Re: Upgrade as Change in Meta-state

From: Kenneth Downs <firstinit.lastname_at_lastnameplusfam.net>
Date: Thu, 21 Oct 2004 20:00:04 -0400
Message-ID: <4ii9lc.n3s.ln@mercury.downsfam.net>

Tony Andrews wrote:

> Kenneth Downs wrote:

>> Here is something I do not see discussed often, or else I have missed

> it.

>>
>> Is there much interest in discussing the upgrade operation on an RDB?

>
> <SNIP>
>
> Sounds rather like a case for CASE to me. The project I currently work
> on (maintaining a complex product that is used by many customers) uses
> the Oracle CASE tool called Designer to define the database and rules,
> with much of the code (data entry screens) being generated directly
> from the CASE tool.
>

Right, but does Oracle build MS SQL databases also?

>> Consider point #3.  Consider a system under active development, with

> a dozen

>> or so programmers pushing changes through.  Some of them are steadily
>> altering several large tables.

>
> That wouldn't happen here. Programmers don't push database changes
> through: database changes have to be routed via the database designers
> - there is a formal process. The changes are made in the CASE system,
> which can then generate the "alter table" DDL by comparing the new
> definition with the old.

Somebody with power pushes the changes through. I've seen shops where its the programmers, which leads to certain predictable problems, and other shops where its the DBAs, which leads to other predictable problems.

But I have a question, is the CASE tool running at the customer site? What happens if the customer has made changes that will cause the DDL to fail?

-- 
Kenneth Downs
Use first initial plus last name at last name plus literal "fam.net" to
email me

Received on Thu Oct 21 2004 - 19:00:04 CDT