Robert Bernier wrote:
> Didn't somebody mentioned some time ago that there was a postgresql
> benchtest suite that could be downloaded and anybody could use it
> against other engines?
I'm sure there is, but performance is only one part of the puzzle. In
fact, PostgreSQL is so fast and flexible it is relatively easy for me to
demonstrate that. The issues which are not so easy to demonstrate are
reliability, administrative benefits, and quality of commercial support.
Right now I am trying to convince my boss to migrate our COBOL
accounting/factory management app off of an ISAM proprietary filesystem
unto PostgreSQL utilizing a hand-coded ISAM driver. This is a big step
for us and my boss has serious (and legitimate) concerns about the
hidden costs and dangers of such a move. By far, the biggest factor is
uptime and disaster recovery. My pitch is for them to pick up a bronze
level contract support from PostgreSQL, Inc. (are there any testimonials
from psql.inc customers?)and see where that takes them. This is more
than offset by the saved cost of the Oracle/MS SQL license and the
proprietary COBOL driver for those databases.
As the in house 'PostgreSQL guy', I'm fairly aware of these issues, but
I'm having a hard time convincing my boss (we are a Microsoft shop) that
open source is 'the way'. Some reputable studies with a detailed
technical and cost analysis would be extremely helpful in this regard.
If and when we adopt PostgreSQL, I plan to write a detailed study about
the cost and benefits of using the database as a backend for COBOL apps.
From my point of view, as a former SQL programmer who had to take a
COBOL development position to get work, this is a fascinating and
exciting project.
Merlin
