[Ex-plain] NISO Metasearch TG2
azaroth at liverpool.ac.uk
Mon Apr 26 18:23:47 CEST 2004
On Mon, 26 Apr 2004, Mike Taylor wrote:
> > However, it should be possible to describe other less complicated or
> > less describable services using the same or very similar schema. In
> > particular OAI seems like a good candidate. The MetaSearch
> What else is on the table besides OAI-PMH?
The list that was drawn up as being in scope:
HTTP/HTML (probably undescribable, so just a pointer)
HTTP/XML (eg proprietary SOAP, potentially more describable)
X.500/LDAP (Looks describable from a quick skim)
GRID IR (If such a thing exists)
GIS (If there's a standard protocol/service)
Important to note is the distinction between protocol and service. We
don't want to try to describe -how- to extract information out of HTML,
just what's available if you know how to process it. The same way as
we're not describing Z39.50 or SRW messages, just what's available through
Service descriptions are used for to:
1) Configure access to a service
eg: here's the information for how to connect to the LC Z server
2) Enable discovery of a service based on its supported features
eg: find me all services that support a Series Title search.
> > There are two levels at which ZeeRex may need to change to accomodate
> > 1) Minor changes. The notion of OAI sets does not easily fit into the
> > 2) Profiles. The metasearch taskgroup only wants to describe the
> > 3) Semantics. Some of the semantics may be fuzzy when describing
> Have trouble counting to three? :-)
Apparently so :)
> I don't understand the significance of "The metasearch taskgroup only
> wants to describe the service, not the collection." Otherwise, this
> all seems reasonable to me. Even Google's "link:" as an index.
For example, the collection of objects which is being searched will have a
separate schema to the service through which that collection is searched.
In abstract, a collection is a group of items. For example a group of
MARC records that makes up a library catalogue. That collection needs
some description to enable discovery as an appropriate collection to
search, but that's different from a service which enables a client to
search it. One collection may have many services, some of which are
unknown to the owner of the collection. But each service makes one and
only one collection available. This fits the ZeeRex model very nicely.
Service is the 'how', whereas collection is the 'what'.
,'/:. Dr Robert Sanderson (azaroth at liverpool.ac.uk)
,'--/::(@)::. Special Collections and Archives, extension 3142
,'---/::::::::::. Nebmedes: http://nebmedes.o-r-g.org:8000/
I L L U M I N A T I
More information about the Ex-plain