Three quick questions

[John Duncan]

> >I was wondering a few things:
> >    1. Does anyone have any experience doing genetic algorithms
using
> >Clean?
>
> Don't know. Chris Clack has used functional languages successfully
for
> genetic algorithms. If I remember well he used good old Miranda.
>

Yes, I've been looking at some work he and Tina Yu have written on
genetic programming in Haskell. I'm not exactly sure, but it seems to
me that GP needs either lots of batch compilations or a reflective
language. In this sense, it should not be too much of a pain to use
clean.

> >    2. How "open" is clean. What would it take to get the src?
>
> I principle the source code (C) of the current system is not
available. We
> are working on a new implementation of Clean in Clean (called Clean
2.0).
> The Clean parts of this implementation will be available.
> Why do you need the source anyway? As far as i can see, genetic
algorithms
> can be done without changing the language.

Yes, but for the applications I have in mind, it may be desired to
compile down to platforms that Clean does not support, such as the ARM
THUMB processors. I'd hate to have to turn down a good language simply
because of a lack of support. Or, perhaps the support team would be
willing to write new code generators and runtimes on demand, for pay.
But, as you have said in another message, you seem to be
oversubscribed.

In Clack & Yu, they show that one of the problems using Haskell is
that large data structures get copied over and over again. This may be
preventable with Clean and its unique types. But if I can't get the
support for Clean, I really can't justify using it.

On the other hand, the Clean team should really consider producing a
version of Concurrent Clean with pluggable code generators and a
runtime that can be targeted to at least one embedded RTK. This
wouldn't help my application specifically, but I think it could help
the adoption of a functional language in that lucrative embedded
systems market. Clean seems perfect for that market, among other
languages, considering its high efficiency and low system
requirements.

If a real-time GC were added, then it could get the additional
real-time embedded market.

Cheerio,
John