why both a domain specialization and a specialization of a map (ortopic)?

[John Accardi <John.Accardi@Sun.COM>]

In today's SC meeting I raised the question about the wisdom of 
implementing both a domain specialization and a specialization of a 
topic or map for the same structure...

Specifically, we discussed creating a specialization of map for a 
learningobject and then also doing a domain specialization of same.

Other than the obvious power of a domain specialization (being able to 
use across all info types) what are the advantages of doing this "double 
implementation"?

The analogous question referring to general DITA is ...

If bookmap.dtd is a specialization of map.dtd, why hasn't there been a 
domain specialization of a book(map) for use across all info types?  Is 
it because it doesn't make sense to be able to use a "book" anywhere 
(across all the infotypes)?

I'm hoping that knowing this for bookness in general DITA will help us 
determine if we in the SC should follow example with our learning object 
implementation.

Any insights? thoughts?

Thanks,
John