Re: OOP
- From: Traveler <traveler@xxxxxxxxxx>
- Date: Wed, 21 Sep 2005 10:02:40 -0400
On 21 Sep 2005 01:33:08 -0700, makc.the.great@xxxxxxxxx wrote:
>
>Traveler wrote:
>> Having said that, I am not sure what the original poster is looking
>> for in a different inheritance mechanism. The OOP paradigm is more
>> than an adequate approach to software composition.
>
>Original poster thinks that "OOP paradigm" doesn't allow for software
>to evolve efficiently. Once low-level classes are coded, base classes
>cannot change substantially, unless whole thing is re-written. This is
>much more expensive way that I feel it ought to be.
As long as one keeps the same parameters for the base virtual
functions, any modification to the actual underlying code should
automatically be used by the higher level classes. At any rate, it
makes sense that changes to base classes can have major effects on
the higher level classes. How can it be otherwise?
Inheritance is just a compositional mechanism. The problem is that
raditional text-based programming makes it hard to visualize and
modify the composition of a software object. Re-writing (recomposing)
an entire class should not be a chore, IMO.
I think the problem that you seem to be alluding to has very little to
do with inheritance. It probably has to do with algorithmic
programming in general: modifying one part of the code can have nasty
side effects somewhere else. This is something that I call blind code.
It is the bane of managers and programmers who are given the task of
maintaining complex legacy software. More often than not, a project
manager will try to find ways around a bug that will not involve
modifying the code. The COSA software model eliminates blind code and
makes it easy to modify a complex program without introducing
unforeseen side effects. The reason is that any change to a variable
in a COSA program is immediately broadcasted to every object in the
program that is potentially affected by the change.
Louis Savain
Why Software Is Bad and What We Can Do to Fix It:
http://www.rebelscience.org/Cosas/Reliability.htm
.
- References:
- Prev by Date: Re: DARPA Grand Challenge
- Next by Date: Re: The structure of a self-conscious mind
- Previous by thread: Re: OOP
- Next by thread: Re: OOP
- Index(es):
Relevant Pages
|