See also: IRC log
<trackbot> Date: 09 May 2013
<janina> Meeting: HTML-A11Y Task Force Teleconference
<MarkS> scribe: Mark Sadecki
<MarkS> scribeNick: MarkS
<Suzanne_T> 785 is me
<MarkS> JS: any thoughts we should consider before diving in and testing HTML spec
<MarkS> …recap. HTML WG walked the entire spec, outlining what needed to be tested and what could be passed, emphasis on getting out of CR. If feature was widely supported, it need not be tested.
<MarkS> …possibly a good approach for us
<MarkS> …identify features that are well supported, like ALT, there are at least two browsers that support it successfully.
<MarkS> …in the relatively near future, perhaps we should walk through the spec to identify things that need to be tested
<MarkS> JF: it took HTML WG a minimum of 3 hours to do that at the F2F.
<MarkS> …we would have to carve out a time to do that.
<MarkS> …should we try to setup two 90 minute sessions at a different time to accomplish something like this.
<MarkS> JS: we did this before and that is the approach we took.
<MarkS> …spec is long
<MarkS> …not eager to do this in the coming month, but perhaps in the next couple months.
<MarkS> JS: is this not the way to proceed?
<MarkS> …no objections. will try to schedule in the coming weeks
<MarkS> JS: HTML markup for rowgroup does not seem to be in the spec.
<SteveF> Rich is in the house
<MarkS> RS: what happened for whatever reason is that Firefox introduced rowgroup, into the API mappings. Turns out they were doing this for thead to have a rowgroup.
<MarkS> …looked at native semantics in HTML5 and it wasn't even in the list.
<MarkS> JF: was wondering about colgroup?
<MarkS> RS: ARIA doesn't currently have colgroup, but could add colgroup to ARIA 1.1
<MarkS> CS: does it matter?
<MarkS> RS: Firefox wanted to have everything for what they have in HTML5
<MarkS> …that is how it showed up.
<MarkS> CS: how big of a deal is this? is it a necessary thing.
<MarkS> JF: thead, or rowgroup?
<MarkS> CS: rowgroup
<MarkS> RS: what does thead map to?
<MarkS> CS: it can change the mapping of its children, but otherwise not specified
<MarkS> JF: i think you could fix the position of thead and tfoot and scroll the body of the table
<MarkS> …could be good for mobile
<MarkS> JS: i think we negotiated this back in 2008
<MarkS> SF: IE does not expose even a TR, at least not in MSAA
<MarkS> RS: firefox added a rowgroup role in. not sure what their motivation was.
<MarkS> …thought they needed it for thead
<MarkS> JF: has anyone talked to dbolter?
<MarkS> RS: they wanted to have every HTML element to have a mapping.
<MarkS> JF: rowgroup has been around for a really long time. perhaps they just wanted to get everything in there. is there anything we really need to do?
<MarkS> RS: do we want to do that consistently across the platforms.
<MarkS> CS: MSAA is not supposed to have random strings in the role.
<MarkS> RS: you can do it in UiA
<MarkS> CS: yes
<MarkS> JS: suggest Rich ask the HTML WG to ask if there was a working group decision or approach to implement this, or if they were on their own.
<MarkS> RS: from what I understand, they were on their own.
<MarkS> RS: we have a table mapping HTML elements in ARIA that never included these elements. Was this an oversight?
<MarkS> CS: maybe they weren't in the spec at that time.
<MarkS> SF: thead has never been taken out of spec, nor tbody. I don't understand what is being asked? what are we suggesting the HTML WG do in regards to rowgroup?
<MarkS> …sounds to me like you have elements that group rows and we have mappings to them in Firefox, but this cannot me addressed in the HTML spec.
<MarkS> JF: whether or not we have a way of expressing these elements to the API? i would say yes… but don't know if there is a semantic that we can use.
<Suzanne_T> html 4 allowed scope="rowgroup" for a header scope
<MarkS> CS: if it can be done in UiA it can be done for all the elements in HTML
<MarkS> RS: what we have to do is 2 things: any elements that are missing in the table, we need to define an ARIA semantic for
<MarkS> …next we have to decided what the user agent has to do with it. toss it away, or use it.
<MarkS> …there are elements in the HTML 5 spec that don't have native host semantics.
<MarkS> …happy to let the browsers decide what they want to do with rowgroup.
<MarkS> …cynthia if you have a thead, would you expose the scrollable interface control pattern?
<MarkS> CS: good idea, not sure if that is how its implemented.
<MarkS> RS: either way you would want to have the user agent tell the API what it is
<MarkS> …scrollable or not, doesn't really matter from my perspective.
<MarkS> JF: sounds reasonable.
<MarkS> RS: you need to have a look through HTML5 to find elements that don't have host language semantics.
<MarkS> …user agent implementation needs to add support for these
<SteveF> https://dvcs.w3.org/hg/aria-unofficial/raw-file/tip/index.html#recommendations-table
<MarkS> SF: using ARIA in HTML doc. list of all the HTML elements. second column defines which ones have ARIA mapping or not.
<MarkS> …some have multiple elements in them. all elements should be there.
<MarkS> RS: transfer this to HTML5 spec
<MarkS> SF: i've been arguing for this.
<MarkS> …will get it done
<MarkS> JS: should we create a bug?
<MarkS> SF: yes, that is what should be done.
<MarkS> …it shouldn't be a high traffic bug.
<SteveF> https://www.w3.org/Bugs/Public/enter_bug.cgi?product=HTML%20WG&component=HTML5%20spec&priority=P3
<MarkS> RS: I will file the bug.
<MarkS> JS: lets put a TF keyword in there.
<richardschwerdtfeger> https://www.w3.org/Bugs/Public/show_bug.cgi?id=21985
<MarkS> JS: is this for 5.0 or 5.1?
<MarkS> JF: writing an extension spec that has no support from browser vendors isn't very productive.
<MarkS> …i plan on reaching out to the vendors. perhaps we can have a meeting and prepare something to present to them.
<MarkS> …i think this is a 5.1 item
<MarkS> JS: ok, lets bring the Media sub-team back into action.
<chaals> scribe: chaals
<MarkS> https://www.w3.org/Bugs/Public/show_bug.cgi?id=9485
MS: 2 bugs came up in Triage that we want to bring to TF
… 9485: Requiring conformance checkers to refer to WCAG regarding alternative text stuff.
… There is a comment from Martin Leihm, Henri Sivonen says there is no reason for the spec to micromanage the UI of tools.
… Triage felt this could not be tracked, and should be re-assigned to validator product.
scribe: ideas?
Steve: You could file a bug on the validator.
JS: Is this one we want to pursue?
LW: Don't think so.
SF: Don't think so
JS: How do you assess quality heuristically?
LW: Think this bug was filed when the spec was weaker on accessibility.
JS: agree.
<MarkS> https://www.w3.org/Bugs/Public/show_bug.cgi?id=13451
RESOLUTION: close 9485 Rejected.
MS: 13451 - Don't disallow image maps on object.
… refers to an object element containing an image map but the UA cannot / does not render images.
… note in spec says to refer to fallback content instead of alts in the image map.
… proposal was to remove the note. Hixie said that won't solve the issues.
… In processing, where images are disabled the fallback of object gets rendered, which ruins the ability to show the alternative content where it is an image map.
JF: Won't this be solved by the links being the content?
MS: No, this is the plaintext that follows the image.
… so the image gets removed from the tree.
JF: But you can still expose the links that are in the image map
MS: How do you do that if the image is removed?
JF: Whether you see the image or not you can still find the links for the regions. They are somewhat independent.
… the links are there even when the image is gone. How you render that I don't know, but it seem like a list of links.
… this is an issue of presentation vs structure.
… if the images are disabled, the user still needs access to the links.
<MarkS> CMN: ability to use block content was removed from the spec and changed to something else. don't recall exactly how it was meant to work, but it is kind of strange.
<MarkS> …used to be able to build a map that was just some html that had some links in it.
<MarkS> …now you can add bits of html that show up as invisible block content.
<MarkS> …as i recall, you can still use area elements as fallback content and it would work.
<MarkS> …should we be happily going along with the new model, one of the issues I had at the time was that it wasn't properly implemented at the time, so we would need to test, or should be go back to the 4.0 model
<MarkS> …the issue is if we go with HTML5 we need to change all the instructional material. which is going to be difficult.
<MarkS> JF: we need to do some testing.
<MarkS> …+1 chaals on changing the educational piece. its not easy, but it is doable. and we need to capture the significance of that
<Zakim> chaals, you wanted to offer to add this to the list of triage bugs I need to look at...
<scribe> ACTION: chaals to look into bug 13451 [recorded in http://www.w3.org/2013/05/09-html-a11y-minutes.html#action01]
<trackbot> Created ACTION-180 - Look into bug 13451 [on Charles McCathie Nevile - due 2013-05-16].
<MarkS> MS: keep original bug the same, but file a new bug on how the object element will handle images when browser cannot display images or has them disabled.
<MarkS> MS: no more changes to wiki to date
<MarkS> CMN: If you received an email regarding open action items, please reply with a status update.
<MarkS> …i only looked at items that are more than 3 months overdue.
<MarkS> …i changed the date on one and closed some others.
<MarkS> JS: try to reduce the list.
<MarkS> CMN: a lot of these jobs are worth doing. most effective to have the assigned person take a look and decide what state they are in.