On Fri, 2005-12-02 at 11:14 -0800, Nathan Young -X (natyoung - Artizen
at Cisco) wrote:
> Hi.
>  
> > My point was: use the best format for the task, and don't be afraid to
> > invent a new format if it's others are fitting a square peg 
> > into a round
> > hole, but never, no never play the 
> > tunneling-markup-in-protean GI game
> > of
> > 
> > <div class="monty">
> >  <span class="python"/>
> > </div>
> 
> >From a core of basic agreement I have a couple of objections.
> 
> The first, which I think is echoing things Robert has also said, is that
> if part of the task in "use the best format for the task" is sending out
> documents to be edited by XML illiterate users, then making up my own
> markup language is just not an option.

And why on earth would I expect an *Office document* user to write a
document like

<monty>
  <python/>
</monty>

???

I would expect an office user to create *ODF* using, say OpenOffice, or
Office XML using, say MS Word.  The above is a completely dofferent
case, and it is for situations where you *are* dealing directly with
XML.  No one is saying that average users should deal directly with XML
in any scenario.  Heck, even if you're talking XHTML, average users
wouldn't deal with the XHTML.  They'd deal with the Dreamweaver UI, or
whatever.

> The second is that there are times where the kind of
> schema-tunneling-through-attributes is a good design pattern.

I think there are almost no cases where this is true.

> Layering
> semantics on XHTML using attributes is a big step up in expressiveness,
> and a necessary design pattern if you want to use CSS for formatting
> complex pages.  To me this is a perfectly valid use case:

Umm.  CSS works perfectly well with custom XML, thank you very much,
and this is supported on almost all browsers.


> <ul class="navigation">
>  <li><a href="home" class="home-link">home</a></li>
>  <li><a href="next">next</a></li>
>  <li><a href="previous">previous</a></li>
> </ul>
> 
> Many microformats are also valid design patterns to my eye.

I think most microformats in most cases are a hopeless stretch, and in my
experience when you try to apply them to real world tasks, you run into the
*exact* same semantic problems as you do with custom markup.


-- 
Uche Ogbuji                               Fourthought, Inc.
http://uche.ogbuji.net                    http://fourthought.com
http://copia.ogbuji.net                   http://4Suite.org
Articles: http://uche.ogbuji.net/tech/publications/