Re: New issue: IMG section of HTML5 draft contradicts WCAG 1 & WCAG 2 (draft)

I would have to agree with you that it would make more errors possible.
noalt, or wathever the chosen solution for the situation when an alt text is
not available would be used in a similar way to the font tag with the
wysiwyg meta tag (I know that is also debated)

The idea here is that when used in conjunction with a meta generator tag,
that noalt would not be ambiguous. I would agree that does not precludes bad
copy paste by authors and that those errors must be accounted for. (They
could also copy paste the meta, which would leave us nowhere good)

if alt is not required in the case when it is not available, would it be
possible to force the inclusion of a meta tag that basically says that the
page is generated and that some @alt are not available ?  This would at
least give some information to user agents. It does complicate everything
else though ... I am not sure this is a good idea

@noaltavailable maybe ? it is certainly less confusing... but not a lot
better...


Tough decision indeed....


2008/4/11, Ian Hickson <ian@hixie.ch>:
>
> On Fri, 11 Apr 2008, Benoit Piette wrote:
> >
> > Personally, I think something along the lines of
> >
> > <img src="toto.png" noalt >
> >
> > in pages that have a meta tag saying that the content is generated and
> > all semantic information is not necessarily available to user agents
> > would not be ambiguous in any way.
>
>
> Why would it be any less ambiguous than not having an attribute at all?
>
> I considered "noalt". The problem is that if the concern is validation,
> how do we distinguish cases where the user has forgotten to provide alt=""
> text and has accidentally copy-and-pasted the noalt marker, from cases
> where the user has intentionally marked the image as not having alt text
> available despite needing some?
>
> All adding "noalt" does is increase the number of possible errors -- now
> we have to deal with no alt, noalt at the wrong time, alt with the wrong
> value, and both noalt and alt provided at the same time. With just "alt",
> we only have to worry about no alt and alt with the wrong value. The net
> effect isn't an improvement.
>
>
> --
> Ian Hickson               U+1047E                )\._.,--....,'``.    fL
> http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
> Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'
>