[1149] in Public-Access_Computer_Systems_Forum
Re2: Library Automation
daemon@ATHENA.MIT.EDU (Research SW Design, D Goldman,PRT)
Fri Sep 4 10:59:59 1992
Date: Fri, 4 Sep 1992 09:53:28 CDT
Reply-To: Public-Access Computer Systems Forum <PACS-L%UHUPVM1.BITNET@ricevm1.rice.edu>
From: "Research SW Design, D Goldman,PRT" <RSD@AppleLink.Apple.COM>
To: Multiple recipients of list PACS-L <PACS-L%UHUPVM1.BITNET@ricevm1.rice.edu>
----------------------------Original message----------------------------
Bernard --
Thanks for your thorough reply!
Accuracy
--------
Yes, it would be great if there were one master bibliographic data source that
was easily accessible and highly accurate. At the moment we instead have
hundreds of specialized bibliographic databases, each available through a
number of different vendors via dial-up, Internet, CD-ROM, monthly diskettes,
etc. Each database is created in its own unique style, and each vendor then
further mutilates the appearance of the data.
This current level of anarchy, combined with the incredible volume of data
involved, makes me pessimistic about expecting any single organization (Library
of Congress?) to take over the entire job in my lifetime.
A universally agreed-upon standard format would help a lot. But each different
specialization has its own requirements and traditions (eg, Medlars/Medline has
its system of Subject Headings, Chem Abstracts has its chemical Registry
Numbers, etc.), so I am pretty pessimistic about even this! Still, if a de
facto bibliographic standard took root on the Internet, perhaps things would
eventually improve.
Large databases
---------------
I agree with you that the unique "bibitem" identifier method becomes
impractical once you've gathered a few thousand references in your database.
The way PAPYRUS (and some of the other PC-based bibliographic programs) deals
with citing a reference is to allow you to jump from your word processor to the
bibliography program, locate the reference(s) of interest via an
author/year/etc search, and then jump back to your word processor to paste in a
unique identifier (typically a number). When the manuscript is ready, you then
have the bibliography program read the manuscript and replace these identifiers
with the correct footnote numbers or author-year citations (as appropriate for
the journal style you specify).
Most bibliographic programs will be much faster than BibTeX with medium to
large databases. BibTeX is simply working with a huge text file, while other
programs use standard database techniques to store and index your data.
[Crass commercial interlude: By the way, PAPYRUS is able to import existing
BibTeX databases, and to interact with TeX manuscripts.]
Links
-----
To date, personal bibliographic programs have concentrated on producing printed
bibliographies in the correct style, and allowing a user to easily search
his/her personal reference collection. PAPYRUS is gradually evolving in the
direction you are interested in, in which additional information and links can
be added to the underlying citation information.
Allowing separate databases in order to correlate multiple versions of an
author's name, multiple versions of a conference title, etc is a good idea. As
you point out, though, this might not be a terribly useful addition right now
if each end-user has to enter this information her/himself.
I am *very* interested in anyone's thoughts on this subject, as we are working
on a redesign of PAPYRUS right now. What sorts of automated links between
references or parts of references (eg, authors) would help you in your work?
Would you, as an end-user, have to enter this information yourself, or could
there be some mechanism established for a group of specialists to maintain a
downloadable set of linked references?
UNESCO
------
Do you have the details on this document, so that I can obtain a copy?
Further info on PAPYRUS
-----------------------
Rather than clutter up PACS-L with specifics about PAPYRUS, we will be happy to
e-mail or physically mail our product literature to anyone interested.
-- Dave Goldman
Research Software Design
