Whither z39.50?

[Genny Engel]

> I would see it being used more as a mechanism for multiple databases to be
> searched simultaneously rather than anything else.  I also expect that
> z39.50 clients will mainly be used by power users and the bulk of access
> with be via http-z39.50 gateways.

While I would agree that only "power users" would actually go to the
bother of installing Z39.50 clients on their desktops, for us it hasn't
been true that most Z39.50 access is via the web.  In fact, our Z39.50
access to OCLC and RLG databases isn't yet supported in our web interface.
For right now, our major Z39.50 users are those who telnet to our
line-mode catalog interface and select a Z39.50-accessible database as if
it were any locally-loaded database.

Currently, the user must select the database in order to search it (i.e.,
they can't enter a search and then select a list of databases to perform
the search against in a single step; they must search each in turn).  But
that's just the way our search works.  It doesn't really have anything to
do with whether the database in question is Z39.50-accessed or locally
loaded.  I don't know of any library systems that make this distinction in
parsing the search to determine whether or not more than one database can
be searched in a single statement.

Libraries supporting the Z39.50 gateway on a central server will probably
get the bulk of Z39.50 use through HTTP only if the bulk of their overall
OPAC access is via HTTP.  That's a moving target right now.

Genny Engel
MELVYL System User Services
University of California
gen at dla.ucop.edu