On Nov 10, 2007 1:30 PM, Bert Freudenberg <bert at freudenbergs.de> wrote:
> On Nov 10, 2007, at 11:57 , Jason Johnson wrote:
>
> One could argue that if you set out from a restricted environment,
> you can only enhance it by adding complexity.

Or realize that if something is so complex it's probably wrong and
look for another solution.

> That certainly is what
> happened - Java advocates always laughed about C++ templates, and now
> they have them too as generics, because you actually need them in
> that system to be expressive.

Yep, but if you switch to a late bound type system or a more effective
type system (e.g. Haskell's inferred type system is much easier to
work with while being stronger typed).

> If you start with a powerful simple model, the urge to extend it is
> much smaller - see Lisp or Smalltalk, you can basically do anything
> you want within the system.

That's true, but what would happen if Lisp or Smalltalk had the people
Java has now?  I'm obviously not optimistic about what would occur.

> Does anybody here still subscribe to the idea that Smalltalk should
> be a system in particular for non-experts? That certainly was its
> original motivation. It also has largely failed in that regard,
> though not necessarily so on technical reasons.

Well, I wasn't thinking about children when I said "people who should
not be programming".  I was more thinking about people who have been
brain-washed with horrible systems for decades that are either not
capable or not willing to see that there are better options out there
then what they know.

A child does not have this preprogramming so I would expect them to be
potentially good contributors.