W3C home > Mailing lists > Public > w3c-wai-gl@w3.org > July to September 2001

Re: f2f session is over (8:50pm boston time/5:50pm local time [US Pacific])

From: Aaron Swartz <aswartz@upclink.com>
Date: Mon, 10 Sep 2001 20:02:49 -0500
Message-Id: <200109110105.f8B15qb01460@theinfo.org>
Cc: "Charles McCathieNevile" <charles@w3.org>, "Wendy A. Chisholm" <wendy@w3.org>
To: w3c-wai-gl@w3.org
On Monday, September 10, 2001, at 07:51  PM, gregory j. rosmaita wrote:

> aloha, aaron - could you please send your IRC log of #wai directly to
> w3c-wai-gl@w3.org?

It's included below. Hope this helps,
- [ "Aaron Swartz" ; <mailto:me@aaronsw.com> ; 
<http://www.aaronsw.com/> ]

Log file opened at: 2001-09-10 3:40:14 PM
[3:40] *** Mode change "+i" for AaronSw by AaronSw
[3:40] *** You have joined the channel
[3:48] *** TimLa has quit IRC ()
[3:48] *** TimLa (~timlaranc@ has joined the channel
[3:50] <Jo> GV why is the assignment database not part of the 
project store?
[3:51] <Jo> typically the manager wants to be in control over 
what's in project store. Assignment area is temp database until 
things get approved by mgr to go into project store.
[3:53] <Jo> Is Exchange required?
[3:53] <Jo> Not any more. Web-based.
[3:54] <Jo> (above question was from RN)
[3:57] <Jo> WC On screen shots, which are ActiveX controls?
[3:58] <Jo> (shows) The benefits of the ActiveX controls: much 
faster performance.
[3:59] <Jo> WC What are you trying in moving from ActiveX to HTML/DHTML?
[3:59] <Jo> (inaudible)..trying to leverage other work being 
done within microsoft.
[4:01] <Jo> CMN the activeX controls are standardized?
[4:01] *** mcmay has quit IRC (Connection reset by peer)
[4:01] <Jo> There are 5 or 6, not from a library of standard controls.
[4:02] <Jo> CMN Are you using one standard Rich Text entry thing?
[4:02] <Jo> Mostly standard to other MS products.
[4:02] *** TimLa has quit IRC (Ping timeout)
[4:04] *** TimLa (~timlaranc@ has joined the channel
[4:04] *** cyns (~cyns@ has joined the channel
[4:05] *** mcmay (~Matt@ has joined the channel
[4:06] <Jo> WC Any recommendations or thoughts on dealing with 
older browsers?
[4:07] <Jo> RN to WC we don't worry about IE 3.
[4:09] <Jo> MS Project still viewed as a niche product, it's 
hard to get AT vendors to support it (as opposed to Office and 
other products with larger user base).
[4:09] *** mcmay has quit IRC (Connection reset by peer)
[4:10] <Jo> LGR Access keys?
[4:11] <Jo> LGR Using standard access keys?
[4:11] <Jo> Pretty similar.
[4:12] <Jo> GV trouble with collision with hotkeys in IE?
[4:12] <Jo> Yes.
[4:12] <Jo> Try to put access keys around the page rather than 
in one location.
[4:13] <Jo> Try to group the controls logically and provide 
access keys for each section to make tabbing easier.
[4:15] <Jo> CMN notes problems with collision with access keys
[4:16] <Jo> GV Problem with visual representation of access 
keys. One thing to think about is to expose access keys to 
screen readers.
[4:17] <Jo> GV With AT, maybe we have to say this assumes that 
you have a reader that works with current technologies and that 
will display these in standard way.
[4:17] <Jo> GV "We assume the AT has the following capabilities..."
[4:18] <Jo> TL Access Keys are also exposed in document object 
model. Up to AT vendors to decide how they're going to display 
[4:18] <Jo> WC What would you consider the applications?
[4:19] <Jo> Device independence. Not table layout. Working on 
that in future versions.
[4:20] <Jo> WC ActiveX controls?
[4:20] <Jo> GV Nested headers?
[4:21] <Jo> No, not on single page.
[4:21] <Jo> WC Any other things to warn us about as we move forward?
[4:22] <Jo> Standard for access keys. repetitive navigation. 
ability to skip long lists of links when tabbing.
[4:22] <Jo> CMN Browser manufacturers pushed back on access keys.
[4:23] <Jo> CMN smart browsers re-map access keys.
[4:24] <oedipus> smart browsers offer the user the option to 
pick a triggering mechanism for accesskey
[4:24] <oedipus> they also allow users to chose between 
establishing focus on the element or activating it, which needs 
to be a UA configuration option
[4:25] <Jo> GV Have you thought of having alt key combinations?
[4:26] <Jo> CMN Have you been looking at web authoring guidelines?
[4:26] <Jo> No, maybe good idea.
[4:26] <Jo> Note: very hard to hear back here...minutes 
incomplete and possibly inaccurate. Sorry.
[4:27] <oedipus> jo, something is better than nothing
[4:30] <chaals> do you want us to dial you in?
[4:30] <Jo> Demo over. Back to Guidelines.
[4:30] <Jo> GV proposes we make a list defining what we agree on 
vs. what we don't agree on.
[4:32] <Jo> GV refers to e-mail message he wrote this morning, 
on consensus.
[4:32] <Jo> GV That our guidelines should be usable by people 
who are writing regulations or requirements or policies. This is 
not the only group, but one group we need to address. Seems to 
be consensus on that.
[4:32] <Jo> GV That our guidelines should not necessarily be 
directly usable or adoptable as regs.
[4:33] <Jo> That our guidelines should not be limited to only 
that information that could or should be required today. They 
should talk as well about what would make web content more 
accessible even if it is not possible today.
[4:33] *** wendy (~wendy@tux.w3.org) has joined the channel
[4:35] <Jo> CMN In order to meet this checkpoint, there are 
things that are required and things that are nice to do. 
[4:35] <TimLa> <jo's phone rings.  Timla takes minutes
[4:36] <TimLa> Normative vs informative vs required vs possible 
vs applicable.  Still in email subject
[4:36] <TimLa> Above from GV
[4:37] <TimLa> GV Trying to find consensus on these items.  Are 
any candidates?  What about topics and issues list?
[4:38] <Jo> GV Do people feel that there was consensus on the 
things in the e-mail?
[4:38] <TimLa> 1) do perople feel there was consensus on  any items?
[4:38] <Jo> CM dangerous to put things that aren't possible yet 
in guidelines. lessens credibility.
[4:39] <TimLa> GV No consensus on more important vs less important
[4:39] <Jo> LGR We shouldn't require anything we don't have 
decent techniques for. That's agreed, right?
[4:39] <Jo> GV And examples?
[4:39] <Jo> LGR Yeah, techniques and examples.
[4:40] <Jo> MM People not in this room will disagree with this, 
perhaps. Can techniques be provided for 3?
[4:41] *** d-van (~boydonova@tide75.microsoft.com) has joined 
the channel
[4:41] <Jo> Sorry, that was KHW, not LGR.
[4:41] <Jo> KHS
[4:42] <Jo> KHS we must demonstrate whatever is normative.
[4:42] <Jo> CM do we have to demonstrate informative?
[4:42] <Jo> GV No. Put it out, might stimulate someone to come 
up with the solution.
[4:42] <TimLa> Cookies haver arrived
[4:43] <chaals> Mmmm
[4:43] <cyns> coooookie
[4:43] <TimLa> must..... focus.....
[4:43] <Jo> CM We shouldn't include anything as normative that 
we can't provide success criteria for them.
[4:43] <oedipus> damn, and my browser is set not to accept cookies
[4:43] <TimLa> That's not my fault.
[4:43] <d-van> we shouldn't include cookies that we can't 
provide milk for.
[4:43] <Jo> GV dissents.
[4:44] <Jo> GV did success criteria evaluation. Can say success 
criterion is the guideline itself. It's subjective whether 
something is objective or not.
[4:45] <oedipus> success criteria should only be quantitative at 
the technology specific level - at the higher levels they are 
merely expressions of necessary functionality/access to the "C" 
word (content)
[4:45] *** TimLa has quit IRC (Connection reset by peer)
[4:45] <Jo> CMN How we actually use this idea that we shouldn't 
be including this or that. Warn against using this to throw 
anything out before last call.
[4:45] <Jo> CMN just means we haven't done our work.
[4:46] <Jo> MM Where we have trouble with things being 
objective, removing them from the doc itself isn't the solution. 
The place is not necessarily within the core of the guidelines.
[4:46] <Jo> MM users need educational information that they need 
to read and know about. Not remove 3 because it can't be proven 
objectively. Put it where we can tell them how to succeed.
[4:47] <Jo> GV As informative they wouldn't need to be removed.
[4:47] *** Liddy (~Liddy@ has joined the channel
[4:47] <Jo> MM Are they going to be informative guidelines, and 
then we get back to conformance.
[4:47] <Liddy> oops! have I arrived?
[4:47] <Jo> MM if it's just there as informative, and you have 
someone whose goal is to achieve a certain level of conformance, 
then they can ignore it. need buy-in from stakeholders that 
they'll care about stuff we put down as informative.
[4:48] <Jo> GV we'll hit conformance in a bit.
[4:48] <Jo> CM As exercise, can we separate out what's testable? 
What's normative must be testable. Things that aren't testable 
should be in different document.
[4:49] <wendy> s/CM/CS
[4:49] *** TimLa (~timlaranc@ has joined the channel
[4:49] <wendy> initials going around the room:
[4:49] <wendy> CMN, CM, KHS, MM, CS, DH, ASW, JM, TL, LGR, RN, 
[4:50] <oedipus> what's the definition of "testable", please?
[4:50] <Liddy> mmm...!
[4:50] <Jo> thanks wendy -- trying to type too fast!
[4:50] <wendy> right on
[4:50] *** Jo has quit IRC (Connection reset by peer)
[4:50] <cyns> testable means that you can devise a test to 
determine that you have done the thing.
[4:51] <cyns> doesn't have to be automated, just verifiable
[4:51] <oedipus> machine-testable? or are there classes of testability?
[4:51] <wendy> gregory - do you want us to call you?
[4:51] <wendy> RN both qualitative and quantitative in a 
checklist, but they are subjective when you measure them.
[4:51] <cyns> doesn't have to be machine testable.  you can 
classify it as automatable or not if you like
[4:51] <oedipus> yes, please - i thought chaals was going to 
initiate the call
[4:52] *** Jo (~Jo@ has joined the channel
[4:52] <cyns> jo, do you want me to minut?
[4:54] <wendy> br people are already ignoring checkpoints, e.g., 
clear and simple.
[4:54] *** oeddie (~oedipus@ns.hicom.net) has joined the channel
[4:54] <wendy> gb implementation versus accessibility.
[4:54] <wendy> .. heuristics. is it valid alt? is it suspicious? 
is it testable automatically? no.
[4:55] <wendy> .. a checkpoint might require human judgement by 
a tester or actual users - simple language.
[4:55] *** d-van has quit IRC (Connection reset by peer)
[4:55] <wendy> .. I agree w/normative checkpoints taht we 
associate easily testable conditions.
[4:55] <wendy> .. other guidelines/checkpoints should be there as well.
[4:55] <wendy> .. perhaps suggest other ways of testign them.
[4:55] <wendy> .. e.g. user testing.
[4:56] <wendy> .. capture in a separate document.
[4:56] <oeddie> cyns - there sitll is the problem of verifying 
the "appropriateness" of equivalent alternatives, longdescs, 
simplest & clearest, etc.
[4:56] <wendy> GV Testing group developing tests?
[4:56] <wendy> WC No, WCAG technology-specific checkpoints.
[4:57] <wendy> GV As soon as we start talking about objective 
and subjective, also a continuum.
[4:57] *** Geeze (~love26@ has joined the channel
[4:58] <Geeze> whazzup
[4:58] *** Jo has quit IRC (Ping timeout)
[4:58] *** TimLa has quit IRC (Ping timeout)
[4:58] *** cyns has quit IRC (Ping timeout)
[4:58] <oeddie> can't hear much via the phone
[4:58] *** chaals has quit IRC (Ping timeout)
[4:59] <wendy> not much we can do to help.
[4:59] <oeddie> ok
[4:59] <wendy> big room. small phone. loud fans.
[4:59] <Geeze> 1.617.252.1859 is just me
[4:59] <wendy> if we've got mystic today and more than GR 
interested in participating, then we ought to call in to that.
[5:00] <Geeze> I thought that was the mystic?
[5:00] <wendy> we've called GR directly.
[5:00] <wendy> oops, i need to minute.
[5:00] <wendy> GV said something about addressing issues.
[5:00] <wendy> anyone can help?
[5:00] <Geeze> well, you'll call him but you won't call me!
[5:00] *** KatieHS (~kshea@ has joined the channel
[5:00] <wendy> CMN stuff that can be tested by a tool is almost 
worthless subset.
[5:01] <wendy> .. even a basic test by a tool is interpreted by 
someone.  you need to think about it to say, "it's fine."
[5:01] <wendy> .. some tests we are bad at describing that we 
could describe better.
[5:01] <wendy> .. part of problem is language complexity.
[5:02] <Geeze> (509) 773-4885
[5:02] <wendy> Restricting to testable by a tool is basically giving up.
[5:02] *** oedipus has quit IRC (EOF From client)
[5:02] <wendy> geeze - hold on.
[5:02] <wendy> cmn is waxing away.
[5:02] <wendy> once done i'll interrupt to get us all connected.
[5:02] <Geeze> I'm holding, I'm holding!
[5:02] <wendy> :)
[5:03] <Geeze> don't use ascii art without a d-link
[5:03] <wendy> you want a description of cmn waxing? :)
[5:03] <Geeze> No of :)
[5:04] <Geeze> I should stop this; Aaron will think I've gone senile!
[5:04] *** oedipus (~oedipus@ns.hicom.net) has joined the channel
[5:04] <wendy> ah ha!
[5:04] <wendy> yes, i usually use <grin/>
[5:05] <oedipus> geze - we're moving to mystic
[5:05] <Geeze> You mean "too mystic"?
[5:05] <oedipus> yeah...
[5:05] <oedipus> the mystic bridge, which i think you dialed before
[5:06] <Geeze> how can oeddie still be on with oedipus here?
[5:06] <oedipus> beats me - i keep losing connection due to 
local weather conditions
[5:06] <Liddy> can someone please tell me the schedule for this 
meeting ... and is it OK to stay on IRC? not phone?
[5:07] <oedipus> mystic: tel:+16172521859
[5:08] <wendy> liddy - the scheule is at: 
[5:08] <Liddy> thanks!
[5:08] <oedipus> liddy - the agenda is at 
[5:09] <Liddy> and this is east coast timing?
[5:09] <wendy> this is West coast.
[5:10] <oedipus> all times local to seattle (left coast USA)
[5:10] *** oeddie has quit IRC (EOF From client)
[5:10] <Geeze> not Perth!
[5:10] <wendy> WC in response to BR, people have been intersted 
in suggestions of granulated conformance claims. ERT WG working 
on EARL to do that.
[5:10] <wendy> .. that way, can ignore checkpoints, but won'nt 
through out more than might.
[5:11] <wendy> MM Put checkpoints into perspective.
[5:11] <oedipus> accessibility is merely specific instance usability
[5:11] <wendy> .. anything put forth in guidelines is not 
complete enough.
[5:11] <wendy> .. people need to be educated. can't do it in our 
[5:11] <wendy> .. comprehensive "how-to."
[5:11] <wendy> .. to increase accessibility.
[5:11] <wendy> .. WCAG 2.0 core is going to be appropriate for 
regulatory agencies, but for the people implementing it,
[5:11] <wendy> they would like more of a conversation.
[5:12] <wendy> .. practices for making the content related 
[5:12] <wendy> GV You said, "not removing the non-norm stuff, 
but making it ..?
[5:12] <wendy> MM It still needs to be prominent, but whether it 
belongs in the document as it is structure or crafted around 
those guidelines.
[5:14] *** chaals (~charles@tux.w3.org) has joined the channel
[5:14] <chaals> Hi all
[5:14] <chaals> wireless should be alive. But I don't have a log 
of it from lunchtime onwards :(
[5:14] *** cyns (~cyns@ has joined the channel
[5:15] <cyns> GB: We need to distinguish between guidelines for 
making a page accessible and for testing compliance.  We should 
have both normative, testable guidelines and informative 
guidelines.  Maybe those can be tested manually.
[5:15] <wendy> what about loggy?
[5:15] <cyns> GV: does the testing working group handle testing 
[5:15] <cyns> WC: no, that's why we have tech-specific 
checkpoints in this group
[5:15] <cyns> GV: as soon as we start talking about objective 
and subjective, we have to realize that that's a continuum.
[5:15] <cyns> just because something is subjective doesn't mean 
we should take it out, even if people blow it off.
[5:15] <cyns> CM: just testing with a tool isn't enough.  Does 
the text tell you what the picture is about.  There are some 
tests we could describe a lot better than we have.  Equivalency 
is hard to determine.  We shouldn't restrict ourselves to 
testable by a tool.
[5:15] <cyns> WC: A+ conformance
[5:15] <cyns> mm: need to have a comprehensive readable how-to.  
someone to explain to them how to do it.
[5:15] <cyns> better served by
[5:15] *** Jo (~Jo@ has joined the channel
[5:15] <chaals> Loggy isn't on :(
[5:15] <wendy> no way to turn it on?
[5:15] <chaals> Do you have a complete window wendy?
[5:15] <wendy> complete?
[5:15] <chaals> (could do copy/paste...)
[5:15] <chaals> the buffer since lunch...)
[5:15] <wendy> i've got from 17:49.
[5:15] <chaals> Hi Liddy
[5:16] <wendy> jo, e-mail me from:
[5:16] *** TimLa (~timlaranc@ has joined the channel
[5:16] <wendy> hmm, no timestamp. but I do have discussion of project.
[5:16] <wendy> my  machine was off at the time.
[5:16] <Jo> wendy, e-mailing you my bbedit notes now.
[5:16] <wendy> BR A reading of the document that takes into 
account __, if we include hard to test
[5:17] <TimLa> (thanks for getting the wireless back up, chaaaaals)
[5:17] <wendy> .. give someone reading just stand alone 
guidelines, what is being sought after in a checkpoint.
[5:17] *** mcmay (~Matt@ has joined the channel
[5:17] <wendy> GV Whether easy to measure or not, should be easy 
to understand the intent?
[5:17] <cyns> I pasted in my notes
[5:18] <wendy> GB Feasibility of test.  We draw lines at a 
certain point between normative and informative checkpoints, 
based on
[5:18] <wendy> .. activity required to test.
[5:19] <wendy> .. specialized skills, time, etc.
[5:19] <wendy> .. identify techniques and their cost, use that 
info to determine where the line is between normative and 
informative checkpoints.
[5:20] <wendy> GV Normativity determined by how easy it is to 
test or not.
[5:20] *** Geeze has quit IRC ()
[5:20] *** WHOIWAS Unknown command
[5:20] *** Geeze was ~love26@ (William 
Loughborough) on channel *private*
[5:20] *** on irc via server irc.w3.org (Mon Sep 10 18:20:15 2001)
[5:20] *** geeze: No such nick/channel
[5:21] <oedipus> if normativity is dertermined by how easy it is 
to test, how can the appropriateness of an equivalent 
alternative be tested? this is a BAD idea
[5:21] <wendy> TL Bad to base checkpoints on ease.
[5:21] *** Geeze (~love26@ has joined the channel
[5:21] <wendy> CS I want to clarify "easiness" of testing.
[5:21] *** Geeze has left the channel
[5:22] <wendy> .. not so much if it is easy, or costly, but if 
it is easy for people to agree on the result.
[5:22] <wendy> .. agree for people to say, "yes, it has 
alt-text."  it is is not easy for people to agree on, "this site 
is clear and simple and appropriate for the audience."
[5:23] *** Geeze (~love26@ has joined the channel
[5:23] <Geeze> I'm back
[5:23] <wendy> hi geeze.
[5:24] <wendy> cmn if we can't describe a test very well, we'll 
have a hard time making a checkpoint.
[5:24] <wendy> initialmania again:
[5:24] <wendy> cmn charles mccathienevile
[5:24] <wendy> cm charles munat
[5:24] <wendy> khs katie haritos-shea
[5:24] <wendy> mm matt may
[5:24] <wendy> cs cynthia shelly
[5:25] <wendy> asw andi snow-weaver
[5:25] <wendy> jm jo miller
[5:25] <wendy> tl tim lacy
[5:25] <wendy> lgr loretta guarino reid
[5:25] <wendy> rn rob neff
[5:25] <wendy> gb giorgio brajnik (sp?)
[5:25] <wendy> br bob regan
[5:25] <wendy> gv gregg vanderheiden
[5:25] *** mcmay has quit IRC (Connection reset by peer)
[5:25] *** TimLa has quit IRC (Connection reset by peer)
[5:25] <wendy> wc wendy chisholm
[5:25] <wendy> i apologize for any spelling mistakes.
[5:25] <wendy> or typos.
[5:26] <wendy> but, hope that helps y'all on irc only.
[5:26] *** mcmay (~Matt@ has joined the channel
[5:27] <chaals> jl == Liddy
[5:27] *** TimLa (~timlaranc@ has joined the channel
[5:27] <wendy> GV Perhaps a diff between writing guidelinesa nd 
regulations is: this is what do to make more usable...
[5:27] <wendy> .. claim to make ours at a certain level, and 
make sure they have.
[5:28] <wendy> CS described grading guidelines.
[5:28] <wendy> JM In educational testing, the stakes are high.
[5:28] <wendy> .. the graders have standardized grading criteria.
[5:28] <wendy> .. they have to back up the subjective grades.
[5:28] <chaals> Liddy, you're not on the phone, right?
[5:28] <wendy> CM There is a point in the Rec process where 
implementation comes in.
[5:29] <wendy> .. we might have devised great ways to test.
[5:29] <wendy> .. i sketched out a quick graph and made it into 
something pretty.
[5:29] <oedipus> can the mike be moved to chaals or chas when/if 
the graph is discussed?
[5:30] <wendy> .. it is a sequeway from where we are now to 
conformance testing.
[5:30] <wendy> GV talk about conformance or techniques?
[5:30] <wendy> CM Perhaps split the time?
[5:31] <wendy> KHS suggests we have homework.
[5:31] <wendy> laughter.
[5:31] <wendy> GV Vote: elephants or technqiues after break.
[5:31] <wendy> more elephants than ants.
[5:32] *** cyns has quit IRC ()
[5:32] <wendy> therefore, take one hour after break to discuss 
[5:32] <chaals> the graph: two axes - practicality and usefulness
[5:32] <chaals> A rough split - some things are first, some are last.
[5:33] *** TimLa has quit IRC (Ping timeout)
[5:33] <wendy> * GV walking through the lsit of conformance 
items he's been jotting down as we've been discussing *
[5:34] <oedipus> WL: is clear & simple normative
[5:34] <chaals> there are some things that are high on one scale 
but low on another - where do they come from
[5:34] <chaals> sorry - where do they come in the priority of 
doing stuff...
[5:34] <chaals> (this is my interpretation of chas' picture.
[5:34] <wendy> "candidate consensus statements"
[5:34] <Jo> GV If you can't decide what the success criteria 
are, we shoudn't include it. Feels like it should be true but 
seems worrisome. Can we hold back on this one?
[5:35] <chaals> http://www.w3.org/WAI/GL/2001/09/use-prac.svg
[5:35] <Jo> GV "Normativity is determined by how easy it is to 
test." Many objections. No consensus.
[5:35] <Geeze> that was sure a short break!
[5:36] <wendy> right, no break. before we break we're going to 
finish this up.
[5:36] <Jo> GV "Normative-ity is determined by objectiveness -- 
ease of establishing fulfillment." People are nervous, need to 
think about this.
[5:39] <chaals> ========
[5:39] <chaals> \break 15 minutes
[5:41] <oedipus> going to sub-shell for the next 15 minutes, but 
am hanging onthe phone line
[5:47] *** oedipus has quit IRC (Ping timeout)
[5:48] *** oedipus (~oedipus@ns.hicom.net) has joined the channel
[5:50] <Geeze> I asked if "clear and simple" is "normative". I 
might also ask if "normative" (the word, not the concept) is 
clear or simple?
[5:50] <Geeze> If the document being worked on is going to be on 
the Web, must it not conform to itself?
[5:51] <oedipus> that's what i'd call a rhetorical question - 
waxing anyone?
[5:51] <Geeze> perhaps rhetorical but not pedantic, IMO.
[5:52] <oedipus> clear
[5:52] <oedipus> clearly
[5:53] <Geeze> This document is like the "nutritional 
information" on a dog food can and the first ingredient has to 
be *meat*
[5:54] <Geeze> Of course the guy who ate dog food was killed by 
an 18-wheeler when he sat down to lick his balls on the freeway.
[5:55] <oedipus> actually it is like the perscription cat food 
that i once needed to buy for a feline - you couldn't buy it 
without a perscription from a vet, although i never caught any 
of the cats selling it to junkie alley cats
[5:55] <Geeze> A market for loopholes to the conformance criteria?
[5:56] *** TimLa (~timlaranc@ has joined the channel
[5:56] <Geeze> Where's Joe Clark now that we need him?
[5:57] <Geeze> Was all that logged?
[5:58] *** mcmay has quit IRC (Connection reset by peer)
[5:58] *** TimLa has quit IRC (Connection reset by peer)
[5:58] <Geeze> The phone is adequate for listening but doesn't 
qualify for talking, which in my case is probably for the best.
[6:00] *** chaals has quit IRC (EOF From client)
[6:01] *** wendy has quit IRC (Ping timeout)
[6:03] *** Liddy has quit IRC (bye... )
[6:05] *** oedipus has quit IRC (EOF From client)
[6:06] *** Jo has quit IRC (Ping timeout)
[6:07] *** wendy (~wendy@tux.w3.org) has joined the channel
[6:07] <Geeze> Did I drive everybody off the IRC?
[6:07] <wendy> :)
[6:07] <wendy> <grin/>
[6:08] *** SanchoP (~oedipus@ns.hicom.net) has joined the channel
[6:08] <Geeze> Miguel Cervantes welcomes you
[6:09] <wendy> chaals - pls bring up the list of issues.
[6:09] <Geeze> quixotically?
[6:09] *** mcmay (~Matt@ has joined the channel
[6:09] <wendy> sancho - nice nickname.
[6:10] *** chaals (~charles@tux.w3.org) has joined the channel
[6:10] *** TimLa (~timlaranc@ has joined the channel
[6:10] <chaals> we are back.
[6:10] *** SanchoP has quit IRC (EOF From client)
[6:10] <Geeze> what does "different content" mean?
[6:11] <wendy> * discussion about content, serving specialized 
views or not *
[6:11] <Geeze> If it means repurposed version of "actual 
content" OK but "different" very suspect.
[6:12] <wendy> KHS one for developer, etc...as long as all 
available in one place.
[6:12] <wendy> CM Like a database?
[6:12] <wendy> LGR Do they have to be tied together?
[6:12] <wendy> BR Guidelines in diff forms or pages evaluated by 
guidelines avail in diff forms?
[6:12] <wendy> KHS not guidelines, content.
[6:13] <wendy> GV We already allow for text equivalent page.
[6:13] <wendy> .. as last resort.
[6:13] *** oedipus (~oedipus@ns.hicom.net) has joined the channel
[6:13] <Geeze> charles is short-handed?
[6:13] <wendy> CM Issue of equivalency.
[6:13] <wendy> .. using grahpics and other things.
[6:14] <wendy> .. multiple redundant media.'
[6:14] <wendy> .. audio, video, text, etc. can not come from the 
same source.
[6:14] <wendy> .. 2 sep. items supposed to be equiv.
[6:14] <oedipus> thinking of dropping off, as i keep losing 
connections, and have a grinding headache, which the noise on 
the bridge is worsening
[6:14] <wendy> .. equiv alt text same idea.
[6:14] <wendy> noise?
[6:14] <wendy> what nosie?
[6:14] <wendy> gv source - a server or code?
[6:15] <wendy> cm a database.
[6:15] <Geeze> I keep hearing voices.
[6:15] <oedipus> don't hear conversations, so much as lots of 
background noise
[6:15] <wendy> hmmm from the phone or the phone line?
[6:15] <Geeze> get a proper headset!
[6:15] <wendy> nothing is near the phone...although the fan is 
kidn of close.
[6:15] <oedipus> geeze, you've been hearing voices since you 
were dropped on your head
[6:15] <oedipus> as a child
[6:15] <oedipus> problem probably at my end - weather is quite nasty
[6:15] <Geeze> Yeah but not GV's!
[6:16] <wendy> cm real question: how do we force equivalence. it 
is subjective.
[6:16] <wendy> GV serving content in diff forms acceptable way 
to conform as long as all available through same URI, tho may be 
linked to it.
[6:16] <wendy> CM Fine except that not all info avail in all 
forms. no true equiv between image and text.
[6:17] <wendy> .. key info might all be there...
[6:17] <wendy> GV Some people think that my pda is accessible 
since the keys are accessible.
[6:17] <wendy> .. the phone keys.
[6:17] <wendy> CM A description of graph tends to give all the 
info in the graph.
[6:18] <wendy> GV Phrase to use as term as art. As long as equiv
[6:18] <oedipus> can you read people the UAAG definition of 
conditional content?  i think that it might help - perhaps also 
UAAG GL2 intro text and CP 2.3 - VERY GERMANE and good to have 
harmonization of terms
[6:18] <wendy> oedipus - help with UAAG ?
[6:18] <Geeze> "functionally equivalent"?
[6:18] <wendy> gregory - URI?
[6:19] <TimLa> Wendy:  
http://www.w3.org/TR/UAAG10/guidelines.html#gl-feature-on-off I 
[6:20] *** Jo (~Jo@ has joined the channel
[6:20] <wendy> could someone minute, while i check that out?
[6:20] <Jo> yes
[6:20] <oedipus> wendy, i just cut-and-pasted the pertinent bits 
into an emessage & sent it to you
[6:21] <Jo> CM I think this is the biggest stumbling point in 
whole thing. Nothing is more subjective than "equivalence"
[6:21] <wendy> thx gr and jm!
[6:21] <Jo> CM Take "write clearly and simply." As you simplify, 
you are losing information.
[6:21] <Jo> CMN We are confusing equivalence and identity. (TL agrees)
[6:22] <oedipus> write clearly and simply applies to 
alternatives, as does "appropriateness
[6:22] <Jo> GV If you're playing a game of "name that tune," 
what is the equivalent of the tune?
[6:23] <Jo> CMN The written score then.
[6:23] <wendy> GR - don't see that that fits here in that this 
discusses which content to display. we're talking about what 
makes things equivalent and then how to represent that in markup.
[6:23] <Jo> GV we have same problem trying to caption spelling tests.
[6:23] <Jo> CMN the point is that equivalence is situational. 
For a given situation, you can achieve the goal, They are not 
identical, not the same thing. But equivalent.
[6:24] <Geeze> Can't take any more. Try again tomorrow PM at the 
ATAG/GL joint meeting.
[6:24] *** SanchoP (~oedipus@ns.hicom.net) has joined the channel
[6:25] <Jo> GV The gist is that server-side is considered a solution.
[6:25] <Jo> GB Situation depends on specific context and person.
[6:25] <Jo> GV So how is an author supposed to know how to 
create equivalent if he doesn't know the user in advance?
[6:25] <SanchoP> i wish that IRCsome oedipus would go away
[6:25] <Jo> GB No way to do it.
[6:26] <Jo> GV Add an issue: what is an equivalent.
[6:26] *** Geeze has quit IRC ()
[6:26] <Jo> CMN It's a piece of pornography. You know it when 
you see it.
[6:26] <Jo> GV Disagrees totally.
[6:27] <Jo> CMN same thing happens with equivalence. We haven't 
defined it, but we think we can write a text equivalent, and it 
[6:28] <Jo> CM Only in simple cases, and only in one direction. 
When we try to create graphic equivalent for text, we're not 
doing a good job.
[6:28] <Jo> CM agrees with GB that the problem is you don't know 
what's equivalent until you deal with specific person in 
specific context. Otherwise we're limited to usability studies 
and percentages, projections.
[6:29] <Jo> GV proposes test for equivalency. Run it through, 
run it backwards, and see if you get what you started with.
[6:29] <Jo> GV "The spirit is willing but the flesh is weak" 
into Russian, came back "The vodka is good but the meat is 
[6:31] <Jo> CM If you're going to do that, then you need to be 
up front about what information you're trying to communicate. 
Difficult to select.
[6:32] *** oedipus has quit IRC (EOF From client)
[6:33] <Jo> RN "Equivalent": want to separate text of checkpoint 
from testing of checkpoint.
[6:34] <Jo> RN There are exceptions, extreme cases to be 
addressed in techniques document. Add a caveat.
[6:35] <Jo> GV Where something applies most times but sometimes 
does not, we need to capture it. We're finding that we can't 
write flat rules without writing exceptions.
[6:35] *** TimLa has quit IRC (Connection reset by peer)
[6:35] *** mcmay has quit IRC (Connection reset by peer)
[6:35] <Jo> RN We're writing requirements, and we're intermixing 
"how do we test these" too much. In test report, I can note 
exceptions and sign off.
[6:35] *** oedipus (~oedipus@ns.hicom.net) has joined the channel
[6:36] <oedipus> clear
[6:36] <Jo> RN On list and in here we go back and forth without 
focusing on making a good well defined checkpoint.
[6:37] *** mcmay (~Matt@ has joined the channel
[6:37] *** TimLa (~timlaranc@ has joined the channel
[6:37] <TimLa> IRC is driving me nuts, BTW.
[6:38] <oedipus> IRC is quite irking
[6:38] <KatieHS> so........what's your point?
[6:39] *** chaals has quit IRC (Ping timeout)
[6:39] *** SanchoP has quit IRC (EOF From client)
[6:39] <Jo> I'm getting a little lost in the discussion on 
testing, not sure where we're going. Sorry.
[6:39] *** chaals (~charles@tux.w3.org) has joined the channel
[6:40] *** SanchoP (~oedipus@ns.hicom.net) has joined the channel
[6:40] <Jo> CM Idea of baseline conformance. There's a certain 
level that is a minimum, and there are levels above that. As 
opposed to checkpoint by checkpoint.
[6:42] <TimLa> Jo:  We've come back to this topic from this 
morning.  Needs recognition for accomplichment beyond baseline.
[6:42] <SanchoP> many people in the disability community 
perceive triple-A as just adequate
[6:42] <Jo> JM: This brings us back to the idea of degrees, that 
you don't stop once you've made your site baseline adequate.
[6:43] <SanchoP> organization by disability is EVIL
[6:43] <Jo> LGR You should be able to claim conformance for one 
class of disabilities but not all.
[6:43] <Jo> LGR this is valuable, for instance, for regulators.
[6:43] <Jo> KHS Hmmmmmmm, we've been here.
[6:43] <SanchoP> isolating by disability is ghettoization of the 
worst type
[6:43] <TimLa> Many groans from all on that one.
[6:45] <SanchoP> disability specific lists are best left to 
individual disability orgs, who do (and should) promulgate their 
own checklists for the disability type that they represent
[6:45] <Jo> CMN On a practical level, if I know I've got three 
blind employees, it's useful to know what I need to do for that 
group right away. But we don't want to give elephant stamps for 
going just a little way towards conformance.
[6:45] <SanchoP> those are the lists that people seeking to 
accomodate specific individuals in individual circumstances, but 
what happens when they hire a deaf employee - refit again?
[6:46] <wendy> yep, paul bohman brought this up a while ago.  he 
even created a conformance icon: http://www.webaim.org/wcag/logo
[6:46] <Jo> GV please clarify.
[6:46] <Jo> CMN The goal is: ALL. It's helpful for implementers 
to have stages along the way.
[6:47] <Jo> CMN It is also helpful in implementation to identify 
things that help people according to a set of needs.
[6:47] <wendy> UAAG conformance scheme should have been on 
required reading list for this meeting: 
[6:47] <Jo> CMN But these do not make good conformance. 
Situation dependent implementation decisions. Having the 
information about who is winning and losing is important.
[6:48] *** oedipus has quit IRC (EOF From client)
[6:48] <SanchoP> CMN - that is for EO to do, not GL
[6:48] <Jo> CMN What we should do is have the goal. The top 
level. Priority scheme behind wcag 1 was pretty good, centered 
on user needs. Would not change idea that priorities are based 
on level of impact on users.
[6:48] <Jo> CMN I would provide profiling, but not conformance 
level for helping one type of disability group.
[6:49] <Jo> CM to CMN. Disagrees. Thinks there should be 
baseline level, and above that every checkpoint results in some 
increase. If there's A, AA, and AAA, and they can't make it to 
AA, they'll stop at A and not go onwards.
[6:50] <SanchoP> hello - did anyone pick up on my comment that 
there are a significant number of "stakeholders" that consider 
triple-A the baseline
[6:50] <Jo> CM Another issue being raised: use metadata to mark 
checkpoints in different ways. Like what type of disability this 
checkpoint relates to, what device it relates to. To aid in 
generating reports. This is separate from conformance issue.
[6:51] <SanchoP> disability specific conformance should be 
something defined by "stakeholder" orgs, NOT the WAI
[6:53] <SanchoP> hotels that cater to specific disabilities? 
wouldn't cut it in the "real" world
[6:53] <SanchoP> picking and choosing will set off red flags
[6:57] <Jo> CMN to GR There is a mechanism. If you meet this 
list of checkpoints that are important to me, then you've solved 
my problems. Agree that stakeholder groups are the ones to ask. 
But we shouldn't provide those as conformance scenarios. Just a 
selection of profiles.
[6:58] <SanchoP> GJR to CMN - right, they should not - MUST 
not - be conformance scenarios - profiles handled by EO with 
specific stakeholder orgs
[6:58] <Jo> MM Perspective from content-provision side. There is 
a concept that there is an "Accessible" thing that exists, that 
you can attain. The single-A level is the industry standard for 
compliance (or 508 in the case of US gvt agencies).
[6:59] <Jo> MM This is a perception that is widely held outside 
this room.
[6:59] <SanchoP> accessibility is in the eye, ear, and/or hand 
of the beholder
[6:59] <Jo> MM whether an accessibility effort is made depends 
on how they think they can get that "Accessibility" noun on 
their site.
[6:59] <SanchoP> accessibility is a means to an end, not an end 
in and of itself
[7:00] <Jo> MM Think about how people are going to approach this 
with respect to adoption. Baseline needs to be there in terms of 
most important things to be done before you can get that noun on 
your site.
[7:00] <Jo> GV Can you say that in a bullet?
[7:00] <Jo> MM Come back and I'll try.
[7:01] <Jo> To a non-stakeholding entity, an org that doesn't 
have a pressing need (lawsuit, known member of disabled group on 
staff) -- they want to have something tangible to comply to that 
will be the barest mimimum that they can attempt.
[7:01] <Jo> MM Having an A plus is helpful, but realistically 
they're going to go for A.
[7:01] <Jo> LGR One item this morning: user vs. user groups.
[7:04] <SanchoP> can't hear the discussion
[7:05] <Jo> We're going to think of some examples of when user 
vs. user might come into play.
[7:07] <SanchoP> why user versus user?  braille on ATMs don't 
make 'em accessible, unless they provide non-visual feedback, 
and have a mechanism (such as a talking sign) that points the 
non-visual to the ATM, so that they know it is there in the 
first place
[7:08] <Jo> GV A U. Wisc example. They decided to go with level 
2 conformance. Turns out to be difficult, so now there's a 
committee to get it repealed. If they have to do too much, 
sometimes they do nothing. But we don't want to let them do less 
than the minimum.
[7:09] <Jo> CM Is it sequential? Do these in this order?
[7:09] <Jo> KHS Yes
[7:10] <Jo> CM People should be able to choose what they do 
beyond the baseline, in what order.
[7:11] <Jo> GV Physical is comparable to cognitive. Guidelines 
define a reach range that many people cannot in reality reach.
[7:11] <SanchoP> "reach" from a mobility standpoint is something 
that a UA should handle, but there needs to be something to 
reach for!!!
[7:12] <Jo> CMN Anyone who only meets one checkpoint gets an F. 
Great that they started, but they fail. Baseline gets a C. Etc.
[7:14] <Jo> GV You'd think companies would say great, I'll pass 
based on what I report. They'd rather have someone else evaluate 
them, because they don't want to be liable for fraud later. 
Criminal penalties for lying to the government, etc.
[7:14] <SanchoP> you can't choose profiles for physical access, 
why should you be able to do so for cyber-real-estate?
[7:15] * chaals thinks we are overrunning the techniques stuff.
[7:15] <Jo> BR The Univ. of Wisc. example is instructive. Likes 
the idea of using conformance levels pedagogically.
[7:16] <Jo> BR First level represents a starting point. A plan. 
Here's how you build up. So that you're not throwing out the 
attempt because you can't reach level 2.
[7:17] <mcmay> JM Users tend to leap to something that is not 
the first priority. They'll go into headers long before they do 
alt tags on the front page.
[7:18] <mcmay> I agree completely.
[7:18] <mcmay> GV Reviewing proposals for consensus.
[7:19] <mcmay> 1. serving content in different forms is 
acceptable for compliance as long as equivalents are provided in 
different forms, and all are available from the same URI.
[7:20] <mcmay> agreed
[7:20] <mcmay> 2. It is good to have levels of conformance 
rather than all or nothing.
[7:20] <mcmay> JM levels that include a baseline
[7:21] <mcmay> CM, CMN disagree
[7:21] * SanchoP serving identical/equivlet content in different forms
[7:21] <mcmay> CMN It's a necessary evil.
[7:21] <mcmay> CM changes his mind
[7:22] <mcmay> 3. There is a minimum set that conformance should 
not be possible without.
[7:22] <mcmay> agreed
[7:22] * SanchoP agree with 3
[7:22] <mcmay> 4. We want to have recognition for accomplishment 
beyond baseline
[7:22] <mcmay> agreed
[7:22] <mcmay> We should be able to claim conformance by disability
[7:22] <mcmay> several disagree
[7:23] <mcmay> 6. We should provide a way for people to measure 
access against particular disabilities but it should be not for 
[7:23] * SanchoP agree with 4 partially, but disagree with 5 
(conform by disability)
[7:23] * SanchoP 6 is the province/balliwick of EO
[7:24] <mcmay> 6 changed to:
[7:24] <mcmay> WCAG should provide a way for people to see the 
impact of items for particular disabilities, but not for 
[7:24] * SanchoP even with change, 6 is an EO deliverable
[7:25] * SanchoP SanchoP = oedipus = GJR = Gregory J. Rosmaita
[7:26] * SanchoP i can't hear a thing with mute off, so have to 
respond via IRC
[7:26] <mcmay> aww, and we were going to kick-ban you... hee!
[7:27] <mcmay> 6 changed to:
[7:28] <mcmay> GL should provide hooks in WCAG to allow WAI 
(EO?) to provide a way for people to measure access against 
particular disabilities but not for conformance.
[7:28] <mcmay> 7. Conformance levels should be based on user 
needs only (and not ease of implementation or ease of 
[7:28] * SanchoP agree with hooks comment - as stated earlier in 
IRC log (something to reach for)
[7:29] <mcmay> 8. Conformance levels should take user factors 
and testability into account.
[7:29] * SanchoP agree strongly with 7
[7:29] <mcmay> 9. Conformance levels should take user factors 
and testability and ease into account (at least on a gross 
[7:29] * SanchoP data should be collected/collated by the ersatz R&D WG
[7:30] * SanchoP user factors & testability also for R&D activity
[7:31] <chaals> CMN: it must be possible to test a checkpoint, 
but the difficulty of testing is irrelelvant to conformance 
[7:31] * SanchoP what about intersection with QA work?
[7:32] <wendy> R&D?
[7:32] <SanchoP> research and development, as per: 
[7:32] <mcmay> GV Some things are hard to do. Say you want to 
add a vocoder to a phone.
[7:33] <wendy> R&D is not chartered to do research, but to 
coordinate it, enlighten those doing it, stir it up a bit (and 
that's really, REALLY boiled down and a bit off the cuff give 
how long i've been sitting here today).
[7:33] <mcmay> Then they want to rewrite it, which is easier to 
do. Ease of use is a function of opportunity.
[7:33] <SanchoP> R&D isn't yet chartered, is it?
[7:33] <wendy> yes, clearly intersection w/QA.
[7:33] <wendy> our charter is approved, but we have not issued a 
call for participation.
[7:33] <wendy> therefore, it's not really a group yet.
[7:34] <SanchoP> congrats on approval
[7:34] <wendy> thx
[7:34] <mcmay> Analogy to the hotel, where you put the faucets 
in, but not the elevator
[7:34] <mcmay> RN There's a level where certain things are required.
[7:34] *** wendy has quit IRC (EOF From client)
[7:35] <mcmay> GV I worry about ease of use bumping something important
[7:35] <mcmay> RN Priorities can be shifted.
[7:35] <mcmay> GV Not in a normative document
[7:35] <mcmay> RN But in a revision.
[7:35] <mcmay> CM It's been 2 years since the last revision.
[7:35] <SanchoP> analogy to equal discomfort - 
[7:37] * SanchoP subject line of above-referenced post is "can 
accessibility be distinguished from usability?"
[7:37] <SanchoP> what's not going to help gregory (i'm asking 
for it now)
[7:38] <mcmay> 11. We shouldn't be including anything as 
normative that we can't provide success criteria for.
[7:38] <mcmay> 12. Things that are normative must be testable.
[7:38] <mcmay> 13. Normativity is determined by objectivity - 
ease of establishing consensus on fulfillment.
[7:39] <mcmay> End of meeting
[7:40] *** TimLa has quit IRC ()
[7:41] *** Jo has quit IRC (ShadowIRC 1.1b2 PPC)
[7:44] *** mcmay has quit IRC (Leaving)
[7:48] *** chaals has quit IRC (Time to go do something else)
[7:51] *** SanchoP has quit IRC (Leaving)
Log file closed at: 2001-09-10 7:55:25 PM
Received on Monday, 10 September 2001 21:03:01 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 16 January 2018 15:33:38 UTC