RE: CP, ISSUE-30: Link longdesc to role of img [Was: hypothetical question on longdesc]

>> Have a look at a very typical usage (where I, alas, can't watch the video): http://www.bbc.co.uk/news/world-europe-17457845 .  That image is just a better initial place-holder than the first frame - which hasn't been fetched yet, when the page loads.

Good example. The image there is not in fact a frame from the video at all, although similar in theme and clearly from the same event.

I think we have clearly established that at least some of the time there may be an image displayed which is distinct from the video content, and I don't see anyone arguing to remove the possibility of this distinction.

Since we also all appear to agree that the idea here is to provide in text all of the information available visually, then there is a clear use case to allow the means to describe both parts *when they are distinct*, What we seem to have a disagreement on is whether the information needs to be recorded in such a way that the user is aware where the dividing line is. I personally believe that when the image is semantically distinct, that distinction is present in the visual mode, and should therefore be present in the text alternative.

In order to satisfy the use case though, I don't think it really matters if both parts are co-located in a single resource at the end of a single longdesc link, or in two resources at the end of two distinct links. What I do think matters however is that when the image description is not semantically a subset of the video description, there is a clear dividing line so that the browser (or AT) can provide the right amount of information to the user. Two clearly labeled links naturally provides that division, a single link does not, and therefore implies a mechanism would need to be in place in the resource pointed to. I'm happy either way provided the use case is met.

-----Original Message-----
From: David Singer [mailto:singer@apple.com] 
Sent: 21 March 2012 21:16
To: janina@rednote.net
Cc: Sean Hayes; John Foliot; 'Silvia Pfeiffer'; "'xn--mlform-iua@målform.no'"; rubys@intertwingly.net; laura.lee.carlson@gmail.com; mjs@apple.com; Paul Cotton; public-html-a11y@w3.org; public-html@w3.org
Subject: Re: CP, ISSUE-30: Link longdesc to role of img [Was: hypothetical question on longdesc]


On Mar 21, 2012, at 13:59 , janina@rednote.net wrote:

> David Singer writes:
>> Or we simply say the obvious "If the image is not a representative frame of the video, or conveys information in addition to the content of the video, then a description of that information must also be included with the description(s) of the video that are supplied for accessibility (e.g. alt, longdesc, transcript, etc.)."
>> 
> A description of the video is not the same thing as a description of a
> rich image that is published to stand for the video.
> 
> 
> It's not a video until the video is running.

It's not a video when it is paused for any reason, at any time.  And when it is paused, the image is *unpredictable* because it's paused at an indeterminate point.  That's a characteristic of paused videos, they look like images when time is not advancing. Any initial paused state is just one of those paused states.

> Until then it's an icon for
> the video, or a poster for the video, or the magic doo-hickie that tells
> you something about the video (or not).

Actually, it's a frame of the video. You left out the 90%+ case.  Have a look at a very typical usage (where I, alas, can't watch the video): <http://www.bbc.co.uk/news/world-europe-17457845>.  That image is just a better initial place-holder than the first frame - which hasn't been fetched yet, when the page loads.

There is some history here.  QuickTime movies have a field in their header "poster time" - the time of the image to show as the initial image, when the first frame was unsuitable (e.g. because it's black, or the typical trailer green-frame). Clearly in the HTML case we were looking for something to show while loading the resource, so indexing into the resource was not suitable.

Occasionally, in QuickTime, people wanted to author custom frames (to be clearer, for example, or juxtapose some elements that didn't come together in any single frame of the video); for them, we provided custom poster frames as well.

> Two different things, semantically disparate. Therefore, they obviously
> need semantically distinct description.

This is an assertion that does not reflect that the media element is a single element, which needs a description.

> This should be obvious,

It ought to be obvious that a paused video looks like an image.  But apparently it's not.

It ought to be obvious that we should have, long ago, had non-timed provision for the video itself.  Short text? No. Description? No. Link to a transcript? Um, not yet.

Instead, we debate how to arrange the deck-chairs. If we had already dealt with the major problems, spending time on this as a tidy-up might be warranted; even then, all it takes is saying that the image must be included in the alternative(s) to the element (description(s), transcript, etc.).

David Singer
Multimedia and Software Standards, Apple Inc.

Received on Wednesday, 21 March 2012 22:09:25 UTC