Very Computer

I am involved with a project (data base of recent law cases in New Zealand)
at the moment there are aprox 7000 cases from the last 4 years.
The database contains identification information about the hearing,
(name of parties, date, court, judge etc.), a very brief summary of the
judgment, and a keyword field (eg. property law).

The data base is a comercial product and is in use by several of the
larger law firms as well as the Police.

There are approx 300 new cases added and 300 corrections per month.

We are currently using INMAGIC software ( under MSDOS) to index and search
the database. The summary of the case is completely indexed. I.e. one
can search for the occurance of any word in the body of the summary and
key fields.

INMAGIC has several undesirable features:
1/ Indexing is extreamly slow, and speed depends on the number of items in the
   database. It now takes 2 minutes to add an item on a 12MHz AT.
2/ There is no journaling mechanism to allow us to record changes to records.
3/ Threre are short commings in its searching abilities. In paricular it can
   only search one database at a time.
4/ It does not run on Mac. (And some lawyer love their Macs, Yes some lawyers
   do have some sense!! :-))

Items 1 & 2 realy hurt because it means that the only practical way to supply
monthly updates to clients is to ship them a DOS BACKUP
of the entire 10MB of the database.
(It would take 20 hours to load the update if we distributed the updates as
ascii text to be read in.)

To the point, finally. Is there anybody out there who uses INMAGIC who has
found way of addressing these problems? (We have had a look at 7.2 and after
initial hope were disappointed.)
Does anybondy know of other dbs that can handle indexing of text fields.
(We Know about BRS and in fact have a couple of clients that use it but it
is too expensive for most of our clients, besides it runs under Xenix.)

Russell Fulton, Computer Centre, University of Auckland.
--

Russell Fulton, Computer Centre, University of Auckland, New Zealand.

Quote:> (We Know about BRS and in fact have a couple of clients that use it but it
> is too expensive for most of our clients, besides it runs under Xenix.)

We have BRS at the law firm I work for.  Yucky Pooh.  I'm not impress
by it at all.

You have to use dot commands that looks like troff in order for BRS to
find something.

They do have a DOS version I'm told.

--
Tony Holden

I have had a brief look at BROWSE (from the people who bring you PCWRITE),
a shareware program which also introduces elements of hypertext into flat
file searching.  I'm looking forward to some free time for further
exploration.

There seems to be quite a lot of interest in full-text databases, or
full-text retrieval systems. Unfortunately most of the readily-available
database `managers', particularly for DOS, assume a finite amount of data,
or have address maxima. Depending on what is required, here are some other
pointers:

1. Text Retrieval

If your need is text retrieval, rather than a full database with data
dictionary and all the t*s, it's quite reasonable to go with
a text retrieval system. A number for UNIX are reviewed in the August
1990 (I think - it might be September) issue of UNIXWorld. These
include some quite nifty-looking products, although I haven't yet had
the opportunity to try any of them. Rather brief reviews, and some
RSN vapourware.

Text retrieval systems enable you (or anyone else) to create text files
in any way at all, then use the text retriever to select the elements of
the files which are of interest to you. Sort of like a humungous grep(1)
with context.

A DOS product which I have recently been playing with, and have been very
impressed by, is called ISYS, created by the Sydney (Australia) company
Odyssey Developments. It creates a separate index file, which is updated
for any documents which change. The advantage is that users of files, and
those who tend to create them, such as secretaries, are not affected at all
by the presence of ISYS. It's completely separate to the document-creating
software. It's also possible to have ISYS as a TSR, and cut and paste between
several documents using your preferred word processor with hot keys. After
trying out a demo copy, I'll definitely be buying it as soon as the pocket
permits.

2.  Full-text database systems.

I know of InMagic, but I've only had a fiddle with it. Another FTDBMS
which I have acquired is CDS/ISIS, from the Division of the UNESCO Library,
Archives and Documentation Services. I haven't tried this yet, either,
but the manuals are appallingly written in a kind of UNESCO-speak which
looks fairly cryptic. Nevertheless, it may be a goer for you to try.

Just hoping to contribute to that commonweal of knowledge.