# Re: cdt glossary [Graph] (was: what are keys and surrogates?)

Date: Sun, 13 Jan 2008 17:35:29 -0800 (PST)

Message-ID: <6c7030ec-8a37-4922-a0ea-4241916bd863_at_e23g2000prf.googlegroups.com>

On Jan 14, 6:57 am, JOG <j..._at_cs.nott.ac.uk> wrote:

> On Jan 13, 8:04 pm, David BL <davi..._at_iinet.net.au> wrote:

*>
**>
**>
**>
**>
**> > On Jan 13, 3:48 am, JOG <j..._at_cs.nott.ac.uk> wrote:
**>
**> > > On Jan 12, 8:14 am, David BL <davi..._at_iinet.net.au> wrote:
**>
**> > > > On Jan 12, 2:24 pm, JOG <j..._at_cs.nott.ac.uk> wrote:
**>
**> > > > > On Jan 12, 1:05 am, David BL <davi..._at_iinet.net.au> wrote:
**> > > > > > Really! I have seen a (mathematical) relation formally defined as a
**> > > > > > subset of a cartesian product (and not an ordered tuple) on many
**> > > > > > occasions.
**>
**> > > > > Bit confused by this - a cartesian product generates a set of ordered
**> > > > > tuples (over which a function is a subset), and all the hyperlinks you
**> > > > > listed seemed to follow that description.
**>
**> > > > Do you agree that most authors define a binary relation as a set of
**> > > > ordered pairs? In an earlier post you said a function is the ordered
**> > > > triple (D,C,G). How do you reconcile saying that a function is a
**> > > > (binary) relation?
**>
**> > > Relations are formally described by the ordered triple (D,C,G), but
**> > > are often informally described by just G.
**>
**> > So all those authors that define a binary relation as a set of ordered
**> > pairs are being informal? I don't agree with that.
**>
**> > Check out the section under formal definitions in
**>
**> > http://en.wikipedia.org/wiki/Relation_%28mathematics%29
**>
**> > Are you saying that definition 1 is informal?
**>
**> I don't see why you'd think so - in the article, first the domains,
**> x1...xn, are stated. Then the graph is specified as a subset of the
**> cartesian product of x1..xn. Seems relatively formal to me - domains
**> and a graph (albeit for an n-ary as opposed to a binary relation).
*

Definition 1 doesn't mention the words "graph" or "domains". Instead it says that a relation L over the sets X1,...,Xk is a subset of their cartesian product. It follows for example (with that formal definition) that all empty relations are equal and in general for a given relation it is not possible to uniquely determine a list of "domains".

I appreciate that you prefer definition 2 (and I agree it's better for developing a "theory of relations"). Received on Mon Jan 14 2008 - 02:35:29 CET