# Re: armstrong's axioms: augmentation - help pls

Date: Fri, 3 Dec 2004 13:49:40 -0500

Message-ID: <31bqq2F3aeh4aU1_at_individual.net>

"Jan Hidders" <jan.hidders_at_REMOVETHIS.pandora.be> wrote in message
news:nm2sd.8939$g94.550005_at_phobos.telenet-ops.be...

*> loveboat wrote:
*

> > hi,

*> >
**> > in armstrongs axioms, i don't get the augmentation, which is:
**> > (X -> Y) => (XZ -> YZ)
**> >
**> > why is that? i mean when Y is functional depending from X, i can't just
*

add

> > another set of attributes and make them dependent as well? (at least

that's

*> > what i think)
*

> > why can i do that?

*>
**> That's actually quite simple. Try a concrete example. Suppose that X =
**> "name" and Y = "age", then the dependency X -> Y means that if I give
**> you a name then there is only one corresponding age. Let's assume for
**> example that "Fred" has age 27 and Mary has age 29.
**>
**> If Z = "job" then XZ -> YZ means that for every pair of name and job
**> that I give you there is only one corresponding pair of age and job.
**> This is rather obvious because if I give you the combination "Fred" and
**> "secretary" then you know that there is a unique corresponding pair "27"
**> and "secretary"; the first part ("27") is unique because that is
**> determined by the value "Fred" and the second part is unique because it
**> can only be the value "secretary" that was already given in the first
*

pair.

*>
**> See?
**>
**> -- Jan Hidders
*

Or, how about it's just like algebra. If you do something to one side of an equation, as long as you do it to the other side too, the equation remains equal. Received on Fri Dec 03 2004 - 19:49:40 CET