Re: RM and abstract syntax trees

From: paul c <>
Date: Thu, 01 Nov 2007 15:36:32 GMT
Message-ID: <4omWi.167353$th2.6381_at_pd7urf3no>

Jonathan Leffler wrote:
> ... It may also contain the
> same value as a large number of other places in the database.

That is the beauty of the Information Principle. When you add a fact, the existing associations don't go away and reflect the new fact to boot. Likewise, no pointers to remove when you drop a fact.

This all depends on people recognizing that it is extremely useful to recognize the difference between "under the covers" implementation and the logical abstraction that users and apps see.

In another context Codd also said something that I think is related:

"In other words, the sharing of data requires the sharing of its meaning."

I would not want to try to explain to users, programmers or otherwise, that some of their data "points" to some of their other data (although I've seen it tried). Why should a user care? Received on Thu Nov 01 2007 - 16:36:32 CET

Original text of this message