Re: Topic-Organized Object-Oriented Programming


ggroups@xxxxxxxxxxx wrote:
On Feb 8, 7:15 pm, "topmind" <topm...@xxxxxxxxxxxxxxxx> wrote:

The first thing I will put on record is your convenient omission of
your claim
about what the OP wrote. It is therefore fair IMHO to assume that the
following
is true (as I stated) :

you started your usual deluded tree rubbish
- the OP responded to inform you they are talking about graphs
- you made a claim about what the OP wrote, which is shown to be
untrue


I did NOT bring up hierarchies. The opening posting brought it up.
Thus, your implication that I "saw" hierarchies were none were
described nor mentioned is wrong.


So getting back to the rest of the posting ...

On Feb 8, 11:01 am, ggro...@xxxxxxxxxxx wrote:

On Feb 8, 6:40 pm, "topmind" <topm...@xxxxxxxxxxxxxxxx> wrote:

But that is about compiler speed optimization and IDE hints
(programmer search utils), not about the meaning of OO itself.

Known of these things have anything to do with the statement that OO
can
easily provide set-based type classification schemes, based on
concepts such as
strong typing and substitutability, that do *not require type
hierarchies* .

OO languages are Turing Complete such that in theory they can build
and make anything that any other language or paradigm can. However,
writing Smalltalk in Fortran or visa versa does not show us anything
useful beyond TC itself.



TM> But OO *itself* is not based on sets, and that is a drawback of
OO.

Not at all, because the categorisation techniques that type substitutability
provides are trivial to add, which is a benefit, not a drawback. This means the
OO concept is sound because it can be extended orthogonally with such
ease.

Demonstrate its shiney greatness by showing how OO improves biz apps
using objective metrics. Oh, I forgot: you are a dark-age alchemist,
you don't believe in science.

Provide a an example classification that you feel is suitable for
discussion, and
I will show you how *existing* type regimes will treat it with
impunity, and without
needing type hierarchies (one of your deluded "tree" straw-men
perhaps) ...

This probably will not get us past the TC issue, but if you want,
start with the lab-coat example in the nearby "sets" link.



Regards,
Steven Perryman

-T-

.