Re: From the HTML-WG about aria-hidden

Janina Sajka, Thu, 16 Feb 2012 22:30:04 -0500:
> Leif Halvard Silli writes:
>> Janina Sajka, Thu, 16 Feb 2012 21:50:43 -0500:
>>> It seems to me a lot of complication is being concocted to serve a
>>> function far more simply served by longdesc. I still suggest a shave
>>> with Occam's Razor.
>> I am not certain that that is a good point: Even without @hidden, one 
>> can do the same thing - but with more work for the author.
> Well, if the image is ever to be reused on some other page, it's much
> more work to copy all the text and markup than to simply supply a URI
> via an attribute.

I agree. @longdesc is more compact, in that sense.

However, to use <a href hidden> is also quite compact compared with the 
alternative of using <a href aria-hidden=true> plus CSS: If the CSS 
gets out of sync, then the link becomes visible. 

> And, it then becomes a real nightmare to keep such
> text updated correctly.

I in principle agree. And I wonder if that point about authoring, has 
been sufficiently strongly made in the change proposal for @longdesc ...
> That's the point of Occam's Razor--the simplest solutions tend to be the
> correct ones. I just don't see a justification for the complication--let
> alone the question of whether it would actually work as intended (which
> I don't believe it would).

Then one could counter that the outlook for support of @longdesc in UAs 
- and even across AT - has for a long time looked gloomy. 

I gather that you have faith in ARIA as such, but not in this detail. 
It would be interesting to know why.

>> Also, @longdesc only works for <img> - while @aria-describedby and 
>> @hidden can be used all over.
> Well, that's a reasonable goal, but it's also very much in the power of
> the spec creators to expand where any attrib might be used, e.g. in
> Issue-203 we propose it for the media elements.

I saw a message from you which linked to a proposal for @transcript 
attribute - is that what you mean? Or has the 203 proposal n not been 
written yet?

> Bottom line, describedby is the wrong ARIA solution. There will be a
> future ARIA equivalent of longdesc--but it won't be describedby.
> Meanwhile, attempting to shoehorn all of this into describedby is
> putting its legitimate functions at risk, and that just isn't
> acceptable.

How does the shoehorning analogy fit when it is *ARIA* which says that 
native semantics *should* be revealed to the user?

Also, the future ARIA equivalent of longdesc could very well point to a 
@hidden section, could it not? Well - of course - that is for ARIA next 
version to decide. But in theory it could.
Leif Halvard Silli

Received on Friday, 17 February 2012 04:23:53 UTC