>   The other half, if I
> understand him correctly, is that he he wants to have a situation
> where a client asks for the DC schema, or the CIMI schema, or
> whatever, and the server unilaterally says, "No can do, but here's a
> DCX record".  (Right, Theo?)
>
> Now in out-of-the-box SRW/U, that's illegal.  But it's precisely the
> sort of requirement for which we introduced extraRequestData.

How are we conveying in the spec the semantics that Rob describes here?
Specifically I think Rob's saying "the value of recordSchema within record
within records, in the response, must be the same as the value of
recordSchema in the request, if supplied (and if the record is not a
diagnostic) *unless* there is some extraRequestData which modifies the
semantics herein expressed."

Is that about right?

In 1.0 there's no such rule. (There's no extraRequestData in 1.0 but this
discussion implies that there is some implicit rule to same effect except
without the *unless* clause.) At least, there is nothing so stated in 1.0.
And nothing, far as I can tell, written to that effect in 1.1. Personally
I'm happy with the rule but I can't even remember that we agreed to it.
anyway I suppose it should go in the service definition.  Are we comfortable
that we're covering all or most of the semantics such as this?

--Ray