The way I see surrogate diagnostics working in SRW is as follows:

We have one XML format for diagnostic
This can be used in the response message as a non-surrogate.
This XML format can be used as though it were a record schema in the record
list - i.e. the record list could look like

Record (schema=MODS)
  Data

Record (schema=MODS)
  Data

Record (schema=OAIMarc)
  Data

Record (schema=DIAGNOSTIC)
  Data

Record (schema=MODS)
  Data

Etc. (clearly with appropriate XML wrapping).

That seems fairly clean to me?

Matthew



> -----Original Message-----
> From: LeVan,Ralph [mailto:[log in to unmask]]
> Sent: Thursday, June 20, 2002 7:49 PM
> To: [log in to unmask]
> Subject: Re: result set position
> 
> Yes, at least I do.
> 
> We can come to some sort agreement on the structure and content of the
> surrogate to fix your bad feelings.  But I like have the placeholder.  It
> makes life much simpler when records are going away dynamically.
> 
> Ralph
> 
> > -----Original Message-----
> > From: Ray Denenberg [mailto:[log in to unmask]]
> > Sent: Thursday, June 20, 2002 2:33 PM
> > To: [log in to unmask]
> > Subject: Re: result set position
> >
> >
> > "LeVan,Ralph" wrote:
> >
> > > Why not a surrogate diagnostic in the middle of the
> > records?  I like them.
> >
> > It's caused complexity for Z39.50 particularly in the ASN.1.
> > I suspect it
> > would cause similar complexity in the XML.  You have a
> > sequence of records,
> > each is a  "response record" which is either a "retrieval record" or a
> > "surrogate diagnostic".  The retrieval record is accompanied
> > by a record
> > syntax (for srw, a record schema) and the surrogate
> > diagnostic has its own
> > funny format.
> >
> > Do we want this in srw?
> >
> > --Ray
> >