RE: Our OO Verus Procdedural Discussions and New Forum

> From: Paul Holm
>
> 3.    IMHO:  OO is paradigm changing and is not easy to initially
grasp
> however
> it has the capability to reduce development times by orders and
magnitude.

This is simply not true, certainly not on a blanket level.  Code reuse
from OO is phenomenal, and as such it may allow you to get SOME
benefits.  But to say orders of magnitude means that you insist it will
cost an average of 1% of the development time to develop an OO solution
as opposed to a procedural one.


> OO is not necessarily about a product, it is the fundamental approach
one
> takes to developing software.

Again, I disagree.  It's simply one more way of programming.  Nearly
everything you do in OO can be done in one shape or another in other
languages.

By the way, Paul, you're not the first one to say this.  The SmallTalk
guys were saying it a long time ago and so were the C++ folks.  But it
never really translated to the real world.  That's because the benefits
of OO really only add up when you're talking about high levels of reuse,
and in the real world, reuse is not so simple.  Past data entry, there
just isn't that much reuse.


> 4.    IMHO:  It is my assertion that procedural approaches cannot
solve
> business problems with anywhere near the productivity and flexibility
that
> an OO approach yields.  My guess is it may take 2 months in procedural
to
> solve Nathan's problem to the level that a OO approach has.

This is the part I have a problem with.  You constantly talk about the
time it took to put together Nathan's example.  And here's the sticking
point: you keep saying "72 minutes" or "91 minutes".

How many of your classes did it reuse?  How long did it take you to
develop those classes?  WOW has been in development for years.  If I
took two years to develop a skeleton RPG framework, I bet I could do the
same.  In fact, I remember churning out basic file maintenance programs
in under a day, and they had a lot more validation steps than yours.

Heck, I can put 1700 green screen programs on the web in six hours.
Does that mean it only takes 12 seconds per screen?  Well yeah, provided
you buy my product. But if you amortize the cost of developing the
framework, it's a little higher.

How long would it take you to develop the entire bit from scratch, do
you think?  Or how much of your stuff are you willing to release into
the Open Source community so that someone else can do the same thing?
Because if it's not freely available, then your time numbers are a
little skewed.


> We stand today at a major cross road for iSeries and industry wide
> development with truly paradigm changing possibilities.

Let's see if that's the case.  Maybe we can come up with just the sort
of frameworks that allow us to do that.

Joe