> -----Original Message-----
> From: Theo van Veen [mailto:[log in to unmask]]
> Sent: Tuesday, May 14, 2002 8:22 PM
>
> What (default) prefixes should be used in distributed searching?
> Ralph will return an error message if I try "dc.title:abc or
> bath.title:abc". Why isn't the default just no prefix?

That's just not true!  I've already said that I'd support an agreement to
accept dc as a prefix by all SRW servers.


> I have the strong feeling that we are currently on the wrong track.
> We are mixing up Z39.50 attribute sets with dc name spaces, while
> the solution is quite simple: use user understandable names for
> search fields.

What are those fields and how will you keep them from colliding with those
same names in other IndexSets?  Be specific!


> It is possible in Dublin Core for description,
> why is it not possible in CQL for searching?

It is possible for DC because all their elements are in their namespace.
Put the indexes in the DC namespace and that will work to.


> The abstract Z39.50 attributes were usefull in case of MARC
> descriptions, but in line with Dublin Core  I think we should map the
> Z39.50 search attributes to user understandable names instead of
> sticking to the attributes.

And that is exactly what we are trying to do.  We will map DC.title to (for
example) Use=4, Position=3, Structure=2, Completeness=1.  Relation will be
specified in the query and Truncation is statically defined as 104.  Any
client that sends me a DC.title search will get the same behavior as if it
had send that combination of attributes to my Z39.50 server.

Ralph