[geeks] something funny emerges from all the BS on comp.os.vm s

[joshua d boyd]

On Tue, Jul 17, 2001 at 03:55:40PM -0400, Ken Hansen wrote:
> 
> My "problem" was that I had never learned how to apporach a programming
> problem in such a way that it was clear how object-oriented programming
> would make a significant difference in the effort involved. I look at a
> problem and think FOR...NEXT, WHILE, and IF...THEN, not overloaded
> operators, classes and whatever else was involved.
> 
> At the 10,000 foot level, I guess it could be beneficial, but the bulk
> of my programming (trivial "hacks" these days, basic business
> "stuff" back then) is/was satisfied by the tools I already had.

I'm increasingly of the opinion that C++ doesn't very well lend itself to
proper OO programming.  While I learned the ideas a long time ago, they
really hit home when I was studying 3 tier programming using MTS (heresy,
I know).  Suddently, no longer were you supposed to think of objects being
something that you sent messages to, but objects litterally could be
seperate things, and yes, you did send messages to them.  I thought it was
a rather nice way of looking at business logic, and if done right, it can
make things more maintainable.  Unfortunately, MS is encouraging people to
emphasize scalability instead of maintainability.  For instance, MTS
objects are SUPPOSED to be stateless (even though you can write statefull
ones).  This makes maintainence so much more difficult than opening an
object and keeping it till you are done.

In my personal time, I'm starting to get more into functional
programming.  While OO was nice for business objects, what I look at now
is more mathematical rather than transactional in nature.

Anyway, to get back to what I was saying about C++, C++ makes objects so
hard that people naturally want to write large objects to ease the
burden.  However, when you do that, you loose most of the benifits of OO
programming, so why do it?

-- 
Joshua D. Boyd