RE: Napster/Docster + Libraries - OCLC/RLIN = MARCSTER?

["Frumkin, Jeremy" <frumkinj@u.library.arizona.edu>]

Karen Coyle writes:

> My first concern would be for search quality. Can 
> Napster/Marcster achieve
> the detailed level of searching that we need in large 
> bibliographic files
> and still serve a large number of users? I presume that as it 
> exists today
> Napster makes use of only a small amount of searchable metadata and
> probably only one kind of searching (keyword). With hundreds 
> of millions of
> MARC records, simple searching gives overwhelming and often 
> innaccurate
> results.

I don't think MARCster would need to be limited to simple searching; I don't
see why it couldn't use advanced searching using full MARC metadata. In the
napster world, this might be more of a problem, because of the increased
factor of clients/servers, but I think we would have a manageable set of
systems within the library world which would allow for timely searching in a
more complete manner.

> 
> My next concern would relate to duplicate retrievals. Does 
> Napster de-dup
> at all? Yes, you get some duplicates on OCLC, but in a 
> distributed model,
> every 3-letter code on an OCLC record would result in a 
> duplicate retrieval. 

If needed, I'm sure MARCster could be built with a de-duping feature
(perhaps configurable by the user). In fact, this is an excellent idea.

-- Jeremy