database management systems, cont. (48)

Thu, 30 Mar 89 21:01:08 EST

Humanist Mailing List, Vol. 2, No. 786. Thursday, 30 Mar 1989.

Date: 30 March 1989, 09:06:29 EST
Subject: Databases

Sebastian makes a good point regarding what can be done with an XT.
We inevitably have to make do with whatever hardware and software we
have immediately available. In many cases an XT will work quite nicely
even for some larger databases.

But --
I still think some caution, caveats, etc. are needed. In such a
case the database would probably have to be very well designed, if the
performance is to be acceptable. Designing a database is not horrendously
difficult, but it does require some understanding of the potential
problems, and some hands-on experience is to be recommended before
attacking a research project. I have worked myself with a project
that came into my hands with the `physical data' arranged in such a
way that the only effecient way of doing the data entry was to adopt
a database design that is rather less than perfect. This has created
a variety of problems that will eventually be solved only when the
data entry is complete and the database can then be redesigned.

I was also working on the assumption that an AT class machine
is now readily affordable by most academics. In Canada and the US
this is true, I think, but in Europe I understand that the cost
would be more prohibitive. A fast hard disk is still a little
expensive but not hopelessly so (ca. $1000 Canadian).

The performance that an AT can provide might be essential
in some circumstances. Imagine a project where the research
required constant querying of the database. If each query takes
even 3 minutes, the work would proceed very slowly. This has
actually happened to me in the past year and I was paying my
research assistants $15 per hour, 8 hours per day. Even smaller
time losses can grow to be significant.

Bob Sinkewicz