David Cressey wrote:
> "Brian Selzer" <brian_at_selzer-software.com> wrote in message
> news:1iQEg.8919$1f6.2533_at_newssvr27.news.prodigy.net...
>
>
>>I think that there is a fundamental difference. Over the years I've
> worked
>>with many databases that used surrogates and many that didn't. In all
> cases
>>where surrogates were used, it was always assumed that surrogates could
> not
>>change.
>
> PMFJI. I'm offering a point of view that doesn't necessarily conform to
> either yours or Bob's.
>
> I also have worked with many databases, some that used surrogates, and some
> that
> used so called "natural" keys.
>
> The distinction in my mind, is whether "we" control the key or whether
> "they" control the key.
> If "we" control the key, then it will only get screwed up if "we" screw it
> up, and "we" aren't going to do that, are "we"?
>
> If, on the other hand, "they" control the key, then sooner or later they
> are going to screw it up. That's the way "they" are.
> If "they" screw it up, and "we" have a problem, then "we" have an
> intractable situation.
>
> I can give you real world examples of exactly this happening, but this
> comment is getting long enough.
You and I are in complete agreement. I note that the distinguishing
characteristic between natural keys and surrogates keys is familiarity,
and I note that all objections to natural keys relate to control.
Received on Thu Aug 17 2006 - 09:50:50 CDT
