Re: Plural or singular table names
Date: Tue, 2 Sep 2003 11:25:35 -0400
Message-ID: <RN35b.377$hL1.39717997_at_mantis.golden.net>
"Paul Vernon" <paul.vernon_at_ukk.ibmm.comm> wrote in message
news:bj1o0m$1gla$1_at_gazette.almaden.ibm.com...
> "Bob Badour" <bbadour_at_golden.net> wrote in message
> news:cZO4b.317$8v7.33990779_at_mantis.golden.net...
> [snip]
> > > > From a theory standpoint, it's just a name. X and Y are as good as
any other
> > > > names.
> > >
> > > That depends on where you draw the theory line in the sand. If you are
interested
> > > only in say the mathamatics of relational algebra, then indeed the
names are not
> very
> > > important, but even then 'X' would be prefered to say
'XAS$%£^AF"sgrSFPI' by
> most.
> >
> > The point is the preference is psychological and not based on any
particular
> > theory. Different people will have different psychological needs leading
to
> > different preferences.
> >
> >
> > > If you prefer to not draw a line between the theory and the practical,
> > > then there is no reason why a 'theory of names' cannot be considered.
> >
> > I don't draw a line between theory and practice. I draw a line between
> > theory and biological empiricism.
> >
> > > Such a theory would surly be
> > > able to tell us that
> > >
> > > { Loves, Loved }
> > >
> > > would be better attribute names than
> > >
> > > { X, Y }
> >
> > Why?
> >
> >
> > > or
> > >
> > > { Cabbage, Potato }
> >
> > Why?
> >
> >
> > > and certain much better than
> > >
> > > { Hates, Hated }
> >
> > Why?
> >
> >
> > > for a tuple purporting to represent predicates of the form
> > >
> > > "Juliet loves Romeo"
> >
> > The predicate and not the name provides the semantic context. Consider
the
> > following relation:
>
> Others (some quite notable) would agree with you Bob, but I don't.
>
> My objection to this emphasis on predicates providing semantic contexts is
the
> implication (to me) that then they must therefore be *mandatory* for every
tuple in
> every database. For tuples that are the result of a query, either the
system must
> then derive the resulting predicate for us (which I know of no general
method for,
> especially in the face of user defined operators), or we need a way of
giving out own
> predicates for every query result!
I don't see how user-defined operators impede the derivation of a derived relation's predicate.
> If predicates are not mandatory, then how do we know what a given database
means?
Predicates are mandatory.
> How
> can knowing what a database means be a 'nice to have'?
See above.
> Where are Chris Date's
> 'external predicates' stored? In the heads of individual users?
Yes. And this is why psychology and usability are valid, if not theoretic, issues.
> > Rottweiler:
> >
> > Cabbage Potato Turnip
> > ====== ===== =====
> > Juliet loves Romeo
> > Romeo loves Juliet
> > Osama hates America
> >
> > It won't read like english in the predicate calculus, but that won't
affect
> > the objective meaning of the predicate.
>
> But you have not told me the predicate of that relation. It is meaningless
to me. By
> your scheme, the predicate can just as well be
> "Person Cabbage has the word Potato tattooed on their body because of
thing
> Turnip"
> as it could be
> "Person Cabbage looked up the word Potato in a dictionary while
thinking about
> thing Turnip"
Yes, you are right. I have not specified the domains or any constraints, and I have made no attempt to describe the external predicate. The domains and constraints do affect the objective meaning of the predicate whereas the external predicate affects only the subjective meaning. The names, however, do not affect the objective meaning. The names affect human ability to apply subjective information for reasons of human biology.
I disagree.
> By your scheme a relation means nothing without specifying it's external
predicate,
> and therefore 99% of the relations ever sketched in books and 99.999% of
tables in
> SQL databases have no meaning.
I disagree.
> I believe than needing to define an external predicate for every tuple
type is much
> too much work.
It is not possible in any case. The external predicate comprises the semantic elements not suitable for machine processing.
> Therefore I say the that the tuple type itself should be (a reasonable
> approximation) for the predicate.
I agree.
> > I am not saying I prefer vegetable names; I am only saying the bottom
line
> > of the issue is it is a matter of preference or of psychology.
>
> We are not blank slates you know. We do have some innate preferences.
Yes, I agree. Psychology and usability measure those preferences empirically.
> The preference
> for meaningful names is not a 'matter of preference', it's just the way
our (and any)
> mind works.
Again, this is biological empiricism and not theory.
> If preferences are universal (especially if they would be universal in any
> conceivable intelligent entity), then they are certainly suitable for
inclusion on
> the theory side of any line in the sand.
Unfortunately, preferences are not universal. There is no accounting for taste. Received on Tue Sep 02 2003 - 17:25:35 CEST