Re: OO Design induces an existential crisis

adaworks@xxxxxxxxxxxxx wrote:
> "topmind" <topmind@xxxxxxxxxxxxxxxx> wrote in message
> news:1121703637.780045.280310@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
> >
> >
> > And, some of the are vague, such as Open-Closed Principle.
>
> There is nothing vague about the Open-closed principle. It is easy
> to understand, easy to implement, and makes the software solution
> easy to extend at some later time.
>
> Open = the design modules are open to easy extension
> Closed = the design modules are closed so they cannot be easily corrupted
>
> This is not simply a principle of object-oriented programming. It has its roots
> in the work of John V Guttag and his dissertation on abstract data types. But
> then, you probably take issue with the value of ADT's, as well.


Using "types" for business modeling has been a common complaint of
mine. Stepanov has made similer complaints about the lack of
multi-factor-handling of "types".

Perhaps I should qualify this better. It perhaps is the *examples* of
Open-Close that were poor rather than the principle itself. (Criticism
of the examples was already given.)

Nothing non-trivial can escape opening when some kinds of changes
happen, at least in my domain. It is a matter of weighing
probabilities. To make software change-friendly, you have to be able to
estimate which direction change will like happen and engineer for such
change, or at least prepare for it. No paradign gives a free lunch in
this area so far. And, polymorphism does not seem to be the "trick"
either, at least not beyond shapes, animals, and device-drivers.

>
> Richard Riehle

-T-

.