fallback<.div>
10:15:45 good on yer Gergory!
10:15:51 thanks
10:16:27 the name of the wiki page is a relic of issues past -- it should be "Long and Short Descriptors for Static Images" or some such
10:18:38 to quote myself from the page:
10:18:40 What is needed, therefore, is a normative list of recommended/expected actions that allow multi-modal interaction with the long description. Treating LONGDESC as HREF isn't the only means of exposing the content of the long description page; the contents -- or the main portion thereof -- could be rendered inline instead of the image or in an IFrame (which has its own accessibility issues) or any other number of means of exposure. The key is that the UA should
10:18:40 * expose in new browser instance
10:18:40 * expose in new browser tab
10:18:41 * expose inline (insert content as object)
10:18:43 * expose inline through the use of IFrame
10:18:45 * expose the contents of the longdesc document in a side-bar,
10:18:47 aligned with the image it describes
10:18:49 and there are many other options, provided a user knows what to do when encountering a long description, then it matters not what assisstive technology she is using, for there is an expected action in the case of browser x for exposing LONGDESC
10:20:01 GJR notes that some of the best examples of LONGDESC are in CSS 2.0 -- 46 longdescs in all!
10:20:18 Shane: "longdesc" attribute is available even when src does not fail.
10:20:54 right: it isn't necessariliy an either/or proposition -- some user groups need guidance through an image
10:20:55 Steven: "longdesc" is the content of element, we don't need "longdesc" attr.
10:22:00 http://www.w3.org/TR/xhtml2/mod-embedding.html#s_embeddingmodule
10:22:06 GJR: point of reference -- are we discussing M12N section 5.7 "Image Module"
10:22:13 Shane: once resource is obtained, the rest of content is not processed.
10:22:23 Steven: a question: not processed or not presented?.
10:22:35 we say: If accessing the remote resource fails, for whatever reason (network unavailable, no resource available at the URI given, inability of the user agent to process the type of resource) or an associated ismap attribute fails, the content of the element must be processed instead.
10:23:44 http://htmlwg.mn.aptest.com/htmlwg/xhtml-m12n-2/abstraction.html#dt_ContentTypes
10:24:15 Shane: I did not look into public version.
10:25:52 Shane: Issue persists. What happens content nested after a document is resolved?
10:26:00 I think that the alternate text is in the DOM
10:26:11 ... I think we can define the behaviour with CSS
10:26:26 .... in fact I'm sure that Jonny Axellsson once demonstrated that
10:26:33 it should be in the DOM so it can be reused/accessed by assistive technology/regular users
10:26:33 s/ll/l/
10:27:12 ACTION: Steven to add text to embedding module saying how it works
10:29:27 Rich says we need some events that say what happens
10:29:35 ... load when a src works I think
10:29:47 s/load/'load'/
10:30:37 Steven: What should you get if the src fails though?
10:31:11 GJR: need to ensure that if user needs side-by-side image and descriptor to process the image (limited viewport, congnative issues, etc.) it is readily and easily available
10:32:01 Steven: reasons for src failing: (1) network down, (2) 404 or similar.
10:34:11 Steven: reviewed error code 406 for HTTP.
10:35:05 Roland: you did not get any event in image load finished.
10:36:15 Roland: DOM3 have anything different.
10:37:33 s/have/don't have/
10:38:03 Steven: DOM went to WebAPI, Rich.
10:39:01 Stenven: DOM3 said images are loaded before you get load event.
10:39:10 "The DOM Implementation finishes loading the resource (such as the document) and any dependent resources (such as images, style sheets, or scripts). "
10:39:18 http://www.w3.org/TR/DOM-Level-3-Events/events.html
10:40:17 Event "error": A resource failed to load, or has been loaded but cannot be interpreted according to its semantics such as an invalid image, a script execution error, or non-well-formed XML.
10:41:43 Roland: Let's see new charter for WebAPI to check DOM3s...
10:42:05 http://www.w3.org/2008/02/06-webapi-minutes.html
10:42:09 minutes of a recent call
10:42:47 Zakim: leaving. As of this point the attendees were Carmelo, Andrew_Emmons, anne, shepazu
10:45:36 Shane: DOM3 Events use qnames?
10:46:53 Roland: no search result for qname.
10:47:01 Steven: BUt they have namespaced events.
10:47:10 s/BU/Bu/
10:48:49 Roland and Steven: We don't need "longdesc" attr. Stenven will add clarifying text for embedded context. (already ACTION for Steven)
10:49:07 s/Stenven/Steven/
10:49:52 Roland: alternative to src.
10:50:13 Steven: Added (3) images switched off, at flipchart.
10:51:13 .. after (2) 404/406 or similar.
10:54:52 In css, you say img[src] {content: attr(src)}
10:55:29 so in principle you can switch between displaying the image and the content
10:55:51 GJR: that "principle" needs to be explicitly stated
10:55:52 (but the CSS doesn't give you control over @srctype)
10:56:28 In any case the alternate content is in the dom, so is available for use as necessary
10:56:45 amen
10:57:26 Steven: poping stack for a few levels, back to "longdesc".
10:57:56 Roland: how we can style both?
10:58:58 Steven: in CSS, content: means replacement with external resource. (Stven went to flipchart to argue)
10:59:10 s/Stven/Steven/
11:02:34 Roland: CSS cannot do fallback at failure.
11:03:06 Steven: *[src]:error, you can invent pseudo-class in that case.
11:03:40 Steven: We don't know it is allowed, but let's assume it allowed...
11:04:13 Steven: *[src]:before
11:04:27 Steven: {content: attr(src)}
11:04:41 Steven: *[src]{display:none}
11:04:55 Steven: body.nosrc *[src] {display:blank}
11:05:07 Steven: body.nosrc *[src]:before
11:05:14 Steven: {content: ""}
11:05:26 Steven: *[src]:error
11:06:18 steven, will that {content: "foo";} make it into the DOM? currently, CSS-generated text isn't in the DOM and isn't accessible to assistive tech
11:06:19 Steven described an error case with src="foo.xdiv" srctype="vide/xdiv"
11:09:19 UAAG (user agent accessibility guidelines WG) is trying to address CSS- and script-generated text, UAAG2 has a proposed requirement that ALL text, no matter what its source, must be made available via the DOM or directly to an accessibility API (such as MSAA, IAccessible2, ATK/AT-SPI, etc.
11:12:33 Roland: we will talk XML Event 2 after lunch break.
11:12:56 s/Event 2/Events 2/
11:13:54 when will we resume? how much time are you breaking for lunch?
11:14:45 s/display:blank/display:block/
11:15:32 Gregory, I expect lunch will be an hour
11:16:43 thanks - now if only i had a left-over canoli (sp?) to eat for breakfast...
11:16:52 Gregory, wrt your content: question
11:17:03 yeah
11:17:10 I didn't put any real content there
11:17:16 in one case it is the embedded image
11:17:40 in the other case, the embedded image is overwritten with 'nothing' (empty string)
11:17:46 the dom remains the same in both cases
11:17:55 (I think)
11:18:18 and the DOM contains the URL of the image, and the alternate content, both of which reside in the DOM without change
11:18:33 do you think the UAAG2 req unreasonable?
11:18:40 and so is accessible to any software that wants to make use of it
11:18:47 ok
11:19:14 I have no problem with the UAAG2 requirement
11:19:26 that's certainly good to hear
11:19:40 In fact it is a pain when you copy a numbered list, and don't get the numbers in the copy buffer
11:19:40 the tricky bit is wording it correctly
11:19:55 because of the generated content problem
11:20:26 Rich: Guideline for browser for markup?
11:20:30 if CSS is used to control list styling, one doesn't get that info from an assistive tech - it uses the "dumb" nesting default algorithm
11:20:45 Shane: we provided default CSS for XHMTL2.
11:20:52