Re: dbdebunk 'Quote of Week' comment

"Roy Hann" <specially_at_processed.almost.meat> wrote in message news:uc-dnWcwKdgVe5jeRVnyjg_at_pipex.net...
> "Frank_Hamersley" <terabite_at_isat.bigpond.com> wrote in message
> news:NclNe.4651$FA3.3258_at_news-server.bigpond.net.au...

> > I guess I should declare 27 years of involvement with gadgets starting
in
> > the age of PDP-8/11's, Altairs, Alpha Micro, CDC Cybers (what a machine
> that
> > was) might have taught me a few tricks in this regard. So all you pups
> > searching for gems should prolly accord 80 in an 80/20 measure of Joes
> > dissertation.
>
> I have been in this field just as long as you. (I still have a copy of
the
> MITS Altair newsletter with a Chirstmas photo of Bill Gates and the rest
of
> the Microsoft team--all 16 of them.)
>
So have I. In my reply, I mentioned the PDP-1.

> In that time I have seen/written/endured untold mountains of spaghetti
code
> do to foolish and misguided things, that had I known and followed the
advice
> of people like Date and Pascal (and maybe even Celko on occasion) I could
> have mostly avoided.

I've learned from my own mistakes, from the mistakes of others, and occasionally, from fundamental principles.

>
> If you were to explain to me how unnecessary surrogate keys silently
conceal
> corruption, and how you guaranteed that they wouldn't, and that it all
ended
> up being cheaper, simpler, and more maintainable than if you hadn't used
> them, then I might be prepared to accept that you haven't suffered any of
> the grief. But since you haven't explained that, I'm more inclined to
think
> you just haven't noticed or recognized it. Not noticing that you have
> cancer is not the same as not having it.
>
> It is not too late to explain now. I keep an open (but properly
sceptical)
> mind.

I suggest you read Ralph Kimball for a good treatment on the benefits of using surrogate keys.

As I said elsewhere in this thread, surrogate keys are not useful when natural keys will do the same job.

Incidentally, most "natural keys" are just as "artificial" as surrogate keys. They are simply assigned externally to the system being disucssed. Received on Sat Aug 20 2005 - 11:14:48 CEST