Re: object algebra
Date: Thu, 26 Feb 2004 03:50:07 GMT
Message-ID: <P1e%b.59961$4o.81635_at_attbi_s52>
"Neo" <neo55592_at_hotmail.com> wrote in message news:4b45d3ad.0402251424.27581264_at_posting.google.com...
> > > Reality doesn't always provide data that fits "rectangularish" relations.
> > > And when it doesn't, RDM needs NULLs, as Codd has correctly recognized.
> >
> > I also find odd the idea that "reality ... provide[s] data."
>
> It was a figurative expression meaning reality can be the source for
> providing values in a relation's tuples. Reality may provide a person
> with two eyes and a person with no eyes. In RDM, the rules say define
> a relation's header at time X. RDM's rules forces us to presuppose
> that the attributes of future tuples will fit the definition at time
> X.
I don't see any room in this explanation for the role that ALTER TABLE plays.
It seems to me a more accurate description would be to say that we are expected to model the data as best we understand it now, and we are given the tools to change our understanding in the future.
In SQL's case, these tools aren't great, but they are sufficient. I think better tools for schema migration is a great opportunity.
Marshall Received on Thu Feb 26 2004 - 04:50:07 CET