subsetting DocBook5

[Nathalie Sequeira <n@n-faktor.net>]

Hello all,

I am currently creating a subset of docBook 5.0 in RELAXNG format, 
primarily to make the choices available for the structuring of docBook 
XMLs for a team of non-technical staff more straightforward.

Though the explanation of how to do this (at 
http://www.docbook.org/docs/howto/#customizations) seems to be quite 
straightforward, I am having doubts as to whether and which related 
definitions of elements (and also attributes) to exclude.

For example:
None of the db.verbatim.blocks will ever be needed, nor will 
db.verbatim.inlines.
However, I also found the db.verbatim.contentmodel, which also 
references db.verbatim.inlines.
And of course, there are also the definitions of the single verbatim 
elements too.

Does that mean I need to remove ALL definitions related to the verbatims 
so bad things don't happen when the subset is in use - as well as 
rewrite any other definitions that contain references to the verbatims?
Or will these just get skipped over if I disallow the groupings?

And if disallowing the groupings does all the magic: is there a detailed 
overview of which groupings there are, or is going through the Schema 
(perhaps searching for grouping terms like "blocks" "attributes" etc.) 
the only way of going about?

Thanks!
Nathalie