See also: IRC log
<trackbot> Date: 05 July 2012
<JF> scribe: JF
<MikeSmith> scribe: MikeSmith
JF: status is that we're at a
point where we're in general agreement
... there's an appetite for find a solution
... comes down to a question of syntax
... both schools of thoughts are that we would introduce a transcript attribute
... Silvia proposes that it should take a URL
... sort of looks and feels like a magic longdesc
... The proposal for Ted and Eric is that it would take a list of IDREFs
... which would allow you to link to multiple transcripts
... I think at this point what needs to be communicated to he chairs is that there will be multiple change proposals
... and we will need to have a WBS survey
<paulc> I have very intermittent Internet access and may drop off at any point.
JF: I think we are going to need some more time, maybe around middle of July
<paulc> Both Silvia and Ted have said they will update their CPs by the end of this week.
janina: can you say more about the disagreement is?
JF: Apple feels the single URL
solution is too limiting
... Either proposal satisfies our primary use case
janina: I wonder if we're worried
about space-delimited sets being robust
... what happens when things don't work right?
JF: I think you'd get a 404
... it's an authoring error
... what happens now if I reference an ID that's not actually in the page?
<JF> scribe: JF
MS: sounds like a survey for Issue 194 is inevitable
so the TF should start thinking about looking at the emergent proposals to determine which the TF could/should endorse
MS: is Rich here?
Steve F: posted some questions to the list about the CP
have some serious reservation around the light-weight object stuff
I've been asking about this, but have heard nothing to date
MS: looking at the next steps from WG side
don't believe there is a deadline
not sure what to do next
<paulc> >John asked questions about current CP. Awaiting feedback from Frank/Ted.
<janina> zaim, unmute me
<paulc> I believe we are waiting for answers to Steve's questions.
MS: when did you pose those questions?
<Stevef> email to html working group on canvas http://lists.w3.org/Archives/Public/public-html/2012Jun/0180.html
SF: asked here
MS: the HTML WG call has been cancelled this week
should bring it up there
JS: perhaps Steve can re-post the question and note that there has been no response to date.
PC: will send a note to Ted/Frank and ask if there is any progress, but note that US holiday was Wednesday
SF: Did also pose these questions after the F2F, so they are not new.
quite happy if someone else can explain this, but with no response at all there remains some concerns
<paulc> My Internet access is very bad so I am at risk of dropping off. Sorry.
MS: who is driving this issue?
JS: believe we are waiting for Ted/Jonas response
<paulc> John F provided a revised CP on Tue Jun 26. Next step is to gather support for this within A11Y TF and with Ted/Jonas.
MS: Jonas is back from vacation now
JF: latest CP is about as far as we can go w.r.t. tab focus issue
next step would be to seek TF backing
JS: need to hear back from Ted/Jonas to see what they think
MS: perhaps another note out to Ted/Jonas?
PC: sending off that note now
MS: this is a new issue from the Chairs, based on what we were calling Issue 31c
deadline for (Counter) Proposals is later this month (24? 27?)
it would be good to see if we have TF support for this CP
MS: likely that there will be alternative CPs to try and address the same use case
but for now, the only CP is to drop this exception
JB: it is specifically to drop the exception around whether or not there is alternative text, and not to drop the generator tag
JS: Mike S are you suggesting that there may be other ways to avoid checking for missing alt
MS: the issue is if the document is machine generated that requires alt text,but no human is involved and images are missing alt, then it shold be recognized as a machine generated page
<Judy> Link to the re-updated re-open request, on which we're requesting TF support, to be followed by a request for PFWG support, and based on which the topic was re-opened: http://www.w3.org/html/wg/wiki/ChangeProposals/Issue31cMetaGeneratorUpdated
believe that henri S believes this is a valid use-case
<Stevef> validator based on validator.nu provides means to filter out any errors not desired author http://validator.keegan.st/
SF: want to point out that a validator can be set to NOT check for anything
The question is what is required for validation
MS: what we want to do on TF side is to consider if we want to take a Resolution to Support from the TF
JB: missed a part of what was being said
one thing discovered is that pages may be processed by multiple tools at multiple times. Very difficult to determine if a page is machine generated alone
not sure what may be coming forward
MS: what Henri has argued in the past is that he does not want the Validator reporting errors if/when a tool is designed to generate pages that contain images (bulk photo uploads), that people who make those tools, and the validator generates errors
then the makers of those tools will respond by adding 'bogus alt content' to silence the validator
JB: this is not a new speculation
hope he provides evidence of that, else it sounds like a 2-layer speculation. Concerned that it will start a loop
MS: would like to get consensus from Henri, but unsure if he will write another CP
SF: Understand the concern, but this is why we have <figcaption>
Authors/machines can put 'some' some differentiating text, that is clearly identified as a caption, and not an ALT
SF: these tools already are providing something so we prefer that they use the caption instead of a crummy alt text. never seen a tool that doesn't provide some text that could be used as the caption
MS: I need to discuss this further with Henri. Note that a document-wide switch is an anti-pattern.
JS: we also need some evidence. this seems to be based on speculation
<richardschwerdtfe> dropping off
JB: still concerned that a partial solution still creates most of the same problems we were concerned about
MS: this has been about validator behavior from the start
one is to look at the link that Steve posted
if handled correctly, i believe this can be "good"
this does not allow / hide any error messages initially, but it also allows the author to go in and filter error messages
gives control over which error messages they have to "hear"
however even if it suppresses the error messages, it does not claim "validation" if there are still errors
MS: believe this is a smarter way to handle it
JB: so the person doing the validation would hve to be pro-active - to consciously choose to suppress error messages
JS: it's a filtered view
JB: perhaps the results would need to have a disclaimer?
that it is not giving a conformance indication
MS: believe that this site is existing proof that this can be done. What we don't want is people shopping around for a validator that doesn't tell them they have errors
so we likely want to try and accommodate the need here.
JF: an error is an error. We should be worried about end-users, and not Validator tools
if a page does not meet user-needs (images with no alt text) then it is an error
SF: the filtering issue allows to have a non-wordy report, but still not generate a "Valid" result
allowing the author/tester to set preferences with filters is a good tool choice
JB: speculation - if there are an annoying number of validation errors, then tools will insert bogus values. What if, instead we speculate that tool vendors will improve their tools
tools are emerging that when people take photos they can tag them at that time
perhaps the error messages will drive innovation and solutions that actually address the problem
MS: agreed, it keeps the problem visible
JB: solves the right problem
MS: think I can talk with henri and see if we can get closer to agreement, and maybe metagenerator will end up being dropped without counter CPs
will defer agenda item 5 to next week
JB: there was no meeting this week. still working on buggy alt guidance and buggy alt text
evolving discussion there,
JS: David MacDonald is back and able to assist
MS: anything else?
JS: 1 point - what to do with bugs that we might want to revisit
lately discovered that the media elements should be directly focusable, and not explicitly rely on controls
JF: suggest that next week we have a generic "media bugs" topic
MS: should invite Bob to the call next week
JF to invite Bob to next weeks call
JS to work on Silvia's media focus bug
This is scribe.perl Revision: 1.136 of Date: 2011/05/12 12:01:43 Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: RRSAgent_Text_Format (score 1.00) Succeeded: s/John/Steve/ Succeeded: s/sending of/ sending off/ Succeeded: s/inprove/improve/ Succeeded: s/and get metagenerator dropped/and see if we can get closer to agreement, and maybe metagenerator will end up being dropped without counter CPs/ Found Scribe: JF Inferring ScribeNick: JF Found Scribe: MikeSmith Inferring ScribeNick: MikeSmith Found Scribe: JF Inferring ScribeNick: JF Scribes: JF, MikeSmith ScribeNicks: JF, MikeSmith WARNING: No "Present: ... " found! Possibly Present: Cooper JB JF JS John_Foliot Judy MS MichaelC Microsoft Mike MikeSmith P15 P6 PC Rich_Schwerdtfeger SF Stevef davidb html-a11y inserted janina joined paulc richardschwerdtfe tm trackbot You can indicate people for the Present list like this: <dbooth> Present: dbooth jonathan mary <dbooth> Present+ amy Agenda: http://lists.w3.org/Archives/Public/public-html-a11y/2012Jul/0006.html Found Date: 05 Jul 2012 Guessing minutes URL: http://www.w3.org/2012/07/05-html-a11y-minutes.html People with action items:[End of scribe.perl diagnostic output]