RE: Roll Back dont roll back counters

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Yes, of course. But you have to have pretty stringent performance
requirements to justify changing to technical IDs as a matter of
rule. Reading an ER model is pretty important too, from a
maintenance perspective, though. I've been involved in doing that
for a while now, and trying to read business meaning into some of the
data models that I've worked with is pretty near impossible, which
makes the maintenance designer's job really difficult. You have to
speak to all the original people involved in the system design, and
sometimes that is worse than simply redoing the whole thing.

Anyway, a discourse on ER design is a pretty subjective area, which
we should probably not indulge in at present ;-)

MikeA

>> -----Original Message-----
>> From: Bruno Wolff III [mailto:bruno(at)wolff(dot)to]
>> Sent: 17 August 2001 13:09
>> To: Michael Ansley (UK)
>> Cc: pgsql-general(at)postgresql(dot)org
>> Subject: Re: Roll Back dont roll back counters
>>
>>
>> On Fri, Aug 17, 2001 at 11:37:33AM +0100,
>> "Michael Ansley (UK)"
>> <Michael(dot)Ansley(at)intec-telecom-systems(dot)com> wrote:
>> >
>> > Um, no, primary keys should not always be opaque.
>> Technical primary
>> > keys should always be opaque. Relational modelling has
>> been built to
>>
>> I think it has more to do with efficiency and whether or not you
>> ever expect the primary key values to change, then how hard it is
>> to read an
>> ER diagram.
>>

-----BEGIN PGP SIGNATURE-----
Version: PGPfreeware 6.5.3 for non-commercial use <http://www.pgp.com>

iQA/AwUBO30TzHympNV/C086EQJaAwCg2bh+rZOH1vMlUy6a42N4T15AJ5EAoKWz
Q1d0jSw6YyxcALUhhktFWb9E
=xJ4l
-----END PGP SIGNATURE-----