Re: Does Codd's view of a relational database differ from that ofDate&Darwin?[M.Gittens]
Date: 9 Jul 2005 09:07:09 -0700
I don't really see the value of the whole possrep thing. The most basic example from TTM wouldn't work very well, because point(x,y) and point(theta, r) *don't* have the same sets of values except in some uncomputable platonic ideal universe. You cannot give a precise conversion of point(1,1) to point(theta, r). In addition, you have to write your conversions routines by hand anyway, so you don't save anything there.
I just don't see what it buys you.
Tagged unions, a la ML, strike me as quite useful, however.
Marshall Received on Sat Jul 09 2005 - 18:07:09 CEST