It looked to me like you had done some programming to customize your local on-line
catalog. Is that right?
Gary Strawn:
Yes, quite a bit of programming.
Ingrid Miflin:
How did you deal with the vendor and get access to the source code?
Gary Strawn:
NOTIS is a different system than most others. There are two things
you need to know about NOTIS: 1) the "N" in NOTIS stands for
"Northwestern"—it was actually our system, which we sold to
Ameritech— and 2) you get the entire source code when you buy a
licence for the software. So all you have to know is assembly language
and you can modify anything you want.
Ingrid Miflin:
But what about Voyager, your new system?
Gary Strawn:
Voyager is a completely different matter. We don't get any source
code; we don't get any help at all, basically. They did provide us
with a map of the database layout and that's about as far as they're
willing to go. That's more typical of other client-server systems.
John Attig:
You've basically done this with Visual Basic or the macro
programming language of your telnet client?
Gary Strawn:
Well, there's a heavy amount of that, to be certain, but there are
times when we have to modify the mainframe system to make a particular
operation work better.
Ingrid Miflin:
So did I understand you that you couldn't do this kind of special
programming in the new system—that you did it with Visual Basic?
Gary Strawn:
What I said was that the vendor wasn't making it easy for us to do
this. We have every intent of continuing our authority and
bibliographic loader program in some way in the next system. We just
don't know what that's going to be, yet. If you ask me in a month I'll
know.
Nancy Morin:
So I am I right in thinking that not all libraries have the ability
to do this 4:30-in-the-morning, upload-download, business, and is there
another way that they can participate in the system?
Gary Strawn:
Maybe I didn't make that clear. (Obviously I didn't make that
clear.) All of the contribution is done through one of the
bibliographic utilities. A typical local library doesn't put out a
file for Library of Congress. They put their new and revised authority records
into OCLC, and OCLC makes them available to the Library of Congress.
The exceptions are really the National Library of Scotland and the
British Library, which are contributing records directly—as I
understand things.
Nancy Morin:
That still suggests though, that for everybody to be playing
they've all got to have the information moving back and forth pretty
regularly.
Gary Strawn:
Absolutely. The files have to be kept in sync.
Nancy Morin:
If we looked at the equivalent for museum collections, or for the
taxonomists sitting around doing their work, there's maybe a little
more unevenness in technical capability.
Laurel Jizba:
The thing that makes it all work is that the MARC flat file
structure can be imported and exported by all of our systems,
regardless of the brand of the system.
Gary Strawn:
Yes.
Joan Swanekamp:
This ability to contribute in an automated fashion is fairly new.
Before that we did paper contribution, and that is how most of us got
into this. We would create paper files and send them off, or fax them
off, and they would be re-keyed when they got to the Library of
Congress.
Gary Strawn:
And that was typically not because of a limitation on our end, but
the Library of Congress was saddled with a system that wasn't amenable
to this kind of exchange. It's only recently that they've acquired the
ability.
Ingrid Miflin:
Those of us who live far out in the country still contribute by
paper, and I think it's important that we continue to contribute in any
way that we can. It doesn't all have to be electronic.
Chris Thompson:
What was the volume figure of changes, per day?
Gary Strawn:
In the name authority side, we get on average 5,500 changes.
That's an average over the last four months. On the subject side, it's
much smaller; in the rage of 250-ish, I think, per week.
Walter Berendsohn:
You mentioned that you have a few inconsistencies; well, some
duplicate records, and some inconsistencies due to data input. How do
you handle these inconsistencies?
Gary Strawn:
Many of them are discovered on an "as encountered"-basis. At my
own institution, we have some programs that can comb through the entire
file and look for inconsistencies and we have we have a reputation for
reporting those vigorously, but there are only certain kinds of things
that we can find. If someone has created two different headings for
the same person, with variations so that there are two records, our
programs won't necessarily find that because there's no way to know
that "John Smith" and "John R. Smith" are the same person. There are
certain mechanical things that we can do, but there are many
intellectual things that we can't do. The subject file is, in general,
much cleaner than the name file for the very reason that all of the
work is done centrally.
Adam Schiff:
I just wanted to add to that that the Research Libraries Group runs
a program against the entire database as well and reports out
duplicates. Those are the cases where the exact same heading appears
on two different records or a cross-reference matches another heading
-- and that's not allowed to happen—and when that happens, those of
us that do the creation locally will get a nasty little e-mail message
saying: "There's a problem here, will you please fix it." The Library
of Congress acts as the intermediary for that.
Anon. (F1):
When submitting subject authority proposals are there guidelines,
and where can they be found? And also, could you comment on the use of
taxonomic terms in LCSH? [GS: "No."] OK, well then, perhaps John Mitchell
could comment on the use of common name versus Latin name. I believe that someone
said earlier that LCSH uses common name, but I know I have also seen
Latin names.
Gary Strawn:
They use whatever happens to be, or appears to be, the current name
for something. There's no principal that says they're always going to
use one or the other.
Anon. (F1):
So they don't have a guideline in place for using one or the other?
Oh, they do?
John Mitchell:
There are guidelines for submitting subject proposals, particularly
for taxonomic or organism names. They're located in the subject
cataloging manual "H-1332"; it's a memo, and just one of many in that
compendium of guidelines. There is an electronic version of the
subject cataloging manual that's available on the "Cataloger's
Desktop." Short of that, we do have guidelines on the Web, under the
PCC, and there's a brochure in the back if you'd like to have the URL.
The guidelines that address the formation of heading for taxonomic
names—I would just like to say, without getting into great detail,
that the Library of Congress prefers common usage over Latin, and when
there is a conflict—say among several common names—then we will use the
Latin. I should also say this policy is based on our users' needs.
John Riemer:
I was just going to point out that the reference to the subject
cataloging manual John mentioned is in the bibliography I put out on
the back table this morning.