Re: OOP - a question about database access

"Topmind" <topmind_at_technologist.com> wrote in message news:4e705869.0311082326.1e704776_at_posting.google.com...
> > > Personally, I do not hang out in comp.object. Having long ago
> > > mastered the technology, I find comp.object infertile ground and a
> > > waste of time. I foresee no important advances there given the
> > > primitiveness of the computational model and the distinct lack of
> > > advances in past decades. You only see my posts because this
> > > thread is cross-posted to comp.databases.theory, where I see very
> > > fertile ground.
> >
> > Coming in to the thread a little late, but, do you feel that OO is
> > infertile and a waste of time for developing, say windowing
> > environments, or a web server? It seems to me that is is a useful
> > disicpline for certian applications. I've always found the Person ->
> > Employee -> SalariedEmployee example to be contrived, but I don't
> > see how one goes about game programming in relational algebra.
>
> People who feel that OO works well in some places but not
> others do not seem to deliver consistent answers about
> when and where OO works best.

The problem I see with your statement above is the difference between well and best. OO can work well in some places without ever working best.

> The only semi-consistency
> I find is that OO may work better where there are
> natural, stable taxonomies or classifications
> or sub-types of things.

This is almost always the case--at least when one considers simpler types.

> This would be like
> chemistry, geometry, etc. I don't
> really have a strong opinion on that aspect because I
> don't work in those domains. I am more familiar with
> human-created domains of business and society.

I don't see how the location-based computational model provides advantages to any application domain. How would it benefit chemistry, for instance? Received on Sun Nov 09 2003 - 14:06:23 CET