>> From what I've seen, "every other programming paradigm so far" has been
>> ad hoc without much in the way of a strong, consistent theoretical basis.

> Actually, I can't think of how a programming paradigm could be ad-hoc
> without a consistent theoretical basis; I mean, that's why it is called
> a /paradigm/ rather than just coding style or some such.
>
> Maybe you're confusing programming paradigms with programming languages
> here.

Yeah, perhaps.

> Yes, virtually ALL mainstream programming languages have something
> ad-hoc-ish about them, in how they're not strictly adhering to any
> single programming paradigm.

The thing is, there /are/ programming languages which stick consistently 
to one programming approach. Unfortunately, it tends to be the mish-mash 
languages which have backwards compatibility to all the pre-existing 
crap that tend to be popular.

> But wait... maybe that's why they ARE mainstream after all - because you
> can mix & match different paradigms with just one language? You know,
> solve the different parts of the software in the way that's most suited
> to each one. Get around some problems with one paradigm by offering
> alternative paradigms to base your software (or module) design on.

I would argue that it's just because they continue the poor design 
choices that came before. Half the complexity in C++ is due to backwards 
compatibility with C. Java and C# just copy the syntax from C++. Visual 
Basic even has "basic" in the name. And so on.

> After all, for practical purposes it is perfectly irrelevant whether
> your language is simple and elegant at its core - all that matters is
> whether it can solve YOUR problems in a way that YOU can wrap your
> brains around after a reasonable amount of training.

Sure. And having a complicated, messy language which lacks internal 
consistency makes it so much easier to learn. Oh, wait...

> http://www.youtube.com/watch?v=WxxG0faDT_M

When I hear [yet another] story about people storing XML in the database 
rather than change the schema, this is what I think it sounds like.

>> I mean, take SQL. It solves only one problem, but it solves it so damned
>> well that it is basically the /only/ language of its type. And oh look,
>> it's based on a theoretical model. Funny coincidence, that...
>
> Yeah, strange though that they didn't use a functional paradigm for
> those databases...

Functional programming is a model of computation, but databases don't 
compute anything. They just store stuff. Far more logical to use a model 
of knowledge for that, no?

Post a reply to this message