Re: Baby Steps or Backwards Steps?

Hi James,
On Aug 16, 2007, at 4:04 AM, James Graham wrote:

>
> Gregory J. Rosmaita wrote:
>
>> 3. how does leaving alt out entirely when an image is not  
>> "purely    decorative" "better" serve someone merely by indicating  
>> the    presence of an image?
>
> The question you have to ask is better than what alternative?  
> Obviously a missing alt attribute is not better than the  
> alternative of well written alt text, and nobody has argued it is  
> (as an aside, the draft goes in to a great deal of detail on how to  
> write good alt text, far more detail than HTML 4 and far more than  
> it goes into for any other attribute, presumably reflecting the  
> great importance that the editor places on well written alt text as  
> an accessibility aid). However a casual glance at the web will  
> indicate that many sites do not have well written alt text. Even  
> where the alt text is present it is often unhelpful either  
> duplicating content that is in the main text (e.g. flickr),  
> containing text that is not an alternative to the image (your photo  
> gallery) or containing blank values for non-decorative images. Is  
> the user really better off with the unhelpful alt text than with  
> the information that an image is missing and whatever attempt the  
> browser can make at extracting metadata from the file itself (e.g.  
> by reading the EXIF headers in a JPEG file)? My feeling is they are  
> not. Also, maintaining the distinction between alt="" meaning "This  
> is a decorative image" and menaing "I wanted to validate but didn't  
> think of alternative text" should prevent non-graphical browsers  
> from having to treat all instances of alt="" as signifying  
> potentially-important-but-inaccessible images.
>

There's definitely a use-case/problem-statement to discuss here.  
Among those problems are that the requirement of @alt has detrimental  
effect along with benefits No one is disputing that. And you raise  
many good points about it (especially the extraction of file metadata  
that I think should be a part of our UA conformance criteria).   
However, its not just the problem that alt="" can be either  
decorative or I'm just trying to get my document to validate. There's  
a parallel problem in distinguishing the entire omission of alt for,  
I'm omitting this because the HTML5 recommendation told me too and  
I'm omitting this because I'm didn't bother to enter it or validate  
my document.  Providing keywords for @alt or a complimentary  
attribute allows UAs and users to distinguish between the five or six  
various scenarios outlined in the draft.

Take care,
Rob

Received on Thursday, 16 August 2007 10:42:35 UTC