Add a feature to MEX to allow a client to specify a list of dialects (and
identifiers).
Need to think about whether to allow the the "format" (or whatever we called
it) to be on a per dialect basis or not.  Thoughts on this would be
appreciated.

Current form:
  <mex:GetMetadata ... >
    (<mex:Dialect>xs:anyURI</mex:Dialect>
       (<mex:Identifier>xs:anyURI</mex:Identifier>)?
    )?
    <mex:Content> xs:anyURI </mex:Content> *
  </mex:GetMetadata>

Proposal:
 <mex:GetMetadata ... >
   <mex:Dialect uri="xs:anyURI" 
                identifier="xs:anyURI"? 
                Content="xs:anyURI"? .../> *
   ...
 </mex:GetMetadata>

This gives you the flexibility of choosing different formats on a per Dialect
basis.  What's nice about this is that since we've defined the "all" Content
URI, you could ask for all forms of a certain metadata with a single element. 
However, if you want just 2 of the forms (e.g. URI and EPR) then you'll need to
use 2 elements.  Not a huge issue and I'm not sure one that will happen too
often.  It seems that if you want a reference to the data you'll probably plan
on only using one way to retrieve it - not both.

We could make the Content attribute a list of URIs, but that seems overly
complicated when adding a new Dialect element gives you the same end result
with only a few more bytes on the wire.

proposal in comment #1 accepted

Done