W3C home > Mailing lists > Public > public-html@w3.org > August 2007

Re: Baby Steps or Backwards Steps?

From: Lachlan Hunt <lachlan.hunt@lachy.id.au>
Date: Thu, 16 Aug 2007 13:43:04 +1000
Message-ID: <46C3C7C8.3070100@lachy.id.au>
To: "Gregory J. Rosmaita" <oedipus@hicom.net>
CC: HTMLWG <public-html@w3.org>, wai-xtech@w3.org, James Graham <jg307@cam.ac.uk>

Gregory J. Rosmaita wrote:
> james wrote, quote
> For millions of people with flickr accounts I suspect imposing a 
> requirement to provide alt text for each image would quickly drive them 
> to a different service that did not impose such a requirement. Yet the 
> _point_ of flickr is the images. You cannot force people to enter 
> metadata, much less the kind of high-quality metadata needed to explain 
> the meaning of something visually complex like an abstract photograph.
> unquote
> 
> 1. if you want to describe your pictures at your flickr account, you 
>    should be able to do so -- tersely AND in detail; you personally 
>    don't HAVE to add an alt attribute, but if you don't your page 
>    won't validate

IIRC, one of the problems with that approach is that encourages 
authoring tools wanting to output conforming markup to generate useless 
alt text, which is often worse than providing no alt attribute at all.

The spec just trying to make a clear distinction between an empty alt 
attribute and no alt attribute.  In the former case, the empty alt 
indicates that it's decorative and the browser can safely ignore it.  In 
the latter case, the lack of alt attribute indicates that none has been 
provided and the browser can still notify the user that an image is present.

It's a matter of trying to find the right balance between the pragmatic 
and idealistic approaches.  In an ideal world, everyone would provide 
good quality alt text for every image.  In reality, though, that's never 
going to happen and there's little benefit gained by the spec trying to 
enforce idealism.

However, I do agree that the current draft does suggest a little too 
strongly to omit it.  It should more strongly encourage it to be 
provided if at all possible, but I think it should still be optional in 
some cases.

The disadvantage of that approach, however, is that conformance checkers 
will have difficulty distinguishing between conforming and 
non-conforming omission of alt.  But it's really not much worse than the 
present situation.  Conformance checkers have never been able to 
distinguish between conforming use of an empty and non-empty alt attributes.

> [ snip - about Gregory's photo album]
> http://my.opera.com/oedipus/albums

That photo album is a perfect example of extremely poorly generated alt 
text.  All of the images contain the alt text: "perception - photography 
- image interpretation - blindness".  Although it's margially better 
than an empty alt attribute because it gives some indication of there 
being an image, it seems to be worse than no alt attribute, especially 
since it says nothing at all useful about the images and is needlessly 
repeated on every one.  That's the use case the draft is trying to 
address by making alt optional in some cases.

> 4. i have a question for you: is it "needless hyperbole" or just 
>    an "urgent re-statement" of ideas,

No, it was needless hyperbole.  There's no need to write a long rant, 
fighting for accessibility every time something is added to the draft 
which wasn't as well written or thought out as intended.  If you just 
attribute a little bit of good faith to Hixie's intention, and calmly 
point where the spec can be improved, the discussion and outcome will be 
much more productive.

-- 
Lachlan Hunt
http://lachy.id.au/
Received on Thursday, 16 August 2007 03:43:22 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 9 May 2012 00:16:04 GMT