URLs, PURLs & TRULs: Link Maintenance in the Web-accessible OPAC

[Thomas G. Tyler]

******* cross posted ********

The paper I presented at CIL'99 -- URLs, PURLs & TRULs: Link Maintenance
in the Web-accessible OPAC -- is now available at:

	http://www.du.edu/~ttyler/cil99/urlsetc.htm

If you are interested in this subject you may want to take a look at the
visuals prepared for CIL which are available at the URL listed above.

Web displays from more than ten different ILS web implementations (CARL,
DRA, Endeavor, Innopac, Melvyl, InterCat, Notis, PALS, Sirsi, etc.) are
shown in "best case" (relatively simple 856 field data) and "worst case"
(Starr Report - 12 URLs in four 856 fields) examples.

If you are only interested in what "TRULs" are I'll tell you here.  I
invented TRULs (an acronym for Tribal Rules) to represent the all too
common "work-arounds" that are required by all of us who have to work
with and maintain automated library systems.  

In the context of the CIL paper I was attempting to describe what
librarians have to do to maintain URLs in their OPACs in the absence of
approprate tools for the task from their ILS vendors.  Example: The DRA
tribe generally has to move note information ($z) to the subfield for
materials specified ($3) to have it display in the DRA web opac; those
in the Innopac tribe have to do just the opposite if they want $3 data
to display.
Neither these ILS' nor any of the others I looked at support real-time,
ongoing, systematic URL validation.

Special attention was paid to the cataloging efforts of the Government
Printing Office which is singularly responsible (I believe but can't
really prove) for creating most of the linked Marc records in American
OPACs today - 8,000 records issued with one or more 856 fields since the
Spring of 1995, 5,000 of which are unique.  When the URLs were tested by
LinkBot in December 1998, nearly 20% presented problems!

One group of visuals illustrates common GPO cataloging errors that made
URLs bad from the start (e.g. use of $a and not $u for URLs, "http:///",
commas instead of periods, unsubfielded notes appended to the URL, badly
formed PURLs, etc.).  With no URL validation (or quality control) at the
point of cataloging instant URL problems have been created for dozens or
hundreds of libraries batchloading these records from OCLC, Marcive,
Auto-Graphics, or LC.

The general thesis of the paper is probably this:  link validation in
library OPACs is a growing problem, but few libraries are dealing with
the problem in a comprehensive way at the present time because the tools
available to them are unsatisfactory for the task.
Those few libraries that have worked out successful link maintenance
strategies, have often had to accomodate their activity in this area to
limitations imposed by their automated library system.

Tom Tyler
Associate Director for Budget & Technical Planning
University of Denver Library
Denver, CO  80208

303-871-3334
ttyler at du.edu