[251] in Public-Access_Computer_Systems_Forum
Cataloging Internet Resources
daemon@ATHENA.MIT.EDU (Paul Buchanan)
Thu May 14 11:13:19 1992
Date: Thu, 14 May 1992 10:10:28 CDT
Reply-To: Public-Access Computer Systems Forum <PACS-L%UHUPVM1.BITNET@RICEVM1.RICE.EDU>
From: Paul Buchanan <buchanan%wulib-jpb.wustl.edu@RICEVM1.RICE.EDU>
To: Multiple recipients of list PACS-L <PACS-L@UHUPVM1.BITNET>
----------------------------Original message----------------------------
I think Kathleen Burnett's comments on cataloging internet resources are very
valuable and deserve some follow-up discussion. It seems to me that one of the
first questions we need to ask is whether we will ultimately need a catalog of
network resources. As more and more source material is available in electronic
form, can we simply throw it all into WAIS servers scattered around the network
and let people search the sources directly without going through finding aids
and surrogate descriptors? Personally, I have real doubts about scaling up
this approach, but it seems to be at least implicit in much of the current
enthusiasm for WAIS technology (it also echoes some very old debates in
information retrieval about keyword retrieval vs controlled vocabulary).
If you assume that finding aids will still be necessary in the network
environment, then I think you need to ask what level of network resource
people really want access to. Many of the current projects in this area are
compiling lists of catalogs, directories, and similar kinds of "indirect"
resources (i.e. we are creating catalogs of catalogs); it seems to me that
the discussion of MARC formats for network services also focuses on these
kinds of services. I would suggest that this effort, while it is somewhat
valuable in the short-term, is perhaps not the long-term system we are
looking for. I think only a limited number of our users are interested in
searching other catalogs per se; I think what they really want is access to
the resources: books, articles, files, images, particular programs.
In the Z39.50 future, I hope that people will be able to search a "virtual
catalog" which is dynamically assembled from one or more servers in much the
same way that WAIS allows multiple searches to be combined at the user's
screen. There are obviously technical issues here (eliminating duplicate
records it one non-trivial problem), but in this environment the catalog of
catalogs becomes a meta-map which is used internally by the client software but
perhaps is not actually seen by the user (whether MARC is the appropriate
format for this kind of file seems problematic to me). I think the user would
want to "assemble" his virtual catalog in ways that do not really correspond
to the physical division of the database between different servers: by subject,
by cost (show me everthing I can get for free), or by "network distance" (show
me everthing I can get today), etc.
The second important part of this future is actually linking the catalog
records with an electronic delivery mechanism which gets the information across
the network. This may take several forms: a telnet session, an email message
for some server, an ftp request, etc. The important point is to let users
move as easily as possible from descriptions of actual resources (individual
documents or files) to the acquisition of these resources. This doesn't
require a new MARC format, it requires extensions of our existing catalog
records to incorporate electronic pointers to supplement the manual pointers
represented by call numbers and location codes. Note that this is a different
approach than the existing document delivery systems like CARL: in those systems
you search THEIR indexes to order THEIR documents. I would like to see these
functions separated by a network protocol so that all of our local catalog
records eventually become access points to an electronic delivery system.
Paul Buchanan <buchanan@wulibs.wustl.edu>
Washington University Libraries
St. Louis, MO
