> I completely agree with Ralph that "only one implementation needs it"
> is not a good reason to remove a diagnostic! However, I'm
> still puzzled about what you're going to use this for, Ralph,
> when the SRW/U model so clearly states one DB per server. I
> quote from the "SRW's Relationship to Z39.50" document at
> http://www.loc.gov/z3950/agency/zing/srw1-1/z3950.html
>
> No distinction between server and database
>
> SRW does not distinguish between a server and a
> database; it is hoped that elimination of the database
> concept will effect significant simplification (since
> the multiple-database concept in Z39.50 has caused
> such complexity), for example Explain is significantly
> simplified (and hopefully it will therefore become
> more widely implemented).
>
> So under what circumstances would diagnostic 8 be used?
In the case of Theo's x-collection parameter, and/or other ways is which
a collection (database) might be specified in an SRW extension (i.e.
either through out of band ways such as SOAP headers, part of the base
URL etc. or as a new CQL index).
As this is a local SRW extension, it really should have a local
diagnostic (if it needs one), not an SRW standard defined one.
(IMHO)
Matthew
|