Re: Objects and Relations

From: dawn <dawnwolthuis_at_gmail.com>
Date: 20 Feb 2007 19:39:37 -0800
Message-ID: <1172029177.861368.224120_at_v45g2000cwv.googlegroups.com>


On Feb 20, 4:05 pm, "JOG" <j..._at_cs.nott.ac.uk> wrote:
> On Feb 20, 8:05 pm, "dawn" <dawnwolth..._at_gmail.com> wrote:
>
>
>
> > On Feb 16, 9:32 pm, "Marshall" <marshall.spi..._at_gmail.com> wrote:
>
> > > On Feb 16, 4:37 pm, "Keith H Duggar" <dug..._at_alum.mit.edu> wrote:
>
> > > > I would like to claim that this very discussion reveals one
> > > > of the advantages of trying to think without entities. It
> > > > encourages us to think about the /problem/ instead. That is
> > > > to think about our goals, our requirements, our knowledge,
> > > > etc. It forces us to consider the facts at hand and those
> > > > that may arise and design solutions for handling them.
>
> > > Yeah.
>
> > > I used to have to deal with this vaguely uneasy
> > > feeling that terminology was an indicator of some
> > > piece of wisdom that I didn't have access to.
> > > So I'd hear talk of, say, UML or OOAD or whatever
> > > and think, oh, heck, I better learn what that is.
> > > So I'd buy a book and it'd be really hard to understand.
> > > I'd push and push, and eventually I'd figure out
> > > that they were just doing something straightforward,
> > > like "programming" or "data modelling" or something,
> > > but they had dressed it up in some fancy clothes,
> > > added some extraneous concepts, applied some
> > > arbitrary rules, etc. The intent was to obscure
> > > rather than to reveal. Make it look like more than
> > > it was. Really annoying.
>
> > > I remember reading a guy on comp.lang.functional
> > > describing going through the same process, but
> > > over the phrase "dependency injection." After a
> > > week of reading he figured out it meant "abstraction"
> > > (as in "lambda abstraction.") In other words it was
> > > just the process of parameterizing code.
>
> > > The thing about entities is, what does it buy me?
> > > I've got relations; I know how they work. Now I'm
> > > supposed to layer this "entities" concept over the
> > > top of that. What do I have now that I didn't before?
>
> > > Seriously, what?
>
> > > Marshall
>
> > I use entities in conceptual modeling. I do not see a need for it in
> > logical modeling, although it is just a level of abstraction away and
> > some products (or standards such as XML) use the term in the logical
> > layer as well. Relations do not fall from the sky, so how did you
> > know what relations to design?
>
> Ah yes, but propositions do fall from the sky.

I offer mine as a counter-example. smiles.

> > You likely looked at nouns/entities/
> > things at some point, with "entities" being the preferred term (as
> > best I can tell).
>
> > I'm in the middle of writing up tips for going from entites and their
> > properties to relations and attributes when working with a non-1NF 2VL
> > DBMS, since there is already plenty written on this in the case that
> > the target DBMS is an SQL-DBMS. If anyone knows of anything written
> > on this subject (the one I'm writing on), please pass along any
> > relevant URLs. I have been highly unsuccessful in finding other write-
> > ups on this.
> > Thanks --dawn
Received on Wed Feb 21 2007 - 04:39:37 CET

Original text of this message