Oz?
<chaals> possibility: Hawaii for Web 2002
<chaals> HS is getting married that week.
<chaals> possiblity - not a lot of support for it?
<chaals> next teleconference: Monday?
chaals is worried about still travelling
<chaals> Monday 24?
<Liddy> I think we were going to see if we could make a word doc, save it
as html, clean it up with HTMLTidy, read it back into Word, and still have
XHTML..
<chaals> Monday 24 is next call.
<KatieHS> k thanks
*** oedipus has signed off (Ping timeout)
*** wendy has signed off (...sunny days, sweeping the clouds away...)
<chaals> thanks to Heather for organising and reorganising this.
<Liddy> thanks Heather for the coffee and donuts...
<chaals> Thanks to Microsoft for hosting.
<chaals> Talk to you all on Monday 24.
<Liddy> oh, toast and coffee .... :-)
<chaals> Bye......
<chaals> Mmm, and vegemite.
<chaals> Bye folks. We're going to be exhausted.
<Liddy> I had p-nut B - for Ollie ...
--
Charles McCathieNevile http://www.w3.org/People/Charles phone: +61 409 134 136
W3C Web Accessibility Initiative http://www.w3.org/WAI fax: +1 617 258 5999
Location: 21 Mitchell street FOOTSCRAY Vic 3011, Australia
(or W3C INRIA, Route des Lucioles, BP 93, 06902 Sophia Antipolis Cedex, France)
Received on Thursday, 13 September 2001 19:51:48 UTC