> I think we've learned our lesson about overloading parameters 
> in the name of
> compatibility, and the solution to this problem isn't terribly
> complicated  -- however I remain skeptical about whether this is a
> problem - real in the sense that if we solve it (in a 
> reasonable fashion)
> people will implement the solution.

To be honest, I'm not convinced that this is needed just yet. In the
example I gave another solution rather than asking for a METS with
OpenOffice, MODS and REC would be to follow the approach I mentioned
earlier which would be

To get the object, do a retrieve requesting OpenOffice XML schema
To get the record, do a retrieve requesting MODS schema
To get the metadata about the record, do a retrieve requesting the rec

This is the approach we are taking with a learning object repostory
whereby retrieving the learning object or the learning object metadata
is simply determined by whether you request a learning object schema, or
the learning object metadata schema.

However, others may have a different view (e.g. Mike wasn't too happy
with the record, rec example I gave). If we do need to be able to handle
aggregate schemas then as Ray indicated, I believe we need to revise the
parameters for requesting recordSchemas rather than overload the
existing one.