University of Florida/FCLA and NOTIS
The University of Florida (Gainesville) was the first U.S. library to purchase NOTIS; UF staff converted the Northwestern VSE programs to run under MVS; and FCLA was the first big consortium using the NOTIS software.
In the late 1970s the different Florida State University System libraries were interested in an automated circulation system. UF (the University of Florida – Gainesville) had an old IBM punched card system that would not be supported for much longer. Nolan Pope had been to Northwestern and had talked to Jim and Velma and was interested in bringing NOTIS to UF, however, it was decided that UF should be aligned with the other campuses’ initiative. The bid process resulted in the purchase of a CLSI system for each campus. The other campuses implemented CLSI [for Circulation], however, CLSI was unable to handle the UF libraries’ requirements, which led to a separate procurement process for UF.
Key players in these early days were Gus Harrer (UF Director of Libraries from 1968-84), Max Willocks (Associate Director for Public Services), and Nolan Pope (Head, Systems and Computer-Based Operations).
Integrated Computer Support for the University of Florida
- PROPOSED SYSTEM PROCUREMENT [sections 1,2, and 3 only] describes the procurement
process. It weighs the strengths and weaknesses of NOTIS, DOBIS/LIBIS,
Cincinnati Electronics’ CLASSIC, and the Dataphase ALIS system at that time
(May 1980). [Thanks to Rich Bennett, former Chair of the Circulation Department, for locating this
document.]
One can see that, while the NOTIS Acquisitions was rated highly and Cataloging, very highly rated, Circulation (prior to the 1985 “New Circulation”) was less highly rated. However, Northwestern was planning to write a new circulation module, and UF worked jointly on the functional specifications.
Nolan notes: “In addition to being a fully integrated system, I think a major attraction was NOTIS' ability to deal with multiple libraries with their individual practices and policies. Most large universities have several individual libraries, some being totally autonomous. Yet libraries recognized the value in a single catalog showing all the library holdings. At that time, other automation systems did not offer that flexibility.”
“In terms of assessing NOTIS, initially we focused on understanding, acquiring and installing it as it existed, other than a few changes to accommodate local requirements and creating software to load the records from our OCLC tapes. The joint or shared developments came later.”
This (page 1) is a Dec. 16, 1980, letter from John McGowan describing a 30-day “trial use” contract with UF.
A memo in the NUL Archives from G. A. Harrer indicates the signing of a trial use agreement (“Confidential Disclosure Agreement for Computer Software”) on Feb. 4, 1981.
This (pages 2-3) is a May 19, 1981, letter from G.A. Harrer saying that the purchase of NOTIS had been approved by the Governor’s Cabinet.
In April, 1981, Mark Hinnebusch started the conversion of the NOTIS software from running on IBM’s VSE operating system only to run on MVS.
Mark notes: “I started the conversion in April [1981], commuting from Tampa. In June, I moved to Gainesville and continued the conversion…. I don't remember when we implemented but I do remember an anecdote. When we first put out two public terminals, they were in the lobby of Library West. I was going out to lunch and noticed both terminals in use, one by a Hare Krishna, in robes. I thought the juxtaposition interesting. If I remember right, we put terminals in cataloging and acquisitions first,… then OPAC, and, finally, circulation. “
Nolan says: “While implementing the existing modules (cataloging, acquisitions and OPAC) we worked on circulation requirements for the new module that Northwestern was developing, and Richard Bennett did much of that work. It was implemented in 1986/87. About 1990, the FCLA created keyword indexes and enhanced the OPAC search capabilities.”
In 1985, FCLA (Florida Center for Library Automation), under the leadership of Jim Corey, Michele (Dalehite) Newberry, and Mark Hinnebusch, was formed, and the other universities’ libraries joined UF in using NOTIS. This is a chronology of the FCLA/NOTIS history .
Note in particular the “New Circulation” entries (also described above). Nolan and Richard (Bennett) played an important role in that. And the “MFHL” entries in 1988/9. FCLA created the programs for this major “MHLD” function which were then incorporated by NSI into the distributed package.
Notes on indexing. [i]
[i] Gerald Snyder writes: “My memory about the index updates is that, in the early versions of NOTIS the index changes were not complete. Only some of the primary fields were updated. There was a more technical problem with the indexes and the way the entries were compressed into the index blocks. As entries were added the blocks were split. The way the blocks were keyed, there was a limit to the number of times the block could be split, then the index had to be regenned. At FCLA, we did some work to improve the way block splitting was handled, so that limit would not be reached. We also did work with deleting entries, adding additional index types. I’m sure NOTIS was also making changes to make the index updates more complete as the MHI index was introduced.
“The general attitude was to just run the index regen programs once per week or so. At FCLA, with all the universities and somewhat larger bib files, running those index generation programs frequently had some problems fitting into our overnight batch update window. I think we were probably more motivated to make the dynamic index updates better than many of the other NOTIS installations. Some of the FCLA index code went back into NOTIS.”
Jim Aagaard notes: “We modeled the structure of the MHI after the structure of IBM’s VSAM (they probably could have sued us for patent infringement!) but not quite as elaborately. So particularly if there were a lot of additions that generated index entries close together, it was necessary to regenerate (or reload) the index to restore the free space. It should have been possible to do this by just reading out the index to a tape and then reloading it (without going through all the bib records and actually recreating the entries) but I don’t remember if we did it that way. I seem to remember that a similar procedure was desirable for performance reasons (if not actually required) for normal VSAM files.