Re: fear of "invisible metadata"

On Jun 18, 2007, at 6:38 PM, Chris Adams wrote:

> I didn't quite mean it like that: what I should have said is that  
> as long as a page is designed to be semantically correct then minor  
> errors given by a conformance checker such as missing alt tags can  
> be ignored.
>
> 100% validity is usually a good thing but can cause unneeded headaches

I think it would be better to design the document conformance  
criteria so that you can reasonably achieve 100% conformance without  
headaches. There's a big difference between having a tool where you  
can fix 100% of the errors it reports, and ones where you have to  
mentally keep track of some number of errors that are pointless and  
should be ignored. We have the opportunity now to make the set of  
errors reported by conformance checkers a good set that will help  
authors.

Regards,
Maciej

>
>
> On 6/18/07, Maciej Stachowiak <mjs@apple.com > wrote:
>
> On Jun 18, 2007, at 6:16 PM, Chris Adams wrote:
>
>> This also goes to show that conformance checkers are tools that,  
>> while useful, should not be the
>> be-all-and-end-all of error checking.
>>
>> If a page is designed to be semantically valid then a conformance  
>> checker is not really needed.
>
> That doesn't make sense to me. It's like saying if you design a  
> program to be correct then you don't need testing.
>
>  - Maciej
>
>>
>>
>> On 6/18/07, Maciej Stachowiak < mjs@apple.com> wrote:
>>
>>
>> On Jun 18, 2007, at 10:24 AM, Gregory J. Rosmaita wrote:
>>
>> >
>> > Lachlan Hunt wrote, quote:
>> >> No, even if the summary attribute were added to HTML5, it
>> >> certainly shouldn't be required.
>> > unquote
>> >
>> > why not?  CAPTION is akin to ALT text - it provides a terse
>> > description
>> > of the object that cannot be visually perceived; the summary  
>> attribute
>> > itself serves the same purpose as LONGDESC (which provides a  
>> detailed
>> > description, orientational material, etc.)
>>
>> ALT should not be required either. It leads to pointless alt="" on
>> images that have no reasonable text equivalent, just to satisfy
>> conformance checkers. And that is actively harmful, because AT can't
>> tell the difference between a semantically null image and a
>> semantically meaningful image with no text alternative.
>>
>> Regards,
>> Maciej
>>
>>
>>
>>
>>
>> -- 
>> Chris@tuesdaybegins.com
>> http://www.tuesdaybegins.com
>
>
>
>
> -- 
> Chris@tuesdaybegins.com
> http://www.tuesdaybegins.com

Received on Tuesday, 19 June 2007 02:03:03 UTC