Re: object algebra

From: Eric Kaun <ekaun_at_yahoo.com>
Date: Tue, 24 Feb 2004 13:14:00 GMT
Message-ID: <s6I_b.17440$3C3.10580_at_newssvr31.news.prodigy.com>


"Neo" <neo55592_at_hotmail.com> wrote in message news:4b45d3ad.0402231444.2f77c935_at_posting.google.com...
> > > RDM's fundamental design ensures NULLs will occur in some
applications.
> >
> > No, it doesn't. Which applications are you talking about?
>
> In any application where a tuple doesn't have all the attributes
> defined by the relation header. There are many such tuples in
> existance today. See Chapter 20 "Missing Information" in Date's Intro
> to Db Sys.

No, there are no such tuples in existence today - they're not relational tuples, by definition! I just finished reading the book (8th edition, so perhaps things have changed) a few days ago, so I'm more than familiar with chapter 20. Unfortunately, I don't have it here with me.

In any event, a relation is a predicate. A tuple which doesn't "have all the attributes" defined by that predicate is NOT a proposition of that predicate! In short, you've opted out of saying something specific, and want to hand-wave about the other attributes. Which is fine, given adequate subtyping - you can model a predicate over a type, and have subtype values in attributes of tuples in the relation.

  • Eric
Received on Tue Feb 24 2004 - 14:14:00 CET

Original text of this message