Re: Just one more anecdote

dawn wrote:
> Marshall Spight wrote:
> > > >
> > > > I suspect that's simply because you/SQL/the world hasn't figured
> > > > out how to tap into the full power of set theory yet.
> > > >
> > > I'm thinking of only using set operations when putting things into and
> > > taking things out of my purse. After all, that approach is based on
> > > set theory and set theory rules! ;-)
> >
> > These analogies between math and the physical world don't work
> > for me.
>
> OK, now I'm confused. What is that we do when preparing a data model?
> We are making an analogy between math and the physical world, right?
> We are preparing a map between the real world and a mathematical model.
> We are making a "mathematical metaphor" for our problem space, right?

Except we weren't talking about preparing a data model; we were talking about the design of the metamodel. That can be done in pure mathematical
terms, whereas designing the (non-meta) model for a particular problem domain *is* done by "analogy" with the physical world.

But if you insist, I hereby do the comparison.

We have a collection of coins in Dawn's purse. Each coin has a value. We want to remove a penny.

Here are two concrete models for doing so. One is declarative/set theoretic,
and the other is imperative/object oriented. (Neither example has actually
been compiled and tested, but I'm using two languages I know pretty well.)

• SQL delete from purse where value = 1 limit 1;

// Java 5, using generics
Purse purse = Dawn.getPurse();
List<Coin> coins = purse.getCoins();
for (int i=0; i<coins.size(); i++) {
  Coin c = coins.get(i);
  if (c.getValue() == 1) {
    coins.remove(i);
    break;
  }
}

I leave it to the reader to choose for themselves which they would prefer. I would propose they consider the economy of expression of each solution in making their choice.

I would guess that we could make a pretty good case for the second, Java-based solution as being at its core a hierarchical, navigation based solution. I've never worked with a hierarchical database, but I imagine it works something like this. Am I close?

Marshall Received on Thu Aug 04 2005 - 17:18:32 CEST