On 25/08/2009, at 2:24 AM, Frank Mittelbach wrote:

> Right now we have
>
> \DeclareDocumentCommand  - doing both
>
> \DeclareDocumentCommandInterface ... making a stab at separation but  
> not
> thought through really (the names are pretty bad really).
>
> Personally I would like to table it for now and if we return to it  
> with some
> good ideas it isn't too hard to even change anything that by then used
> \DeclareDocumentCommand.
>
> Alternative is to go on through a few more cycles to get a grip on  
> this split
> (layer -1 and 0) but I gurantee that this is going to be difficult  
> at the
> current stage.

Indeed.

> The only problem that I see is the following:
>
> \DeclareDocumentCommand is quite a fine name and if we attach it to  
> something
> that bundles layer -1 and 0 then it will be a little difficult to  
> use it
> later, say, only for layer 0 declaration
>
> So having said this, perhaps tabling it isn't such a grand idea  
> after all and
> we should discuss naming and concepts a little further first.

Well, when I think about xparse and the idea of multiple layers -1, I  
can't help but feel that xparse takes a fairly limiting approach.

Consider Bill Hammond's example:
    \section*[short]{long}
to
    <section numbered="false">
    <shorttitle>short</shortitle>
    <sectiontitle>long</sectiontitle>
    </section>

There is no way we could do that at present with xparse. We'd need to  
assign names to all optional arguments and some sort of mapping  
between boolean flags and tag attributes.

\DeclareDocumentCommandInterface needs to take a step back if it's to  
be at all useful for more markup forms than simply LaTeX2e-like.

Brain switching off so I shan't attempt any more thinking...

Will