Noons suggested that:
> ed.prochak_at_magicinterface.com (Ed prochak) wrote in message news:<4b5394b2.0410131031.49b21ed_at_posting.google.com>...
>
> ....
>
>>Why because that is not relationship between the Surrogate key value
>>and the data it is supposed to connect. It is often just a value
>
>
> This is where your examples and arguments are completely wrong.
> If you do not establish a referential integrity constraint
> between ANY two PK and FK keys (natural or not is COMPLETELY
> immaterial), then you WILL have the potential for wrong data.
>
> It has NOTHING to do with how those keys are generated/populated.
> The problem is NOT the use of surrogate or natural keys, the
> problem is the ABSENCE of a RI constraint!
>
> Can we for once stop invoking absolutely farcical and
> irrelevant scenarios that have NOTHING to do with use
> of surrogate versus natural keys? Since WHEN does the
> use of surrogate keys preclude the use of RI?
Only when you have your head so far up your #%$ that you don't want to
see obvious advantages of both natural and surrogate keys. You are
wasting your time and wisdom on someone clearly not willing to see the
big picture (or perhaps not capable).
Received on Thu Oct 21 2004 - 08:52:16 CDT
