[uf-discuss] ISO Dates and Durations using Style

[Scott Reynen]

On [Sep 27], at [ Sep 27] 4:27 , Toby A Inkster wrote:

>> If any style sheet language can be used, why don't microformats  
>> create
>> their own style language eg:
>> <span class="bday" style="bday.1968-01-04">4th Jan, 1968</span>
>
> By definition, the contents of the style attribute must be in "the  
> default style sheet language". The default style sheet language is  
> by definition CSS unless a Content-Style-Type header (either HTTP  
> header or <meta http-equiv>) is present. There can only be one  
> default style sheet language per document, thus any document which  
> wants to use a non-CSS style sheet language in the style attribute  
> cannot use CSS in the style attribute. (That is, you can't use CSS  
> in some style attributes and non-CSS on others.)

That's certainly a reason not to make this a recommendation for  
everyone, but as we already have two alternative methods (machine data  
as human data and abbr-design-pattern), I'm not convinced we should  
discount this idea altogether.  Conflict with CSS is only an issue  
with inline CSS, which is widely regarded as a poor practice anyway,  
especially among publishers paying enough attention to have concerns  
about the abbr-design-pattern.  And it may not even be an issue there,  
as CSS says user agents "must ignore declarations with invalid  
values." [1]

I'm afraid we may be dismissing this too hastily.  My initial reaction  
to this idea was to view it as semantic abuse of the style attribute,  
but after thinking about it more, I now think it makes a lot of sense  
that "1968-01-04" should be treated as styling instructions for "4th  
Jan, 1968."  We already have different kinds of styling instructions  
in CSS (i.e. visual, aural, and physical).  I'd argue that this is a  
simply another type of instruction, context-dependent, just as much  
explaining how the content should be presented, e.g. it should be  
presented in whatever way ISO 8601 dates are presented in a given  
context.  There may be good reasons this won't work, but I don't think  
the fact that no one has previously used @style for anything other  
than CSS is one of them.  After all, the same was widely true of  
@class before we started promoting the alternative uses allowed under  
the HTML spec.

[1] http://www.w3.org/TR/CSS21/conform.html#conformance

Peace,
Scott