> Date: Tue, 9 Aug 2005 14:32:06 -0400
> From: Ray Denenberg <[log in to unmask]>
> If this is a real requirement [...]
... which I admit it is not, at leaast not yet ...
> [...] I believe the only sensible solution is the straightforward
> one (that I think Matthew is suggesting), that is, add another
> parameter -- next version of course. Call it something like
> 'container', with two sub-parameters (1) name/id of a container
> schema, e.g. mets; and (2) one or more "base" schema name/ids,
> e.g. mods, rec, dc.
I completely disagree: this solution is neither necessary nor
sufficient. It is unnecessary because we already have the ability to
name schemas using arbitrary URIs, which has have arbitrary structure,
and it's insufficient because it doesn't allow a client to specify
more complex record structures such as schema A contained in schema B
contained in schema C.
What will happen in practice is that when someone needs to ask for,
say, A Dublin Core record with some record-level metadata, they will
just make up a schema that contains these things and a URI that
identifies it. If several people do this separately, they will surely
end up with different and incompatible schemes. So it seems better to
me that we provide some guidance.
But, what the heck, no-one actually wants to do this anyway, right?
/o ) \/ Mike Taylor <[log in to unmask]> http://www.miketaylor.org.uk
)_v__/\ "Historically, Taunton is part of Minehead already" -- Monty
Python's Flying Circus.