Re: Getting beyond the ping pong match

Rene Saarsoo schrieb:
> 
> Dão Gottwald wrote:
>> You write "It's not the copyright class and how it's used, it's the 
>> principle", and then start with two other special class names.
>>
>> I agree that "note" and "issue" are too ambiguous. Consequence is that 
>> they shouldn't be predefined classes, rather than that there should be 
>> no predefined classes.
> 
> Agreed.
> 
> But the main point I was trying to make is, that the meaning
> of any name you choose, more or less depends on the context
> it is used. For example I could have a page, that discusses
> different search-algorithms. It looks sensible to use class
> "search" when referring to different algorithms:
> 
>   <span class="search">BubbleSort</span> takes more time
>   than <span class="search">QuickSort</span>.
> 
> I can also use multiple class names to achieve completely
> new meanings:
> 
>   <div class="search results"> ... </div>
> 
> Or there could be police website, that lists different
> warrants - including search warrants:
> 
>   <ul id="warrants">
>     <li class="search">Permission to search Joe's apartment.</li>
>     <li class="kill">Special permission for 007</li>
>   </ul>
> 
> You could have the same class name at many places in one page,
> and it could mean different things depending on the context.
> CSS is also context-aware, supporting descendant selectors.
> 
> But with predefined classes the context is stripped off.
> Program looking for predefined class says: "I see a class
> 'search' there, so there must be a search field."
> 
> (I apologise, that my examples are completely made-up, and this time
> only cover one single proposed class name. But you can't say,
> that they are wrong, and no-one would do something similar.)

All the examples that you gave could be solved by defining 'search' for 
the form element only, which would be a good enough context. I'm not 
sure why WA1 defines it for aside, body, p, section and span.

>> Authors don't even have to know any of the predefined classes -- they 
>> can use names that they find useful! The question is whether or not 
>> some meanings should be exposed to user agents.
> 
> But can I use a predefined class name in some other sense than
> the spec defines it, if I find it useful?

For other elements, yes. The spec is flawed if it defines a meaning for 
one particular element where other meanings would make sense.

--Dao

Received on Monday, 7 May 2007 12:33:48 UTC