Re: (phrase extensions and bidirectional text orverride <span> and <bdo>) part of my review of 3.12 Phrase elements

On Jul 26, 2007, at 4:15 PM, Smylers wrote:

>
> Robert Burns writes:
>
>> phrase extensions and bidirectional text orverride <span> and  <bdo>)
>> part of my review of 3.12 Phrase elements
>>
>> Span SPAN:
>>
>> The <span> element makes mention of special meaning to @title when
>> used with <dfn> but nothing else is said about that here. An example
>> and explanation would be helpful.
>
> True, but the mention of "dfn" is a link to what the spec says  
> about the
> <dfn> element, where it is explained.

I understand. It just would be better to include a brief description  
in each of those subsections. Even with the prose that are there the  
various ways this facility can be used is only becoming clear to me  
after much dialog on this list. You've provided several excellent  
examples and other prose that would greatly improve the draft. It  
needs some attention on these issues. Perhaps there's a matching  
algorithm in another section, but I I don't understand (from the  
draft alone) all the ways these facilities are to be used. Even with  
a UA algorithm, I think the draft needs to pay more attention to  
writing descriptions from an author's point-of-view too. For example,  
the ratios/meter/progress sections describe things only from the  
perspective of how a UA handles content it's presented with. The  
other angle should also be in the draft: that is how an author  
composes that text for the UA to handle. Those are very different  
frames of reference and asking authors to "reverse engineer" the UA  
requirements to figure out how to author is inappropriate.

>> Consider requiring at least one of the global attributes to ensure
>> <span> has some specified semantics associated with it.
>
> Why?  There are things that can be done with a <span> that has no
> attributes, for example:
>
> * <span> can delimit a term which is defined elsewhere in the document
>   (with a <dfn>).
>
> * Even without a class or id attribute, <span> can still be  
> selected in
>   CSS for special styling, for example with this HTML:
>
>     <p class="fruit"><span>Apples</span>, <span>bananas</span>, and
>     <span>cherries</span> are all fruit.</p>
>
>   those <span> elements can be distinguished with something like:
>
>     .fruit > span { color: lime; }
>
>   without requiring the class to be applied to each <span>  
> individually.

OK. maybe it should only be a recommendation. However, those examples  
provide further evidence that we need a TERM element since that is  
the semantic being expressed with the generic SPAN element. This  
special TERM semantic is something used quite often in documents.  
It's a glaring omission from the HTML semantic vocabulary. It's hard  
to imagine use of SPAN without one of the global attributes if it's  
not being used implicitly as a particular semantic. I would say what  
that example really represents is:

   <p><span class="fruit" >Apples</span>, <span class="fruit"  
 >bananas</span>, and
     <span class="fruit" >cherries</span> are all fruit.</p>

or

   <p><span class="term" >Apples</span>, <span class="term" >bananas</ 
span>, and
     <span class="term" >cherries</span> are all fruit.</p>

Or even better:

   <p><term class='fruit-example' >Apples</term>, <term class='fruit- 
example'  >bananas</term>, and
     <term class='fruit-example'  >cherries</term> are all fruit.</p>

Leaving the class off of the SPAN in your example leaves the  
semantics of the element only useful in the context of the paragraph.  
That is if one copies one of those terms and pastes it elsewhere it  
loses its semantics and its presentation because of an authoring  
shortcut. It would be very difficult to create an authoring tool UA  
that could handle such authoring shortcuts intelligently. It would be  
better to avoid the shortcuts.

Take care,
Rob

Received on Thursday, 26 July 2007 21:52:19 UTC