Re: [dita] Clarification of <itemgroup>

["Eliot Kimber" <ekimber@innodata-isogen.com>]

Robert D Anderson wrote:
> The spec shows that p is contained by itemgroup, and that itemgroup
> contains p, but not the other way around. Is this a request to allow
> itemgroup within <p>?

Yes, I must have misread something before, but in implementing my 
specialization I did discover that itemgroup is not allowed within <p> 
and confirmed that that's what the spec says.

So this is worse than I thought, in that there is *no* way to do the 
specialization I need.

I see this as a bug in that as currently defined there is no way to 
directly bind <data> to subelements of <p> that are not also allowed in 
<ph>. Since you can bind <data> to *some* children of <p> the fact that 
you can't do it for *any* seems like a bug (that is, what is allowed for 
some should be allowed for all).

I'm not sure what the most appropriate solution is, as I can see several 
alternatives that would allow the specialization pattern I need:

1. Allow everything allowed by <p> in <ph>

2. Allow itemgroup w/in <p> (and allow itemgroup to nest)

3. Define a new wrapper base type analogous to itemgroup but more generic.

Relaxing the content model of <ph> would be the easiest to implement but 
might surprise tools that currently don't expect figures within phrases 
(although it really shouldn't matter).

Allowing itemgroup w/in <p> would be easy to implement and would 
probably not be as disruptive.

A new wrapper type is probably the cleanest solution but would be the 
hardest to implement.

Cheers,

E.

-- 
W. Eliot Kimber
Professional Services
Innodata Isogen
8500 N. Mopac, Suite 402
Austin, TX 78759
(214) 954-5198

ekimber@innodata-isogen.com
www.innodata-isogen.com