clipka wrote:
> "David H. Burns" <dhb### [at] cherokeetelnet> wrote:
>> I drafted a rather longer reply to this, but thought better of it. I'll
>> say only this
>> programming is *not* hard, but it cane be made hard.
> 
> As a professional SW developer, I can tell for sure that it depends a lot on the
> size (and even more so complexity) of the project, and the operational
> parameters to match (memory limits, performance requirements, stuff like that).
> 
> If your projects are the size and complexity of a typical POV-Ray scene with
> some occasional loops, you're right: That's not hard at all.
> 
> If however your project is a POV-Ray animation employing some mechanics
> simulation, and you don't have access to a mech sim framework developed by
> someone else already, you *might* come to the conclusion that programming can
> also be hard all by itself.
> 
> And if you were to develop software for a bluetooth handsfree unit to be
> integrated into a car, or a modern computer game, you'll certainly find how
> hard programming can be.
> 
> 
> One problem with OOP support in many mainstream languages is probably this:
> They're designed to manage complexity in medium-sized projects (for really
> large projects, even those languages may not be good enough); if they do that
> well, nobody will complain about the additional overhead they impose, because
> it really pays off a lot. However, in very small projects this overhead becomes
> more of a burden than a benefit, because the projects don't really have much
> complexity in need of being managed.
> 
> Mind you, this overhead is imposed not by object-oriented programming as such,
> but by the languages. The more formal it is, the more suited it is to manage
> complexity. But, of course, that formalism makes it also less suited for
> programs you could throw together in a few hours if it weren't for the
> formalisms.
> 
> I do love Microsoft's C# programming language and IDE: It's really one of the
> most powerful environments I have ever used for development, and it's really a
> pleasure to use it. I can't think of anything better suited for projects the
> size an experienced hobbyist programmer could ever hope to pull off alone.
> 
> However, when I develop, say, a small straightforward command line tool I expect
> to throw together in a matter of two or three days, C/C++ is actually my
> favorite weapon: I can start off with plain old-school C imperative programming
> (interspersed with some C++ syntax that has nothing to to with OOP, like line
> comments, references, bool data types and the like); if I then happen to
> encounter some unexpected complexity, I can throw in a bit of C++ to define
> some objects anytime, but keep the program C-style otherwise.
> 
> This is the kind of OOP support I'd expect for a POV-Ray SDL: As long as your
> project is straightforward, it doesn't bother you with any formalism
> whatsoever. But when the task at hand gets complex, that OOP support is right
> there for you to use, with all the formalism you need.
> 
> 
Your points are well taken, I should have said programing is not 
*fundamentally* difficult.
A particular programming task may be difficult because it require a lot 
of time and effort, learning
the idiosyncrasies of all the systems involved and a lot of trial and 
error. A great deal of
experience programming may also be required. But fundamentally, 
programming is not difficult.
But it can be made so.

David

Post a reply to this message