Re: Long descriptions spec - a basic idea

On 09/20/2012 10:34 PM, Leif Halvard Silli wrote:
> Hello A11Y TF. This new direction opens up a possibility for two
> conformance levels w.r.t longdesc use. And, inspired by Janina's
> identification of longdesc as fully functional in certain restricted
> corners of the web, I would therefore like to propose the following way
> forward:

If you will permit me to make an observation: focusing on conformance 
may be putting the cart before the horse.  I'd recommend starting from 
deciding what problem you want to solve.

Roy Fielding recently presented[1] one view on what longdesc is meant to be:

   enhancing the public record so that it better fits the needs of
   those with assistive technologies, without changing the visual
   representation for those who don't use AT.

His view may or may not represent what the HTML WG or that A11y Task 
Force intends to pursue, but the prescribed behavior in the Instate 
longdesc change proposal goes well beyond what Roy described, and does 
so in a way that has not gotten a warm reception by those that develop a 
number of user agents.

Define something that is likely to be universally implemented by next 
year, or define something that is explicitly NOT intended to be 
universally implemented.  Or do both, perhaps on different schedules.

Depending on which path you chose, different levels of validator 
warnings may be appropriate, or no warning at all.

Things with a universal appeal and are likely to be nearly universally 
implemented in the near term stand a strong chance of becoming 
integrated into the HTML5 specification.

In yesterday's A11y telecon, Janina used the normally emotionally 
charged term "ghetto", but did so to make a point: there actually are 
advantages to a "ghetto" specification.  As this was near the end of the 
telecon she did not have a full opportunity to explain what she meant by 
that so I may be misunderstanding, but there is nothing wrong with 
intentionally pursuing a "ghetto" specification, nor is there anything 
wrong with intentionally pursuing integration into the HTML5 
specification.  You just may end up specifying different things 
depending on what problem you want to solve.

- Sam Ruby

[1] http://lists.w3.org/Archives/Public/public-html/2012Sep/0309.html

Received on Friday, 21 September 2012 09:23:05 UTC