<alastairc> WCAG 2 issues https://lists.w3.org/Archives/Public/public-wcag2-issues/2024May/0008.html
I can do it
<knights> I can also jump in if needed
<alastairc> scribe: Detlev
<scribe> scribe: Detlev
Alastair: Introductions?
<Chuck> Welcome Ken Franqueiro!
Ken Franquiero introducing himself
Alastair: any topics for future meetings?
mGower: ChatGPT 4o shows tech is moving fast
<jaunita_george> +1000000
introduce it as point of consideration
Alastair: let's keep it I mind for discussing alt text
<alastairc> https://github.com/w3c/wcag2ict/issues/372
Alastair: Tops is WCAG2ICT review
there were 4 things to review
<alastairc> https://github.com/w3c/wcag2ict/issues/369
Alastair: how to interpret
consistent help
... applies almost as is
... no negative comments
... any further comments?
<alastairc> draft RESOLUTION: Approve issue 369
+1
<PhilDay> +1
<Azlan> +1
<Chuck> +1
<Jennie_Delisi> +1
<knights> +1
<dj> +1
<kevin> +1
<Poornima> +1
<Sam> +1
<Rachael> +1
<jon_avila> +1
<sarahhorton> +1
<ShawnT> +1
<kirkwood> +1
<bruce_bailey> +1
<Frankie> +1
<laura> +1
RESOLUTION: Approve issue 369
<julierawe> IRC kicked me out. Can you please reshare the link to the issues you're discussing?
Next is accessible authentication - little change going into WCAG2ICT
<Chuck> https://github.com/w3c/wcag2ict/issues/368
<alastairc> https://github.com/w3c/wcag2ict/issues/368
<alastairc> draft RESOLUTION: Approve issue 368
<Chuck> +1
<PhilDay> +1
<rscano> +1
<Azlan> +1
<kirkwood> +1
+1
<Makoto> +1
<jaunita_george> +1
<Frankie> +1
Chuck: TF Meeting discussed Shadi's suggestion - group resolved not to include it
<GreggVan> +1
Alastair: :you may still copy & paste
Kevin: Will there be a response to Shadi's suggestion?
MaryJo: Will respond
<mitch11> +1
Alastair: :you have 5 days from to day to still change your mind
<kirkwood> no
Gregg: Many people don't have own computer - then password manager doesn't work
Alastair: you need your own phone
<kevin> qq+
<ShawnT> You can log into a web base password manager
Gregg: : It's about people without devices - or shall we assume that users have password managers?
<jspellman> L
<jon_avila> When I log into Chrome with my Google profile I am able to use the saved Google passwords.
<kirkwood> aging person going to library
<Zakim> kevin, you wanted to react to GreggVan
Kevin: There are web-based password managers where you need to log in
Gregg: then we are OK
Alastair: there are other options
too
... any minus 1?
Gregg: are there USB solutions?
<kirkwood> usb breaks security protocols
Alastair: the site can offer multiple options - e.g. USB or passkey or copy/paste from text file on USB stick - the author's responsibility is not to block it
RESOLUTION: Approve issue 368
<Chuck> https://github.com/w3c/wcag2ict/issues/364
Alastair: was Review before - TF
took comments and reworked it - difference between 2.2 and
older versions with a note - so it now better aligns with the
differences in version
... there was a wording update too
MariJo: there is a note in WCAG
2.2 as well, and a different one in the errata...
... need to check - the text seems to be different in 2..0 and
2.1
Alastair: The editor's note will be removed to the references then make to sense
<bruce_bailey> Errata for 2.0, 4.1.1 is at bottom: https://www.w3.org/WAI/WCAG20/errata/
Alastair: markdown was probably a typo
Phil Day: Markdown is an example of a markup language - change may not be nneeded
Alastair: there is reference to LATEX and markdown - but yes, OK
<dmontalvo> +1 to Phil
Dan: yes the reference to markdown is inntentional
Jon: wasn't apparent, thought it might be confusing
Alastair: The examples do make sense - markdown seems alright
<alastairc> draft RESOLUTION: Approve issue 364, with small amendments for wording on the WCAG errata/note
+1
<bruce_bailey> +1
<Chuck> +1
<mitch11> +1
<PhilDay> +1
<dj> +1
<dan_bjorge> +1
<laura> +1
<Azlan> +1
<Makoto> +1
<maryjom> +1
<Sam> +1
<ShawnT> +1
<JenStrickland> +1
<Frankie> +1
<kirkwood> +1
<jaunita_george> +1
<Poornima> +1
<jon_avila> +0
RESOLUTION: Approve issue 364, with small amendments for wording on the WCAG errata/note
<rscano> +1
Alastair: again you have 5 days to add comments o the issue
<Chuck> https://github.com/w3c/wcag2ict/issues/370
Alastair: : There are definitions, comments one closed functionality, SCs that have difficulties in transposition
MariJo: : was used on page titled (?)
Alastair: any blockers?
<alastairc> draft RESOLUTION: Approve issue 370
<Sam> +1
<PhilDay> +1
<Chuck> +1
<mitch11> +1
+1
<maryjom> +1
<ShawnT> +1
<kirkwood> +1
<dj> +1
<kevin> +1
<Makoto> +1
RESOLUTION: Approve issue 370
<rscano> +1
Alastair: Tired to pin down best kind of structure for text alternatives - taking a decision tree approach
Thread....
tried to pin down
Alastair: (describing traversal
of decision tree)
... how important is image, what ways can alternative be
provided, is there text in the image, etc
... Question to group: Looking at that structure does this seem
like a reasonable approach?
... In this case, tests are similar for different methods - in
others, tests would be different
Wendy: : likes decision tree - wonder if we should bring up higher in the tree what the image is being used for - part of a control?
Bring that up higher is closer to people's intetnions
Alastair: we started that way, gone back and forth on that
MikeGower: Looking at the demos of ChatGPT 4o incorporates decent visual cognition of line graph, did a good job, even when queried. So what should the author's role be in that situation?
<kirkwood> author role - editorial, approval
MikeGower: importance of image is
crucial - currently importance comes after telling whether alt
is concise - but intent / importance is more importat
... it should be primary consideration for authors'
responsibility
<jspellman> Kelly Ford, a blind accessibility expert recently posted on LinkedIn that ChatGPT4 still has a long way to for accuracy.
dj why 4o??
<Zakim> alastairc, you wanted to comment on relying on ML/AI for meeting outcomes
<kirkwood> +1 an author adds an image with intent
Alastair: logic of decision trees is worth checking so they are more future-proof
a thanks
<kirkwood> ‘important’ is useless
<laura> Image Accessibility Creator By Arizona State University. Uses the GPT-4 vision model provided by OpenAI.
<laura> https://asuo-ai-labs.streamlit.app/Image_Accessibility
<alastairc> and wendyreid
Alastair: got methods in place - but we can't rely one demos, don't know how well distributed they will be, how well they respond to different types of images - so that may become part of a rationale - but not there yet
<JenStrickland> +1 to wendyreid
Wendy: AI is "a" method to achieve the goal - should not get lost in assessments how good it is right now
<kevin> +1 to wendyreid
Alastair: it is good to check
that we are not blocking things
... there have been a few comments
<Zakim> mbgower, you wanted to say I am not suggestion 'author's don't need to do anything'. I'm suggesting that assessing for importance seems as valid as 'decorotive' as a top level
Alastair: native iOS Android was just included as example
MikeGower: : not suggesting that
authors have nothing to do - just pointing out that looking at
AI methods is a valid consideration to lead to valid
methods
... our language should allow existing technology to be
updated
... it is a prime concern of COGA - maybe we should engage with
HTML 5 and see what can be added
Bruce: do we have good traction on "informative" as binary decision?
Alastair: currently it differentiates between decorative and everything else
<kirkwood> +1 to Bruce
Bruce: :we need a better language on decorative and informative
Alastair: current question is would it change the understanding of the page if ig was missing
<kirkwood> agree with Bruce
Bruce: img was important to page designer - so not easily binary
Alastair: Dan also asked about
img importance
... we got rid of "medium" (sharing screen)
... (talking about screen, editing text)
Jennie: can we map the importance to the way we cover ARIA levels (?)
Alastair: Bruce is saying instead
of a brief description, a brief identification and a longer
description are provided
... (reading Bruce's further comments)
Bruce: : the methods so far talk about text equivalent but do not cover long description (?)
Alastair: it is in the decision
tree - pick one method - they need to be equivalent - an image
will need a description various ways of providing that
... you can do more than one, but need one to conform
... branching important for that
... reading longish comment by Dan (sorry)
<Zakim> Rachael, you wanted to explain
Rachael: user studies show that sometimes image is put in to be explored and al text cannot cover that aspect
Dan: asking about equivalent purpose....
Rachael: when you have exploratory images, the alt text would also need to be exploratory (?)
Alastair: if we combine things, the methods need to be applicable to these scenarios
<kirkwood> surfacing that the author’s intent for placing the image. rather than leaving the user in the dark of the author’s intent. an attestation statement by author could be binary
Alastair: if a method is in the
same branch the author could choose either to conform - need to
be careful about that
... reads Dan's comment when alt is provided by the platform
(say for thumbs don emoji
<Jennie_Delisi> * ARIA similarities to the question of "informative" could be "relevant"? (from the previous discussion raised by bruce_bailey). Not sure if mapping images in this way could support some of the importance of an image conversation.
dj emojis are text
Alastair: emojis may be images or aria-label may override text - shouldn't d that so platform default should come through
<mbgower> +1 "it's more complicated than that"
<jon_avila> I have come many emojis that don't seem to have textual equivalents.
Dan: more complicated - emojis aren't always announced - implementations differ (say NVDA) verbosity level dependent, varies between Screen readers - comes up all the time as consistency problem
Alastair: intent if this brach
was things with a default that should not be overridden - maybe
take out branch of no and mull over it?
... if image contains text - Dan though it confusing
... this one got orphaned - was intended as addition - but
could apply to any of the branches
Dan: should go into the definition of equivalent text alternatives
<rscano> emojii description has a standard, shall require to refer to alternative official text https://www.unicode.org/reports/tr51/
Alastair: Assertions for siler/gold level - alt following style guide - leading to consistent implementation - should be better kept separate
<Zakim> Chuck, you wanted to ask for scribe change
<bruce_bailey> scribe+
<knights> Sorry slow typing -- but can alternate :)
<laura> Need to drop off the call. Bye.
<bruce_bailey> Alastair: We have a few updated from comments today...
<bruce_bailey> ... Wendy suggested controls first and that was where we started with an earlier iteration. Welcome to others to make proposals.
<alastairc> Alternative: https://docs.google.com/document/d/1CIj5LZA5wSySDSg5zqsDUWyX9-CU8tE9ab_fZN8dcKU/edit
<bruce_bailey> ... Rachael had some notes she wanted to share.
<bruce_bailey> Rachael: Shift is away from tester perspective, this came out of the evolution of Focus Indicator methods discussion....
<bruce_bailey> ... Suggestion from Wilco for what / when / how -- which didn't seem to work at higher level but does seem to working for metihods.
<bruce_bailey> ... So that is the biggest change here, same content as before but different layout.
<bruce_bailey> alastairc: Chair hat off, this seems like methods are instance or type -- what is the thing you are giving a alternatives?
<bruce_bailey> ... Methods could be next level after that.
<kirkwood> +1 that was tripping me up to
<mbgower> Remove the word "Method" and consider that way
<bruce_bailey> Rachael: Agree per our working definition, but that leaves us with awkward middle level. So we are working through what is the right number of levels.
<bruce_bailey> alastairc: I am trying how to get a sense how the tree approach is working.
<bruce_bailey> wendyreid: Looking at this from perspective of person working for guidance. Question is what are you doing with the image on site? Are you conveying information? ...
<bruce_bailey> ... What is designer trying to do? What is purpose of image on the page? If a button, designer knows that from start and will be looking for that in the tree.
<bruce_bailey> ... How can this guide that person?
<bruce_bailey> jaunita_george: From that perspective, does this have to be so explict and normative and closed off? ....
+1 to Wendy's take
<bruce_bailey> We are going to want to add more examples and methods as we go along.
<bruce_bailey> alastairc: Chair hat off, that is what we tried to do with 2.x Techniques, but that might be different question than normative or not....
<bruce_bailey> ... With model working through for 3, we kind of need the top most level and bottom most level to be normative and testable , and the middle parts informative.
<bruce_bailey> ... We have spent more time with first approach. But are folks feeling comfortable with second approach as well?
<jon_avila> e.g. the hide image would need to be normative
<bruce_bailey> dan_bjorge: Refering to example shared on screen, there are normative parts definitely in the middle.
<wendyreid> +1 the methods can inform the requirements
<bruce_bailey> alastairc: We have normative outcomes, some clearly informative parts, and then some methods which are required.
<bruce_bailey> ... Work with both approaches, looking for feed back.
<mbgower> Thanks for this, Rachael.
<kirkwood> thank you Rachael!
<dan_bjorge> I disagree fundamentally - I think this is more complicated to read than WCAG 2 and that it's especially going to create a ton of complexity for folks that need to simultaneously meet WCAG 2 and 3
<bruce_bailey> John Kirkwood: Does seem simpler (as compared with 2.x).
<dan_bjorge> I continue to maintain that iterating starting from WCAG 2 would be better, but that ship has sailed
<bruce_bailey> alastairc: Please think about working with this approach, possibly for years.
<kirkwood> sorry bruce i said doesn’t ;)
<jon_avila> Dan, I was thinking perhaps of some mapping from 3.0 to 2.x for each of the requirements.
<bruce_bailey> ... SC may seem simpler and shorter. But many are very complex in practice.
<bruce_bailey> s/Does seem simpler (as compared with 2.x). /Does NOT seem simpler (as compared with 2.x). /
<bruce_bailey> Mike Gower: These were sent out to listserv about a week ago. Mostly house keeping this time around. Some are just bug which would normally merge with just a note to list rather than bring up on a call...
<bruce_bailey> One item, under Target Size, is really editorial but is in the normative text so want to highlight here....
<bruce_bailey> ... Not changing any interpretation of the SC, but is errata changing a word for clarity and consistency with other phrasing of other SC...
<bruce_bailey> ... That is the most severe of this house keeping. Another example is just removing ALL CAPS.
<alastairc> https://github.com/orgs/w3c/projects/56/views/1
<bruce_bailey> Just 9 of them, should only take 15 minutes or so. Please do affirmatively add a thumbs up.
<bruce_bailey> We will keep any normative erratta to no more than once a year.
<bruce_bailey> alastairc: Anything else?
<alastairc> TPAC survey: https://www.w3.org/2002/09/wbs/35422/TPAC_2024/
<bruce_bailey> Rachael: From TPAC survey, we will be reserving a room.
<Rachael> TPAC information: https://www.w3.org/events/tpac/2024/
<bruce_bailey> Chuck: Reminder on 5 business day review of WCAG2ICT items.
This is scribe.perl Revision VERSION of 2020-12-31 Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: Irssi_ISO8601_Log_Text_Format (score 1.00) Succeeded: s/4.0/4o/ Succeeded: s/MariJo: :Will respond/MaryJo: Will respond/ Succeeded: s/4.0/4o/ Succeeded: s/nit/not/ FAILED: s/Does seem simpler (as compared with 2.x). /Does NOT seem simpler (as compared with 2.x). / Default Present: dj, alastairc, JakeAbma, Jennie_Delisi, dan_bjorge, bruce_bailey, Azlan, Justine, scotto, mbgower, jtoles, wendyreid, maryjom, ljoakley, Nayan, giacomo-petri, mike_beganyi, jon_avila, Graham, Jen_G, Rachael, kirkwood, Laura_Carlson, Kimberly, Ben_Tillyer, Glenda, ShawnT, julierawe, Francis_Storr, Theo-MSFT, Detlev, hdv, jspellman, rscano, Frankie, kevin, Makoto, knights, Poornima, jaunita_george, Chuck, sarahhorton, AlinaV, kfranqueiro, PhilDay, Sam, mitch, tburtin, sohara, GreggVan, Daniel, JenStrickland, MJ, ashleyfirth, GN Present: AlinaV, Azlan, Chuck, JakeAbma, Jennie_Delisi, Justine, Rachael, alastairc, bruce_bailey, dan_bjorge, dj, giacomo-petri, jtoles, ljoakley, maryjom, mbgower, sarahhorton, scotto, wendyreid, kfranqueiro, Detlev, PhilDay, Sam, mitch11, kirkwood, julierawe, tburtin, sohara, GreggVan, Glenda, Daniel, JenStrickland, MJ, ashleyfirth, Kimberly, mike_beganyi, knights, GN015 Found Scribe: Detlev Inferring ScribeNick: Detlev Found Scribe: Detlev Inferring ScribeNick: Detlev WARNING: No date found! Assuming today. (Hint: Specify the W3C IRC log URL, and the date will be determined from that.) Or specify the date like this: <dbooth> Date: 12 Sep 2002 People with action items: WARNING: IRC log location not specified! (You can ignore this warning if you do not want the generated minutes to contain a link to the original IRC log.)[End of scribe.perl diagnostic output]