Re: Basic question?What 's the key if there 's no FD(Functional Dependencies)?

From: Cimode <cimode_at_hotmail.com>
Date: 12 Nov 2006 06:47:05 -0800
Message-ID: <1163342825.078898.56690_at_m7g2000cwm.googlegroups.com>


vldm10 a écrit :

> Cimode wrote:
> > vldm10 wrote:
> > > NENASHI, Tegiri wrote:
> > > > vldm10 wrote:
> > > > > NENASHI, Tegiri wrote:
> > > > > > vldm10 wrote:
> > > > > > [...]
> > > > > > > as far as I understand math a relation doesn't have meaninig
> > > > > >
> > > > > > Is it that you are serious ? The relation is a set. You do not
> > > > > > comprehend what the set is ?
> > > > > >
> > > > > > --
> > > > > > Tegi
> > > > >
> > > > > don't think that the comprehension is the same as meaning.
> > > > >
> > > > > Vladimir Odrljin
> > > >
> > > > You like to play with words and pretend not undertsand the words ? This
> > > > is a game you must play all alone, I am sorry.
> > > >
> > > > --
> > > > Tegi
> > >
> > > I think that meaning is pretty complex concept.
> > > Let me simplify things and say that meaning has an intension and an
> > > extension.
> > > If we speak about a meaning of the relation then (if we can speak about
> > > relation's meaning ?) we need extension. Maybe we can associate the
> > > predicats,
> > > maybe interpretation of the predicates, truth and false and the real
> > > world
> > > to see what is truth there. RM doesn't speak about the real world, the
> > > attributes,
> > > the entities etc.
> > > So my point is: let define meaning of the relation first - then I can
> > > see preciselly
> > > what paol c want to say, maybe he is in right.
> > It looks as if the eternal dilemna of definition/meaning of relations
> > is getting back on track again. In fact, a pretty sterile debate. Two
> > schools of thought seem to dominate. People who define relations as
> > ensemblist combinatory functions according to their transformational
> > properties and the one who define relations according as set of values
> > that populate them and therefore describe them. The latter is a result
> > of computing theory trying to force math abstraction in leading to
> > formulate the axiom that all functions are necessary relations. The
> > first approach OTOH is a total negation of such axiom.
> >
> > My two cents...
> > > Vladimir Odrljin

>

> I didn't have time to respond earlier.
> My intention in the tread was to put accent on
> the entities and the attributes rather then on FD's.
> Their structure may define a large groups of the
> databases. The cases 1) and 2) are about entity
> which attribute's are changeable.
> 3) is about entities with the same attributes.
> 4) is about entity which attributes are not
> FD dependable, but their number in the entity
> can be greater than 5 for example.
> Here FD's sometime doesn't work.
> It will be interesting if you can post in the more details
> about the definition of relation.
> Maybe with the examples related with this thread
> or generally related with RM theory.
I see...
My point was that relations are defined differently according to whether they would be considered specific type of mathematical functions or that functions should necessarily be considered a subset of relations.

In the first case, relations are considered as transformations of N tuple sets into other Ntuples sets and are defined according to their characteristics to perform restrictions and/or manipulation during the transformation process called operation. In such case, relations are primarily defined according to the operations they may be involved into and stronger emphasis is put onto relation variables than relation values. In the second case, relations are characterized according to values without further analysis and conclusions may be drawn over their nature according to observation of possible operations they be involved into. Which makes a huge difference of perspective.

I have found out the first approach considerably simplifies relation characterization but the second approach is closer to computing than mathematics. That's why the second approach is the current more accepted approach in current RM theory.

Hope this clarifies...
> Vladimir Odrljin
Received on Sun Nov 12 2006 - 15:47:05 CET

Original text of this message