Re: [dita] Proposal for DITA namespaces

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

Options for namespaces for DITA.

I think we have the following practical options for using namespaces 
with DITA 1.0:

1. Have exactly one namespace for all element types, including local 
specializations.

2. Have exactly two namespaces, one for maps, one for all other element 
types, including local specializations.

3. Have many namespaces: one for maps, one each for topic, task, 
reference, concept, different domains, locally-specialized elements, etc.

4. Have two namespaces for DITA-defined types (including domains). Local 
specializations are in their own namespaces.

Of these options, I think that (1) will not work, for the simple reason 
that maps and topics use the same element type names with different 
content models. Thus maps and topics are fundamentally different 
document types and therefore represent different namespaces.

I think that (3) is just nuts, but that's mostly because I don't yet 
understand the implications of having that many namespaces, especially 
for schema definition and modularization.

Option (4) is probably the right answer for the long term for the simple 
reason that local specializations are clearly in their own space and 
should, at least intuitively, be in a distinct namespace. What we don't 
yet understand is exactly how XSD-based declarations would work in this 
case.

My intent is to do some experiments to test options 2 and 4 and see if I 
can determine anything definitive.

Cheers,

Eliot
-- 
W. Eliot Kimber
Professional Services
Innodata Isogen
9390 Research Blvd, #410
Austin, TX 78759
(512) 372-8122

eliot@innodata-isogen.com
www.innodata-isogen.com