NetNews Usenet Archive 1992 #31/NN_1992_31.iso/spool/comp/database/ingres/2135

Internet Message Format | 1992-12-22 | 1.6 KB

Path: sparky!uunet!gatech!emory!wupost!zaphod.mps.ohio-state.edu!darwin.sura.net!cs.ucf.edu!tarpit!ge-dab!cho006.cho.ge.com!judy_k
From: [e-mail]
Newsgroups: comp.databases.ingres
Subject: Code management - how much is enough?
Message-ID: <[e-mail]>
Date: 22 Dec 92 10:21:09 EDT
Organization: GE Fanuc Automation, North America
Lines: 23

A debate on database development and code management methodology has occurred
in the midst of a database project here and I am looking for information or
experiences along these lines.

Our application is targeted for a unix workstation with 4gl user interfaces.
We have had four developers working on this for the last few months. It happens
that we have varied backgrounds, 2 working on 4gl, 1 on 3gl (interfacing to
other systems with esqlc and cron) and the other working closely with the
end-users on the design and implementation details. The end-users are internal
here in our company.

The debate is on how much code management we should do and to what extreme it
should be done. The one side insists we break every element out into scripts
to be managed with rcs (or even use a networked vax and cms) in an elaborate
system of alpha, beta and large directory structures for each-- essentially,
an exploded copy of everything for every release. The other side while not
opposed to the release structures is not in favor of the explosion-- rather,
wants only to unloaddb and exportapp as rollback points and views the
explosion as an unnecessary burden.
---
  Ken Judy  ><>
  GE Fanuc  Charlottesville, VA
  [e-mail]