Re: Clean Object Class Design -- Circle/Ellipse

> Discussions of the circle/ellipse type have been coming up every couple of
> months for years. One of the things that strike me about the discussions is
> that people aren't very precise about what they mean by type. You often see
> people offering the definition X is a subtype of Y if every instance of X
> IS-A instance of Y. That's not terribly precise, unless they tell us what
> the meaning of IS is ;-)

Now, don't get Clintonian on me. LOL In any case, Date is very precise about what he means by type and subtype. See below:  

> The intuitive meaning many people seem to be using for subtype is simply
> subset.

At http://www.dbpd.com/vault/9902/date9902.shtml, Date explicity states that a subtype "has a subset of the values and a superset of the properties" of a supertype.

This is consistent with Date's essay at
http://www.firstsql.com/dbdebunk/cjd8a.htm

"For example, all ellipses have an area, while some ellipses, specifically those that happen to be circles,- have a radius as well."

I don't know how anyone could mistake Date for meaning that subtype is simply subset or conclude that he is not very precise in what he means.

> My point is that that is only one (the simplest) of many possible
> definitions. Subtype isn't a natural phenomenon, it's a matter of
> definition. What I'm interested in is which definition is the most useful,
> perhaps depending on the circumstances.

When discussing values and operators on those values, the values and operators exist regardless whether you decide to use them or even know about them.

> Now, about the claim being extraordinary or not. Suppose someone said
> "Richard Nixon, Jimmy Hoffa, Antonin Scalia and Bill Clinton all have the
> Q37 property". Is that a true statement? No, it's a meaningless statement
> unless he tells us what the meaning of Q37 is. Is it an extraordinary
> statement? No, for the same reason, although it can be for sufficiently
> infuriating definitions of Q37. That's all I was trying to say.

Regardless of any additional meanings one might ascribe to the properties of circles and ellipses, the meanings used by Date in his essay are clear and consistent with common understanding of those properties. In any communication, we must assume a minimum level of common background and experience.

For instance, my background leads me to conclude that formalism on its own does not suffice. One must always bring outside knowledge and understanding to the table.

I am not willing to concede that all statements are meaningless. Any reasonable meaning of the original statement regarding Date's understanding of type is extraordinary and ad hominem. Received on Sat Aug 18 2001 - 18:56:38 CEST