if does not evaluate

James Moughan moughanj at tcd.ie
Fri Jun 11 05:02:11 CEST 2004


hungjunglu at yahoo.com (Hung Jung Lu) wrote in message news:<8ef9bea6.0406100710.1a17b11b at posting.google.com>...
> moughanj at tcd.ie (James Moughan) wrote:
> >  I still have to restrain myself from writing PyHaskell now and then.
> 
> Maybe it's time to try with some simple prototype.
> 
> > Secondly, Lisp's syntax doesn't parse well into the way people think,
> > or vica versa.  Python's does; in other words, it's executable
> > pseudocode.  Lisp, fundamentally, cannot change in this regard.
> 
> But Python does have serious problems when one goes to the next level
> of software development. I guess that's why Io and Prothon were born.
> Python is at the stage where you need a fresh re-incarnation to go to
> the next level.
> 

I'm curious as to what those problems are, and what the 'next leve'
is; there are definitely warts, and some, like the scoping rules, get
in the way occasionally, but by and large there's nothing that can't
be worked around.  It lacks severely in compile-time metaprogramming,
but has some very nice run-time features, which are arguably better in
a language which isn't too focussed on performance.

> There are a few directions that need to be explorered, and I am not
> sure creating a real new language is the way to go. I've seen it in
> Io, where once you set things in stone, it becomes just another Python
> with problems that will stay forever. I guess right now it's not the
> moment of creating more languages and set things in stone. It's
> probably better to have some toy languages or language prototypes to
> explorer ideas. Collect enough ideas and experience, and probably
> leave the rest to the next generation of people.
> 
> Frankly, I see a few camps: (a) Lisp and AOP folks, (b) Functional
> folks, (c) Prototype-based folks. Because these are very specialized
> fields, very few people seem to be native speakers of all three of
> them. The next killer language will have to incorporate lessons learnt
> from all three camps. It's a daunting task. It's best to have some toy
> languages... like scaled-down model airplanes, or even just the model
> airplane parts, before one actually build the real thing.
> 

There are a lot of interesting ideas out there - personally I'd like
to see Stepanov cease being so notoriously lazy and write the language
C++ could have been. :)  We'll probably find different paradigms being
useful in different contexts, and to different people.  After all, how
boring would the world be if you only ever needed to learn one
language?

Jam

> regards,
> 
> Hung Jung



More information about the Python-list mailing list