Re: Mandatory and Important

Hi Dave,

On Aug 22, 2008, at 2:05 PM, Dave Singer wrote:

>> Hi Dave,
>>
>> I think it is not a difficult issue. HTML5 can simply say:
>> * the IMG element MUST include an alt attribute
>> * authors MUST include suitable alt text for each image embedded  
>> with the IMG element
>> * authors SHOULD follow WCAG guidelines in composing suitable alt  
>> text
>> * authoring tools SHOULD follow ATAG in assisting authors providing  
>> suitable alt text and MAY automatically generate default alt text  
>> in cases where it is possible (e.g., the replacement of an image of  
>> richly styled text by plain text)
>> * authoring tools MUST NOT add any text that is a placeholder for  
>> alt text (e.g., "this is an image")
>>
>> I don't see the problem then. We have provided suitable guidance to  
>> authoring tools and authors.
>
> As provided, the guidance is fine.  But thisdoesn't seem to address  
> the question that was central to starting the debate:  what to do  
> when alt text is not available at the time of HTML generation?  My  
> perception is that quite a few people believe or hope that this  
> situation doesn't arise, but it does, and it's currently 'polluting'  
> the web;  your second bullet is not always achievable.  I'm unclear  
> as to what you believe should happen in this case;  I assume you're  
> as unhappy as I am with alt="", missing alt, or alt="useless filler  
> text".

I really don't see that as a central question for this WG (other than  
how it is addressed it what I just wrote). From what I just wrote the  
answer is, If the suitable alt text is unavailable the authoring tool  
should make sure the alt attribute is alt=''. Similarly, the authoring  
tool (in the case of Flickr) might add role='photo'. The dilemma is  
solved (at least as far as we HTML5 spec writers are concerned).

>> The alt attribute is always required (though sometimes has a null  
>> value). Authors and authoring tools know their requirements and  
>> recommendations and know which authorities to reference for more  
>> details.
>>
>> As for the curly braces syntax that need is addressed better by the  
>> use of the role attribute and Gregory's recent proposal to use role  
>> keywords to convey the various roles of embedded media[1]. Using  
>> role instead of the curly braces allows localization to occur  
>> through the UA (rather than resorting to HTTP content negotiation  
>> or another mechanism cumbersome to author). It also allows the  
>> conventions among authors, authoring tools, users and assistive  
>> technology to be well defined (which using arbitrary curly braced  
>> content does not).
>
> The role attribute is a possibility, and I see what you are saying  
> also.  One of the reasons I thought that the keywords in {} should  
> ideally come from a well-defined set was to allow  
> internationalization or interpretation by UAs, and not just display.

Agreed.

Incidentally, I meant to reply to continue the discussion on list, so  
I hope I'm not out of line adding the WG  back to the to header.

Take care,
Rob

Received on Friday, 22 August 2008 11:16:49 UTC