Re: Extension Mechanism for HTML

Jonas Sicking schrieb:
> 
> Rene Saarsoo wrote:
>> As I see, it's now the question: do we need all the advantages of
>> @role although it has some additional complexites, or shall we
>> settle with simpler (but maybe too simplistic) predefined classnames?
> 
> Like I've suggested before, why not have both. That way people can use 
> the simple syntax of @class as long as the predefined classes we have 
> defined are enough, and you get all benefits of @class that you listed. 
> And if you want to express more complex semantics or define your own 
> taxonomies you can do that by using the @role attribute.
> 
> The downside of doing this is that implementations that want to extract 
> the semantic meaning from a page will have to look at two attributes. 
> However compared to the complexities of @role, also looking at @class is 
> very little extra overhead.

Makes sense to me. When I was arguing for @class, it was never really 
against @role.

--Dao