Re: Image description extension review

On Sat, 17 Nov 2012 13:03:59 +0100, David MacDonald  
<david100@sympatico.ca> wrote:

>
> Hi Chaals
>
>
> I put up a test page:
>
> http://www.davidmacd.com/test/longdesc.html
>
>
> -JAWS 13 doesn’t support longdesc on the same page in IE9,

So I should clarify what we mean by that.

In your results, JAWS opens a page at the point where the longdesc is,  
right? Then you close it and you're back at the image?

That's what happens in Opera and iCab when longdesc uses an internal link.  
It possibly isn't ideal, but it is consistent so it should not surprise  
users. And it gets them to the longdesc.

>
> - JAWS 13 and FF14, neither example opens.
> -JAWS 13 in Chrome 23.0.1271.64m: Neither example opens.

That's strange. In the test case I had I made the page long - bigger than  
a screenful, with a lot of separation between the image and its  
description so it was clear whether I had arrived at the reference point.

>
>> I think there is a lot of advice to use null alt text for decorative  
>> images. But there is reason such images should not have a longdesc.
>
>
> I think we have our wires crossed. I was commenting on your first  
> example:
>
>
> <!-- pointing to something internal to the page -->
> <img src="http://example.com/image" alt="" title="photo"  
> longdesc="#photo1">
>
>
> My concern is that the example is using null alt text on a  
> non-decorative image.

We don't have our wires crossed. That example was imagined as a decorative  
image that is nevertheless described. I think it would be good to clarify  
that.

cheers

>
> There is currently no precedence for that, I would suggest. Current  
> practice is to provide a short description AND a long description for  
> >images requiring long description. This is what is done in example H45  
> in WCAG.
>
>
> Cheers
>
> David MacDonald
>
>
> CanAdapt Solutions Inc.
>
>  "Enabling the Web"
>
> www.Can-Adapt.com
>
>
> From: Charles McCathie Nevile [mailto:chaals@yandex-team.ru]Sent:  
> November-17-12 3:12 AM
> To: public-html-a11y@w3.org; 'Léonie Watson'; David MacDonald
> Subject: Re: Image description extension review
>
>
> On Fri, 16 Nov 2012 22:38:57 +0100, David MacDonald  
> <david100@sympatico.ca> wrote:
>
> Hi Chaals
>
>
> On the call you had mentioned that longdesc pointing to an anchor on the  
> same page is included in the HTML4 spec. I’m sorry but I’m >unable to  
> find any reference or example in HTML 4 of Longdesc referring to  
> information (an anchor id) on the same page.
>
>
> It's a URI - a link. That can be relative or absolute, including  
> pointing to something else in the same page. It's common practice to  
> have links to somewhere else in >a page. There is very little in the  
> HTML 4 spec about the precise mechanics of the link. But my testing  
> showed that browsers which implemented longdesc >assumed it should work  
> for all URIs including links within the page.
>
>
>>
>> Nor can I find any recommended advice to use null alt text (alt=“”) on  
>> an image that has a longdesc...
>
>
> I think there is a lot of advice to use null alt text for decorative  
> images. But there is reason such images should not have a longdesc.
>
>
>>
>> Nor are there any example techniques in WCAG showing this use of  
>> longdesc or of null alt text.
>>
>> http://www.w3.org/TR/2012/NOTE-WCAG20-TECHS-20120103/H45
>
>
> If I get time I can offer some.
>
>
>>
>> In 10 years on the WCAG team and weekly calls, I’ve never heard this  
>> use case of Longdesc discussed, nor in 14 years in the field >>have I  
>> ever seen it done.
>
>
> OK. I've done it, but I can't think of any examples I have available in  
> public.
>
>
>>
>> I’m fine with the longdesc extension... but I think we either need to  
>> demonstrate that this use case exists in the real world and is in  
>> >>HTML4,
>
>
> The use case of longdesc within a page? There have been arguments raised  
> many times that longdesc is deficient because it doesn't allow for this  
> use case >(which is simply untrue). If the people who are opposed to  
> longdesc think this use case is important, that's good enough for me.
>
>
> As for the use case of long description fora  purely decorative image, I  
> think that's easy to establish. And it is already accepted that such  
> images should have null >alt. The interesting point that perhaps we  
> should make in the spec is to consider this case when ensuring  
> discoverability.
>
>
> cheers
>
>
> Chaals
>
>
>>
>> or we have to acknowledge that it’s a new behaviour for Longdesc.  I  
>> also don’t think we should introduce null alt text for anything but  
>> >>decorative images... so far automated checkers rightly flag null alt  
>> text to ensure it’s purely decorative... in the future we may need to  
>> >>expand the use of null alt text for cases when there is another type  
>> of text alternative such as figcaption, aria-describedby, longdesc,  
>> >>etc.. but I don’t think such a precedence currently exists.
>>
>>
>> HTML4 references
>>
>> http://www.w3.org/TR/REC-html40/struct/objects.html#adef-longdesc-IMG
>>
>> http://www.w3.org/TR/REC-html40/struct/objects.html#adef-alt
>>
>> http://www.w3.org/TR/REC-html40/appendix/notes.html#accessibility
>>
>>
>>
>>
>> Cheers
>>
>> David MacDonald
>>
>>
>> CanAdapt Solutions Inc.
>>
>>  "Enabling the Web"
>>
>> www.Can-Adapt.com
>>
>>
>> From: Charles McCathie Nevile [mailto:chaals@yandex-team.ru]Sent:  
>> November-15-12 9:08 PM
>> To: public-html-a11y@w3.org; Léonie Watson
>> Subject: Re: Image description extension review
>>
>>
>> On Fri, 09 Nov 2012 21:41:55 +0100, Léonie Watson <tink@tink.co.uk>  
>> wrote:
>>
>> Hello,
>>
>>
>> A few thoughts, mostly editorial.
>>
>>
>> Mostly I've adopted or adapted your suggestions.  
>>
>>>
>>> Introduction...
>>>
>>>
>>> The code examples all have null alt attributes. Could they be more  
>>> practical examples?
>>
>>
>> Actually only one had a null value, the others had something but in  
>> parentheses so I removed them and cleaned them a tiny bit. I'd be happy  
>> for people to >>propose better examples...
>>
>>
>>>
>>> Also wonder whether some best practice examples of longer descriptions  
>>> would be helpful? The extension may not be the right >>>place for  
>>> them, but a separate note might do the trick.
>>
>>
>> Steve also suggested this. For now I put it in a note as out of scope,  
>> with a question about whether it would make sense to link to some other  
>> document...
>>
>>
>> Thanks for the review. I updated the document:  
>> http://dvcs.w3.org/hg/html-proposals/raw-file/default/longdesc1/longdesc.html
>>
>>
>> cheers
>>
>>
>> Chaals
>>
>>
>> --
>> Charles McCathie Nevile - Consultant (web standards) CTO Office, Yandex
>> chaals@yandex-team.ru Find more at http://yandex.com
>
>
>
>
>> --
>
> Charles McCathie Nevile - Consultant (web standards) CTO Office, Yandex
> chaals@yandex-team.ru Find more at http://yandex.com



-- 
Charles McCathie Nevile - Consultant (web standards) CTO Office, Yandex
chaals@yandex-team.ru Find more at http://yandex.com