In message <49CB756E.3090608@oucs.ox.ac.uk> Sebastian Rahtz <sebastian.rahtz@oucs.ox.ac.uk> writes:
> * record authorship of an individual odd
> * record dates of changes to classes and elements
> * record status
> * record source of a modification, as you say (cf DOI discussion!)

Yes, some of those, you could argue, belong in the teiHeader for the ODD as a whole, no?  I like the idea of recording dates/natures of changes to individual class/elementSpecs.  

However, am I missing something? biblFull doesn't give you revisionDesc which is where you'd store the dates/changes to classes of elements, no?

> it would be overkill, of course, but allowing
> people to use it there seems logical; better than
> reinventing the info under other names
> 
> I appreciate what Mr Occam has to say, but its equally
> Occamish to simply re-use existing structures and nor
> reinvent any wheels?

I agree.  While usually I'm all for the simple approach, if we can find a way to do this which also includes the ability to record other forms of metadata specific element/class metadata, this would be a valuable addition to ODD.  But what I'd worry about is rushing into doing this.  Perhaps it should be delayed until we discuss the next major revision of ODD (which is needed and important) alongside such things as <constraint>.

-James