Re: DB clasical structure violation
From: Paul Vernon <paul.vernon_at_ukk.ibmm.comm>
Date: Thu, 27 Jun 2002 14:54:52 +0100
Message-ID: <aff5nj$qri$1_at_sp15at20.hursley.ibm.com>
Date: Thu, 27 Jun 2002 14:54:52 +0100
Message-ID: <aff5nj$qri$1_at_sp15at20.hursley.ibm.com>
> Yes, I have. The design principle is: 'Avoid duplication of volatile
>information' (Edsger W. Dijkstra). In a datawarehouse the data is
>not volatile.
Nice principle, I'm glad my CPU ignores it (1st, 2nd, 3rd level memory caches ?)
But more to the point, the Relational Model is described in terms of relation values, which happen to be 'not volatile' (unlike relation variables). So, even if in a datawarehouse the data is _mostly_ not volatile, that is no reason to plain ignore the lessons of the Relational Model.
Regards
Paul Vernon
Business Intelligence, IBM Global Services
Received on Thu Jun 27 2002 - 15:54:52 CEST