W3C home > Mailing lists > Public > public-html@w3.org > April 2008

Re: [html4all] New issue: IMG section of HTML5 draft contradicts WCAG 1 & WCAG 2 (draft)

From: Charles McCathieNevile <chaals@opera.com>
Date: Fri, 11 Apr 2008 17:16:33 +0200
To: HTML4All <list@html4all.org>
Cc: "Michael(tm) Smith" <mike@w3.org>, "Dan Connolly" <connolly@w3.org>, wai-xtech@w3.org, "HTML WG" <public-html@w3.org>, "Al Gilman" <Alfred.S.Gilman@ieee.org>
Message-ID: <op.t9gcpsjlwxe0ny@widsith.local>

On Fri, 11 Apr 2008 14:07:34 +0200, Philip TAYLOR  
<Philip-and-LeKhanh@royal-tunbridge-wells.org> wrote:

> Charles McCathieNevile wrote:
>
>> 4. It means that all legacy testing systems will have to be rebuilt to
>> ensure that they recognise this magic string as being equivalent to not
>> having any alt attribute.
>
> It is /not/ equivalent : that is the whole point.

Ah. This is where we disagree I think. Fundamentally I don't think there  
is such a case as "it is impossible to provide useful alt" - although  
there are many cases of "I couldn't be bothered" and "we didn't bother  
making our application support that" and "the Author didn't give any  
informatin" and so on, that lead to no useful text (in which category I  
include alt="") being present.

And the case where nothing useful has been provided, in the context of  
getting meaningful content from an author to a user is in fact one of  
those that is most important and easiest (currently) to distinguish. FWIW  
I think the current HTML draft recognises this - in line with the  
reocmmendations from WAI on authoring tools, evaluation mechanisms, and  
dealing with this issue.

>> 1. Because that does not mean anything in most langauges of the world.
>
> It means neither more, nor less, than "Mt Fuji, the peak bathed
> in early evening sunset orange, silhouetted against a background
> of pellucid white cumulus clouds", which most of us (and I suspect
> that includes yourself) would regard as a totally acceptable ALT
> text for an image depicting that scene ...

In the right circumstances that would be a fine text.

If you are suggesting that we should generally accept any human-readable  
text explaining that the author didn't provide anything, then I disagree  
(there has already been a thread on default values in alt, and I don't  
think there is any reasoning that hasn't been exposed).

>> 2. Because you are unlikely to come up with a complete text string that
>> people always emit correctly, and get it implemented through the various
>> tool chains in use, with anything like the same efficiency as working  
>> out what the lack of an alt attribute means.
>
> Why ?  These are /automated/ tools, and it is therefore trivial
> to modify them to emit a pre-agreed string in circumstances
> such as these.

Because you have to identify each such tool and convince them to follow  
the new standard. Given the abject failure so far to convince most HTML  
authoring tool developers to follow the relevant standards, I suspect that  
"build it and they will come" is not a useful answer - having spent  
several years when i worked at W3C being *one* of the people pushing tools  
towards standards compliance my experience is that this is the sort of  
thing that takes a lot longer to get into the tool chain than people just  
doing nothing.

cheers

Chaals

-- 
Charles McCathieNevile  Opera Software, Standards Group
     je parle français -- hablo español -- jeg lærer norsk
http://my.opera.com/chaals   Try Opera 9.5: http://snapshot.opera.com
Received on Friday, 11 April 2008 15:17:41 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 9 May 2012 00:16:14 GMT