Questionnaire responses from former NUL (/ISDO) staff members
Note: the views and opinions expressed in these questionnaire responses are those of the respondees only and are not intended to represent those of Northwestern University or the website proprietor. Complete disclaimer .
*= Photo under link below
Contents
Velma Veneziano* (see Interviews)
Photo of Jim Aagaard, Velma Veneziano, John McGowan
Note: This page includes people who were NUL (/ISDO) employees but never NOTIS Office, NOTIS Systems Inc., or Ameritech Library Services employees. The people who were such employees can be found in the “Questionnaire responses from former NOTIS staff members” section of the website.
.
Adele Combs (see Interviews)
Dave Anderson
(ISDO Programmer)
Dave’s answers to questions posed to him by email by Jerry Specht about the NUL timeline….
September, 1987: NUL implemented a batch accounting system. It was even called an interim system. Velma wrote the specs and Dave Anderson did the initial coding. It gathered character string data from the order record. This system was used through August, 1992. There were modifications and enhancements. The output consisted of paper vouchers and management reports. We talked about adding on-line capabilities and also having a tape output for data transfer. The announced move to the commercial version put an end to those plans.
March, 1988: Circulation Statistics reports were rolled out. The circ history file was read and the data munged. Another program was written to provide a list of items charged. There was an input parameter to allow for a list of items for a particular patron category. There were other reports written, but they escape me. These programs were written by Dave Anderson.
1988: Sue Rosko joined the staff and she made enhancements to the on-line program code. She even had to re-engineer these changes for s software upgrade. Sue also designed and implemented a course reserve system.
1988/1989: The library began to feel pressure to migrate to the commercial version of NOTIS. My understanding is that Arnold Weber saw the MDAS product, including Medline, and thought what NUL should have it. Initially, the library was ordered to convert of one fell swoop. After negotiations, we were permitted to convert in two steps.
December, 1990: Work began of migrate the library's version of NOTIS, I think it was called version 4.3. The migration was from an IBM 4341 VSE machine in the library to an IBM 370 MVS machine at University Computer Center. Tests were made with the on-line system and match jobs needed to be converted to MVS/JCL and tested.
July, 1991: On July 3, the old machine went off-line and the data was exported. On July 4, the data was uploaded to the new machine. On July 5, the system was brought up successfully.
1991/1992: Work began on the conversion to the commercial version of NOTIS, which was the 5.0 release. File conversion programs were written, on-line testing and batch jobs were done. Sue Rosko was responsible for the on0line and Dave was responsible for the batch. The production conversion was implemented on Labor Day weekend of 1992. Apart from local modifications, we stayed with this version until we migrated to Voyager.
September, 1992: NUL started using the NOFA and Bill-And-Fine modules. We ran a Fiscal Period Close for the first time.
October, 1992: NUL implemented the Keyword/Boolean feature. Work began on procurring the MDAS products of Medline, EXAC and ABI/Inform.
1993/1994: Sue Rosko had on-line projects.
January, 1993: Work began on writing an interface to transfer NOFA voucher data into the University's accounting system, namely CUFS. Dave Anderson did the analysis and wrote the code.
1994: The university made mention of phasing out the mainframe. As part of that they invited Fretwell-Downing to present their system.
1995: NUL began the process of phasing out the MDAS products. Some subscription were moved the the vendor's server and others migrated to a local UNIX server.
1995: The university announced the the library should select a client-server system. This should be implemented before the end of the century.
1996: A committee was formed named the NOTIS Replacement Task Force (NRTF). Members of the committee included Bill Parod and Scott Swanson. They were given the task of coming up with a "short list". This list consisted of four vendors. The three I remember were Voyager, VTLS and Horizon. I don't remember number four, which might have been Sirsi or III. VTLS pulled out soon afterward. Therefore, the group made a recommendation for Voyager.
October, 1997: Scott Swanson installed version 97.1 as our test server. The server was running AIX.
March, 1998: Work began on writing conversion programs to convert NOTIS data into the appropriate Voyager SIF formats.
August, 1998: The library brought up Voyager 97.2 as their production system, and it stayed up.
September, 1998: The NOTIS system still ran on the mainframe, but only the LUIS OPAC was available to the public. New acquisitions weren't being added, of course. Other users of the mainframe desired the library's DASD space, so NOTIS was shut down in March, 1999.
<end Dave Anderson>
Gary Strawn
[Authorities Librarian, Northwestern University; creator of Cataloger’s Toolkit]
<<js When did NUL start using RLIN? Did NUL use GTO or did they get records from RLIN in some other way? When did you stop loading LC-MARC (bib record) tapes? (I suspect that continued for quite a while.)
I have no idea about when NUL started using RLIN; they were already using it when I got here in 1987. For a very long time, an army of student workers re-typed bibliographic and authority records printed from RLIN into NOTIS. (I clearly remember a row of I think 5 of those big NOTIS terminals just for students.) And for quite a while we were members of both OCLC and RLIN. (I think that one some minimal level we were members of RLIN until the end; but we had long since stopped sending them bibliographic updates.) (For authority records, NUL loaded the entire LCSH file in about 1985 (before me, anyway) and then loaded the weekly updates. We typed, and later downloaded directly, individual name authority records, but didn't start loading the entire LC names file until our big retrospective conversion project which was about 1990 or so, I think …)
This is all very murky, indeed. I can't say any more than that "at some point" we stopped using RLIN as our primary source of bibliographic records and started using OCLC. I'm sure we either initially or very soon after had some kind of transfer going, and I'm sort-of sure that we used GTO for at least some of this but I'm quite certain that we didn't use GTO for *all* cases (more below).
I'm really sorry about the fuzziness of it all, but that's my memory for you.
I'm slightly clearer about my own circumstances. I was hired in 1987 as the Authorities Librarian. One of the first big projects was helping test the final versions of the Merged Headings Index, and soon enough working with the enormous conflict and error detection reports that were spun off of the index (by LB360; see, I can remember something!). I think that probably my continuing questions and suggestions got Jim's attention, and I was eventually asked to move to ISDO. I got along, I thought, quite well, with everyone (I am aware of rumors that this was not always the case); but the stress of sitting at the desk typing away all day eventually got to be too much for my neck and wrists, and after several years I asked that I my job be redefined as half authorities librarian and half programmer, so I could have some variety of position throughout the day. (I eventually got a stand-up desk, which was the probable true resolution of the pain problem.)
Although I very much enjoyed (and still miss!) assembly-level programming, the real fun here started when we installed Windows 3.X and then began to use the HostExplorer terminal software for the NOTIS connection. That emulator had an interface that allowed another program to read and write screens; although writing programs that handled records indirectly through this interface was less efficient that would be the case with more direct contact, the trade-off is that development could be much more rapid.
And so back to record transfer: Though I don't have anything like details or documentation, I did happen to save several big piles of overhead transparencies from talks that I gave over the years about the cataloger's toolkit in its various evolutions, and related programs. One of the slides clearly shows buttons that could be used to download records from OCLC into NOTIS, and upload records from NOTIS to OCLC; this would have been via intermediate MARC files. For downloading from OCLC we had a locally-written screen-scraping program which produced a MARC file; this file was picked up by the toolkit and written line by line to the NOTIS emulator. (My memory is that we asked, and OCLC didn't have a problem with this avoidance of the fee as long as we were eventually attaching our holdings; the export fee was for the processing needed to generate the MARC output, not for the record itself.) I suspect that this process also allowed for the operator to be specific about which record got overlaid in Voyager, so there wouldn't have been what I dimly recall to have been problems with GTO finding the wrong matching record, or not finding any at all.
<<js Am I correct in understanding that there was never any such program for use with RLIN? Bib record creation from RLIN was never done any way except re-typing?
GLS: That's right: everything we got from RLIN we printed and re-typed.
>… the point at which it was decided that the library should start using the generic NOTIS product rather than continuing on a separate path [199?]
Do you have any idea when *that* was ? From what Velma says, it seems that there was a specific date when this happened.
GLS: At no point did we use out-of-the-box NOTIS. There came a point where we needed a bunch of stuff present in a later NOTIS version (was it 5.something? I've forgotten); so we prioritized our local changes and poured the critical ones (mostly having to do with MHI, cataloging and authority control, as I recall, but that may just be my memory) into the chosen NOTIS version and ran with that. (As I recall, in many cases this could be a whole-module swap, but there was also some line-by-line work. This was a biggish project.) And from then on we started to diverge again.
It seems that keyword/Boolean was never implemented at Northwestern as part of NOTIS, correct? But, I’m assuming that, when the library switched to Voyager, they did start offering the keyword/Boolean which came automatically as part of Voyager. Correct?
GLS: No, we did indeed implement KWB at Northwestern. I think this came with 5.whatever (5.2?), but I may be wrong about that. It was needed, as I recall, to implement searching of the local bibliographic database in parallel with (oh, this is really getting dim; help me, Harry!) one or more periodical indexing databases that were mounted locally.
**NUL Switch to Voyager**
From http://www.library.northwestern.edu/about/welcome/history-the-library-0 :
This change had become necessary since, by 1995, the university had mandated a departure from mainframe-based computer systems in order to embrace a more flexible client-server architecture for its systems campus-wide. It also had become clear within the library that a new system was necessary to more fully exploit developments in relational database design and to take advantage of data delivery in a communications environment that was increasingly networked—and gradually poised to exploit the potential of the World Wide Web. Bishop appointed a NOTIS Replacement Task Force in early 1996, chaired by music librarian Don L. Roberts. This task force returned a unanimous recommendation on behalf of the Voyager system, offered by Endeavor Information Systems, Inc.—ironically a product that was largely developed by engineers who had worked on NOTIS at Northwestern during the 1980s.
Does this seem like an accurate record of what happened?
GLS: I think that's fair. My memory is that we had several vendors in for whole-day (or even multi-day?) demonstrations. (At least one of these consisted only of slides showing what the product would look like were it to actually be developed.) My dim memory is that given the constraints there was only the one possible choice.
<<js Do you know what year NUL switched from NOTIS to Voyager?
GLS: That's an easy one, because it's the creation date in the record history of all of the older bib and authority records: July, 1998.
GLS: You may or
may not be interested to know that soon after the switch to Voyager I gathered
up all the NOTIS stuff that I could find, including all of the source code
printouts, and sent them to University Archives; they're still there. (As
I recall they were under some kind of access embargo—staff only—but I think I
set the time period to 10 years.) Archives has a finding list that I drew
up at the time. It may also amuse you to know that in my "ISBN"
folder here at my desk I have a print of the source for LB073TBL; still useful
when someone asks about inserting hyphens into ISBNs.
<end Gary Strawn>
Janet Swan Hill
1. When did you start working for NUL? How did this come about? (If your background prior to NUL is of interest, please include that as well.)
I came to Northwestern in 1978, following my husband, who had taken a job in Chicago. My previous position was Head of Cataloging at the Geography and Map Division of the Library of Congress. I came to the Library of Congress in 1970 as a "Special Recruit" (now called "Intern").
2. What was your job title, or, more generally, what did you do? Were there particular projects you were involved in?
I was Head of Cataloging at NUL. In addition to working on the ongoing development of NOTIS, we became on of the first three libraries to participate in the project that became NACO. we contributed authority records directly to the Library of Congress. Given the technology of the day, we actually had a staff member on our payroll who was stationed at LC. I was also part of the group that designed the first online catalog interface.
3. What were NOTIS’ best features? Any particular weaknesses?
NOTIS was ahead of the field in many areas, including serials control, remote self-checkout, and authority control. Especially notable was the integration of the functions. At that time, IF libraries had any of these functions operating, they were likely to be free-standing and not integrated with each other.
7. When and why did you leave?
I left in 1989 to take a job as Associate Director of Technical Services at the University of Colorado libraries in Boulder. (a position from which I retired in 2012)
9. Are there other things you think your fellow employees would be interested in knowing/remembering?
John McGowan (library director) deserves great credit for having the vision to investigate the creation of a system, and for then giving the creative team (Jim Aagaard and Velma Veneziano) their head. Karen Horny (AUL for Tech Services) deserves recognition for her leadership in Technical Services, and for encouraging all who worked in her division to take an active part in the testing of and influencing the form of NOTIS. Both John and Karen deserve special praise for their encouragement and recognition of people taking an active role in the profession at large, through service, research, and publication.
A memory: coming in to the office in the morning to find Velma cruising the department, just waiting for someone to talk to .... to try out an idea she had had overnight.
10. At NOTIS' peak, 50% of all ARL libraries were using it as their ILS. I believe this degree of dominance of the research library market has not been seen before or since. Why was NOTIS as successful as it was? Some possibilities:
a. NOTIS had functions other ILS’es lacked. (I can't think of any, but maybe you can?)
NOTIS had operational series control well ahead of other ILS's. NOTIS had authority control in advance of most others (ours was working before RLG's). NOTIS had self check-out VERY early.
b. Though other systems had the same module, the NOTIS version was better.... Were certain modules (Circulation? Acquisitions/Serials? Cataloging/Authorities? OPAC? the BRS keyword/Boolean? GTO? MDAS?) especially strong?
c. It's clear that NOTIS programs outperformed those of other library systems. (Someone who worked for a competitor has told me that in 1989 their sites were complaining that, while it took them more than a week to regenerate a particular index, the NOTIS programs were doing the same in 4 hours.) How much of this was due to the superiority of the IBM systems software/hardware the programs were running on and how much to the fact that the application programs themselves were written so efficiently?
My speculation is that it was because of the efficiently written programs and documentation. I would also propose that the systems were strong because they were thoroughly tested by operational staff who felt involved in the testing and ongoing evolution of the system. (We all felt we had a stake in it, because we were using it every day, and because we knew that our "resident geniuses -- Aagaard and Veneziano -- were paying attention. When I left Northwestern, I went to Colorado, where they were working with CARL, and was totally unprepared for the unreliability, opaqueness, slowness, and "unfriendliness" of the CARL system.
d. NOTIS marketers were very effective at identifying the institutions which would most benefit from NOTIS and concentrating their marketing efforts on those institutions.
Well, at first, other libraries came to US asking to purchase the system. There was no intent to market the system until other libraries came asking. Once we had sold the system to a few libraries, there developed a very close "family feel" among the early purchasers.
e. Most NOTIS MVS installations ran on a shared (IBM) campus computer. Except for DOBIS/LIBIS -- not really a factor in the US market -- no other library systems ran on IBM mainframes. As IBM's dominance of campus administrative computing increased, so did NOTIS' dominance of academic library computing.
f. Other factors in NOTIS' success?
The system was purposely written so that it could be maintained locally, and with minimal extra staff or funding. When NOTIS was being developed, other libraries developing systems (e.g. University of Chicago) were using grant funding (which NUL was not), and when the grant funding ran out, they had trouble keeping up .... since their business model had always included external funding.
g. What do you think is the relative importance of these different factors?
Least important: e and d. Most important: It's hard to distinguish among the remaining factors, since they are all so closely interrelated, and reflective of the early history of NOTIS' development.
11. Some people have argued that, if NOTIS had continued the slow-but-steady “word-of-mouth” campaign in progress before the big marketing push in 1984, though not as successful as it was in the broad library market, could still have been reasonably successful, and, given the fact that NOTIS Systems Inc. expenditures on development and marketing would have been so much lower, might have been almost as profitable. Do you think?
I agree with this assessment.
<end Janet Swan Hill>
Jim Aagaard* (see Interviews)
(Photo of Jim Aagaard, Velma Veneziano, John McGowan at bottom of page.)
Karen Horny*
1. When did you start working for NUL? How did this come about? (If your background prior to NUL is of interest, please include that as well.)
I started working at NUL straight out of Library School, July 1, 1966, as Assistant Core Librarian, soon sharing an office with Velma Veneziano.
2. What was your job title, or, more generally, what did you do? Were there particular projects you were involved in?
I was Head, Core Collection/ Undergraduate Services 1968-71, Then became Assistant University Librarian for Technical Services in 1971, three months before the initial NOTIS system was implemented on Oct. 7, 1971. I had become particularly interested in computer applications through conversations with Velma. Coffee breaks were exciting then and Betty Furlong and I actually came up with the acronym NOTS over coffee one day.
3. What were NOTIS’ best features? Any particular weaknesses?
It ran from day one!!! We could even do serials issue check in, although that wasn't so easy on 2740s, where you needed to retype entire lines to do updates. Jim Aagaard is such a genius that when he stopped by the Data Center to see how all was working, if someone realized it would be easier to do a particular task if something worked differently, Jim took note and, when we started up the next day, the change was already made!
7. When and why did you leave?
I left in 1995 to become Dean of Library Services at Missouri State University (which was still running NOTIS at that time).
9. Are there other things you think your fellow employees would be interested in knowing/remembering?
NOTIS WORKED wonderfully for both technical and public services. There wasn't the usual divide.
10. At NOTIS' peak, 50% of all ARL libraries were using it as their ILS. I believe this degree of dominance of the research library market has not been seen before or since. Why was NOTIS as successful as it was? Some possibilities:
a. NOTIS had functions other ILS’es lacked. (I can't think of any, but maybe you can?)
Because NOTIS was developed at and for a large academic library with lots of consultation between the systems analyst, the staff and the master programmer, it matched the needs beautifully.
b. Though other systems had the same module, the NOTIS version was better.... Were certain modules (Circulation? Acquisitions/Serials? Cataloging/Authorities? OPAC? the BRS keyword/Boolean? GTO? MDAS?) especially strong?
c. It's clear that NOTIS programs outperformed those of other library systems.
I think it was a combination of accessible IBM mainframes that were economically sharable and the incredibly efficient programming that Jim did.
d. NOTIS marketers were very effective at identifying the institutions which would most benefit from NOTIS and concentrating their marketing efforts on those institutions.
Academic libraries begged for the system before it was even available for purchase. That was the time when most academic libraries, especially those of middle to larger size became interested in adopting computer applications to enhance their operations. NOTIS practically sold itself.
e. Most NOTIS MVS installations ran on a shared (IBM) campus computer. Except for DOBIS/LIBIS -- not really a factor in the US market -- no other library systems ran on IBM mainframes. As IBM's dominance of campus administrative computing increased, so did NOTIS' dominance of academic library computing.
f. Other factors in NOTIS' success?
g. What do you think is the relative importance of these different factors?
See comments for question a. and d.
11. Some people have argued that, if NOTIS had continued the slow-but-steady “word-of-mouth” campaign in progress before the big marketing push in 1984, though not as successful as it was in the broad library market, could still have been reasonably successful, and, given the fact that NOTIS Systems Inc. expenditures on development and marketing would have been so much lower, might have been almost as profitable. Do you think?
This could well have been true.
<end Karen Horny>
Velma Veneziano* (see Interviews)
(Photo of Jim Aagaard, Velma Veneziano, John McGowan below.)
Photo of Jim Aagaard, Velma Veneziano, John McGowan
Jim (left), Velma (middle), John (right):
