[osis-core] Update on OSIS Schema Issues

[Todd Tillinghast]

<snip>


> > 2. scope attribute on div?
> 
> Steve noticed a post on allowing scope on <q>. Does scope mean "what I
> contain?" Is so, isn't that a duplicate of osisID? But, note you can
> have elements that don't have an osisID where you want to indicate
what
> it contains.
> 
> >
> > Todd wants to have a scope attribute on <div> that uses the same
regex
> > as osisRef to indicate the range of references in a <div>. Chris
> > agrees but wants to add that scope should  allow a list of
osisRefTypes.
> >

I would not want to use osisID to indicate "what I contain" because
osisID is "I am".  The other consequence would be that you would have
VERY lengthy osisIDs where a <div> contains several verse.

osisID on <div> should be reserved for the cases where the <div> IS that
osisID.  (I guess this will not likely happen with a BCV based osisID
now that we have a <chapter> element.)

An argument can be made against the practice of scope, in that the value
should be able to be derived from the osisIDs of the contained elements.
I like the scope attribute because it is easy to tell from the <div>
element what is contained without having to compute the value from the
children.  

(I also like the idea for the future of being able to have only the base
part fo the tree, leaving off the smaller branches and leaves, and then
having the smaller branches and leaves as separate documents that are
loaded as needed.  In this case the scope would be VERY useful.)

<snip>

> >
> > 4. <chapter> as child of <p>
> 
> Partial victory for Todd. ;-)
> 
> Keep the current chapter element (as extension of milestoneable), but
> have a separate chapter element (I know, flood of elements coming into
> the schema) that is the child of <p>, <l>, <speech>, <item>...., but
it
> needs a different name. (Because we can't enforce use of only
milestones
> from the "real" chapter element in these lesser elements.
> 

I think the name for the new milestone only element should also be
<chapter>.

Todd