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

Re: [html4all] Request for review of alt and alt value for authoring or publishing tools

From: Charles McCathieNevile <chaals@opera.com>
Date: Wed, 16 Apr 2008 11:25:17 +0200
To: HTML4All <list@html4all.org>, "Ian Hickson" <ian@hixie.ch>, "Joshue O Connor" <joshue.oconnor@cfit.ie>
Cc: "'Tomas Caspers'" <tomas@tomascaspers.de>, wai-xtech@w3.org, wai-liaison@w3.org, public-html@w3.org
Message-ID: <op.t9o5sfodwxe0ny@widsith.local>

On Wed, 16 Apr 2008 00:21:51 +0200, Anne van Kesteren <annevk@opera.com>  
wrote:

> On Tue, 15 Apr 2008 23:56:36 +0200, Ian Hickson <ian@hixie.ch> wrote:
>> I don't understand. Why can't whatever behaviour will happen for
>> alt="magic vlaue" also happen when the alt="" attribute isn't present?  
>> In both cases we're talking about future tools, and in both cases we're
>> presumably talking about the same behaviour. I agree that the users of
>> legacy tools are screwed either way (magic value or missing attribute,
>> both will result in a poor user experience for these images in today's
>> UAs, though the missing alt case at least typically has user prefs so
>> that the user can tweak those cases, another reason why I personally
>> prefer simply omitting the alt="" attribute rather than introducing
>> keywords).
>
> I think the "opposing viewpoint" is more about today's behavior and
> content than how we can have it in the future. The assumptions seem to be
> that:
>
>   * If the alt attribute is specified it is likely to be correct.
>   * If the alt attribute is omitted the more typical case is that the  
> image does not convey information.

(Leaving aside the question of validity and focusing on the effect on  
users) I am with Ian here. Adding magic values is likely to mess up  
existing workflows from user agents through to authoring tools, system  
evaluation tools (the sort of thing that the people who currently really  
care about getting tehir HTML right actually use), and even websites  
explaining how to write good web pages. Leaving out the alt attribute  
where you don't know anything about what would be a good value (whether  
you are a second-rate tool that never asked, or a lazy or second-rate  
author that never bothere to think about the answer - and I really do mean  
that to sound at least as judgemental as it does) is the simplest approach  
to allowing those who are doing a decent job to improve the web overall.

If the alt attribute is specified it is somewhat likely to be correct. If  
it is ommitted, it is almost certain that this is an error - there is  
*some* useful value (which may be alt="") for every example I have *ever*  
seen, but where that is not made clear, machines guessing it are more  
likely to get it wrong (i.e. a machine heuristic is more likely to be  
harmful than useful).

> Joshue also made the point that AT software skips <img src=...> today
> where they would not skip <img alt=...> today.

Where there is not alt attribute, AT software will often try to do  
something useful with whatever is left (but mostly fail - it's beyond the  
state of AT today, partly as a result of the way people use the Web).

Where there is an alt attribute, the AT will present that. Increasingly  
(for some value of increasingly - AT gets tweaked and adjusted far faster  
than browsers, so it may go up and down depending on the current 3-month  
perception of the most useful way to read the existing Web, but the trend  
seems clear) alt="" is recognised as being *probably* deliberately  
included for images where the best thing is not to say anything.

> I think your assumption is that whether the alt attribute is specified or
> not does not affect the likelyhood of it being correct. (As in, <img
> src=... alt=""> for an image that needs alternate text and <img src=...>
> for an image that doesn't are about as likely to occur.)
>
>
> (This is probably an oversimplification and I'd love for people to make  
> it more clear where they are coming from with this.)

I hope that clarifies where I am coming from at least :)


-- 
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 Wednesday, 16 April 2008 09:26:20 UTC

This archive was generated by hypermail 2.3.1 : Monday, 29 September 2014 09:38:54 UTC