Re: unifying alternate content across embedded content element types

At 01:00 -0500 UTC, on 2007-07-15, Robert Burns wrote:

> On Jul 15, 2007, at 12:13 AM, Sander Tekelenburg wrote:
>> As an author, you should look at the entire document, without the
>> images
>> being loaded, and for each image consider what text would make
>> sense in its
>> place; what text would make you not miss the image, because that
>> text conveys the same as the image. [...]
> It might be helpful if you too could provide some examples of what
> you're saying. For example, in Jon's Fluffy picture example, how
> would you populate the @alt attribute? [...]

I'll try, now that I understand Jon's original point :) I think the problem
is in part with the caption. IMO it would, exactly because it is a context
thing, be entirely appropriate to adjust the caption such that it makes sense
with either equivalent. this might be reasonable:

[context: A page describing my cat, Fluffy, containing one picture]
<img src=cat.jpg alt="Fluffy">
<p>Playing with a ball of yarn</p>

(This is not dissimilar to "<a>click here</a>", which may 'work' in context,
but is useless without context. I think all of us do adjust our text now and
then so that we can provide better link text.)

But as said, it's entirely a contextual thing. If the context is both your
cats Fluffy and Miles, then the above probably won't work, because you may
need to let your audience know which cat you're conveying information about.
So in that case, this might be more appropriate:

<img src=cat.jpg alt="">
<p>Fluffy likes playing with a ball of yarn</p>

After all, if (*if*)) the entire message is already conveyed through the
caption, then that fact means the is image purely decorative.

This is why I usually try to avoid providing examples of proper @alt text. It
quickly gets very subtle. If there is the slightest bit of confusion about
the context, what makes the alt text 'right' may be misunderstood and
incorrectly applied to a different context. It's better to work with real
life cases. (But a lot more work...)


> Also to push this question, I've been trying to get across, why not
> just always put that equivalent content in a document fragment
> targeted by the @longdesc attribute and leave @alt off completely.

I suppose that might make sense, in certain contexts. Maybe like this:

[context: About my cat, Fluffy]
<img src=cat.jpg alt="" longdesc="#id">
<p id="id">Long, richly marked up, description of how Fluffy likes playing
with balls of yarn</p>

> After all, the @longdesc could be short or long? It can be markup
> rich or not. It can be on the same page or on another page. Through
> some CSS and DOM wizardry it can be made available in ways that
> current UAs do not make available for any other fallback (at least by
> default).

CSS cannot be relied upon to be available, so access to content should not be
made reliant on it. Same for "DOM wizardry", assuming you're talking about

Sander Tekelenburg
The Web Repair Initiative: <>

Received on Tuesday, 17 July 2007 03:46:13 UTC