Re: heuristic tests not in guidelines

There are probably a lot of suggestions folks could make for improving
Bobby.  My recommendation is that ALT="" should be illegal inside <A
HREF...

Please check out:
http://www.dors.state.md.us/test.html

A totally inaccessible (but simple) page that flys through the W3C and
Bobby validators with no problems.  Even the verbose "Accessibility
Recommendations" and "Accessibility Questions" and "Accessibility Tips"
don't address what is the real problem with this page.

Please note that I haven't touched this page in a while.  The LONGDESC
is implemented incorrectly (not that the W3C validator complains).

Have other people designed pages to illustrate Bobby's short comings? 
The "Known Bugs" of the Bobby FAQ
(http://www.cast.org/bobby/bobbyfaq.html#sec4) is still blank...

Bruce Bailey


Al Gilman wrote:
> 
> Here is one that just occurred to me.
> 
> Suppose a Bobby-like screening tool were to check for ALT=SRC.
> 
> This is most often an ALT text that a tool has generated, and
> again in the main the author could have done better if she thought
> about it for a minute.
> 
> I was thinking that pages where all ALTs that were present were
> of this form would be in my screening bucket of pages we should
> try the text-equiv techniques on.
> 
> This case is interesting to me, because it seems appropriate to
> me for Bobby to emit a warning on detecting ALT=SRC, but I don't
> see this as something that rises to "guidelines" level of
> significance.  I don't see the Page Author Recommendation
> forbidding this equality.  Particularly when the SRC is a simple
> relative URL, so it might be an OK ALT after all.
> 
> This puts into question the idea that Bobby should check "the
> guidelines, and nothing but the guidelines."
> 
> What do people think of this?
> 
> Al

Received on Tuesday, 27 October 1998 17:05:27 UTC