Re: <bits OT> dbdebunk 'Quote of Week' comment
Date: Mon, 22 Aug 2005 09:00:14 GMT
Message-ID: <yugOe.105$FW1.61_at_newsread3.news.atl.earthlink.net>
"Frank_Hamersley" <terabite_at_isat.bigpond.com> wrote in message
news:HKZNe.6342$FA3.4435_at_news-server.bigpond.net.au...
> > Celko is right in his diatribe, but he lays it on too thick.
>
> Perhaps - he made the best point in another post (excuse my paraphrasing)
> that the problem is not the use of surrogate keys per se, but more the use
> of identity columns for PK's. I guess in the hands of the great unwashed
> there is not much distinction between the two acts, and religious sermons
> might prove to be the most effective way to try and hold back the tide.
>
Look, I'm in favor of education as much as the next man. A significant part
of my career was as an instructor for DEC.
And there's no question that about half the databases being built today
would be better built if the builders knew how to design a database at least
halfway decently. I'm not being a devil's advocate here. But religious
sermons hold back the tide about as well as king Canute's command did.
The deeper question, IMO, is whether databases are still being used as a tool for integrating data, and making it more widely accessible, or whether most of today's databases are private parts of an application, and there is no useful information to be obtained via SQL, knowing only the business model and the metadata. By "more widely accessible" I mean useable in more contexts, not just by more people.
Programmers will always gravitate towards viewing the data in "their databases" as their private bailiwick, and insist that users of the data access it through their own API. Learning SQL is certainly better than learning a hundred programmer's different APIs. Received on Mon Aug 22 2005 - 11:00:14 CEST
