[583] in Public-Access_Computer_Systems_Forum
NOTIS/MDAS, pt.2
daemon@ATHENA.MIT.EDU (Angella Lambrou. Health Sciences L)
Wed Jun 24 10:31:53 1992
Date: Wed, 24 Jun 1992 09:13:09 CDT
Reply-To: Public-Access Computer Systems Forum <PACS-L%UHUPVM1.BITNET@mitvma.mit.edu>
From: "Angella Lambrou. Health Sciences Library. McGill U" <CZSK@MUSICA.MCGILL.CA>
To: Multiple recipients of list PACS-L <PACS-L%UHUPVM1.BITNET@mitvma.mit.edu>
----------------------------Original message----------------------------
Part 2 of NOTIS/MDAS evaluation
----------*
From: ldavids@casbah.acns.nwu.edu (Lloyd Davidson)
We have MDAS at Northwestern with 5 years of MEDLINE
and EXAC (Expanded Academic Index) loaded. We allow
dial-up access to members of the NU community and have
printers attached to some terminals in the library.
Printing is limited to screen prints, not by choice but
because of the limitations of the current version of MDAS.
MDAS has a lot of problems. It is difficult to use, there are
too many inconsistencies (not allowing a truncation symbol in
Author, Title or Subject searches, for example, while requiring
one in Keyword searches), you are not allowed to download entire
sets, it is difficult to combine sets, sets are renumbered
after each search, and so forth. The arbitrary breaks
in the Guide screens also cause much confusion. I could be
more complete but I will simply say - Let the buyer beware.
Some faculty have given up and gone back to BRS After Dark.
Part of the disappointment comes from the inevitable comparisons
with online and CD-ROM products which generally have much
better search engines. In the Science library many of
our patrons are quite sophisticated searchers. Also,
whether because of our mainframe or the software I don't
know, many searches take 5 minutes or longer to complete, a
totally unacceptable length of time. Try a search like:
K HYPERTENSION.SU. WITH MJ.SU. before you decide.
The excuse for using MDAS was that we got the software for
free as part of the NOTIS sale. Definitely a mixed
blessing.
MDAS will be upgraded, of course, and should improve over
time. At some point they are planning on going to UNIX.
----------*
From: "SYSTEMS COORDINATOR ... 855-1958" <JMARME@ucs.indiana.edu>
It's my understanding that there are alternatives to the MDAS
module for use with NOTIS (but not sold by NOTIS) that allow
greater flexibility such as exploding MESHs. I would imagine that
these also take additional CPU cycles as well. Although we are a
NOTIS site, our automation librarian is reluctant to go with MDAS
at this time.
----------*
From: Fred Bogin <fdbbc%cunyvm.bitnet@utcs.utoronto.ca>
We (the City University of New York) have been a NOTIS site since
1988. We loaded MDAS in January of this year--Readers' Guide,
Social Sciences Index, Humanities Index, General Science Index.
We don't have any of them on CD-ROM (only ERIC, PsychLIT, MLA, and
MEDLINE). No printing or downloading here--we use diskless IBM
terminals with no printers. Patrons can dial in from home and
print or download from there. User reaction seems pretty favor-
able. The trickiest thing is going back and forth from the OPAC
to the indexes which requires typing DPAC for the OPAC or DWIL
for the indexes. Also a lot of people miss the holdings infor-
mation. Good luck.
----------*
From: "Jerry V. Caswell" <LB.JVC@ISUMVS>
Iowa State University installed MDAS last winter so that the
library could provide the campus with access to heavily used
bibliographic databases. We first put up AGRICOLA, which is
consistent with our land grant status and the large amount of
agricultural research conducted at the university. We expect to
mount 3-4 additional databases, all of which would be heavily
used or closely tied to the university's instructional/research
mission: ERIC, a general undergrad index (whether Wilson, IAC, or
UMI has not been determined), a newspaper index, and possibly
MEDLINE or a business index.
In addition, we are building a CD-ROM network that will be
accessible campus wide. We have had six Wilson databases available
for the past five months. By next year we hope to have twenty or
more discs online over the campus TCP/IP backbone to Novell
networks that are located in all colleges.
---------*
From: Ed Holtum 319/335-9871 <CADEDHTS@UIAMVS.BITNET>
While the University of Iowa Libraries has mounted 5 Wilson
databases on MDAS, we decided sometime ago to opt for another
system for our Medline needs. We did experiment with Medline/
MDAS and even had some public demonstrations but both the potential
users and the librarians recognized the important limitations which
MDAS would have placed on the use of Medline on the MDAS platform.
The most important of these were:
1. impractibility of loading more than 4 or 5 years of data (to
some extent a local limitation)
2. The inability to explode
3. Difficulty in attaching and using subheadings effectively and
in displaying same
4. Lack of ability to store search strategies
5. Lack of ability to quick limit searches to human and/or
English
6. Lack of ability to perform continous printing at all
terminals
7. Lack of ability to capture continuous set of citations with
one operation
8. Lack of ability to display user-specified fields
9. Misc. other factors
We are currently attempting to muster financial support for the
Plusnet system and, in the meantime, we are making Medline
available via Compact Cambridge (non-networked) and online mediated
searching (the "old-fashioned" option).
----------*
From: Stephanie Allen <SNALLE01@UKCC.uky.edu>
We have had MEDLINE on MDAS for about a year as well as several
years of ERIC. Although I like the search software for the OPAC
and even for ERIC (and for other databases under consideration), it
does not work well for MEDLINE. First of all, the techies were
appalled at the enormous amount of storage space MEDLINE requires;
the 2 yrs. + current package required about 6 gigabytes. The
entire ERIC file wouldn't have taken quite that much room. We were
the first institution to purchase the NOTIS loader and mount/index
MEDLINE ourselves; the computer people have been very good about
this, but again, it is a time-consuming and labor-intensive
chore for such a large database. Vanderbilt offers additional
years, but the older years are in a separate file so that the
entire database does not have to be re-indexed each month; we are
also considering doing that if we mount an additional 2 yrs. While
users are thrilled to have dial-up free access to MEDLINE, they
have several complaints: 2 yrs. +current is not enough, there is
no way to limit to English language, lack of the explode capability
is more than an inconvenience (although most endusers, of course,
don't use that feature anyway - not knowingly, at least), the lack
of direct printing/downloading is a MAJOR drawback, hooks to
holdings hasn't been working properly and may not ever with some
databases (ERIC, for example, for records before the past couple of
years), and there is no "active" or "value-added" search
capability, i.e., natural language terms are not mapped to MeSH,
logical operators are not automatically supplied (our default
operator is the positional operator "with," but you could choose a
logical operator if you wanted to).
Users who are familiar with more user-friendly frontends such as
Grateful Med or COLLEAGUE or even FirstSearch/EasyNet are
frustrated by the lack of many of these features. To save storage
space and mount an additional year of MEDLINE, we ultimately chose
to limit to the English-language subset only (which solves the
inability of the system to process the limit-to-English command -
other languages are, of course, searchable) and to remove the
author, title, and subject indexes; we now have keyword searching
only. Since any search term can be qualified to a certain
field(s), and because only keyword search statements can later be
combined (our patrons NEVER believed that one!), we felt that the
A=, T=, and SM= statements only confused the users in MDAS and
actually interfered with their searches. The only problem with
taking down these indexes has been with saved searches; search
statements containing those searches (from OPAC) cannot then be
re-executed in MDAS, but that almost never happens. We have
changed several technical things, as well, based on timed response
testing we did in-house; response time is fairly good except with
the convoluted way you must make descriptors "major;" this process
seems much too time-consuming.
Hope this helps; I didn't post to the list because I thought you
would probably receive a considerable duplication of info. I'm
currently working on a proposal to install PlusNet on a campus LAN;
I think that this version of MEDLINE (all years, all languages, all
in under 3 gigs of storage space with many user-friendly features)
is much more useable for the average patron. We are considering
PYSCINFO, INFOTRAC, and other databases for MDAS. Again, once
printing/downloading problems are solved, I think MDAS is an
appropriate platform for most databases; MEDLINE is the exception.
I developed a search brochure if you are interested.
