Re: The Latest Scheme

Rob Saunders wrote:
> I think you're getting a little ahead of the game with your prediction 
> that Scheme will oust Python. They satisfy very different requirements 
> and there is easily room for both types of language even within a single 
> game engine.

I meant Scheme will replace python within FEAR.  I do really like python 
as a language, but it's becoming increasingly obvious that it's not very 
suitable for games.  It's a big language, focusing on intuitiveness at 
the expense of implementation simplicity.  A lot of the discussion at 
the GDC Python roundtable was about Lua and how that solved many of the 
issues people were having with python (mainly bloatedness and speed of 
execution).  Scheme beats python hands down at both of these, and 
provides a really elegant foundation for most of FEAR.

I don't think game developers are ready for multiple scripting languages 
per project at this stage (such a hassle), so anything that can be 
converted to C/C++ is a bonus.


> One of the biggest problems when developing with Scheme is managing 
> large scale projects. There have been a number of attempts to develop 
> object oriented extensions to Scheme but none of them seem to have 
> gained much acceptance - which I guess isn't surprising given that most 
> people are attracted to Scheme for its minimalism.

Right, the standard only focuses on things that have proven useful and 
necessary in a broad context.  Scheme is a programmable programming 
language, so it can be extended in any way.  You can borrow lisp's OO 
features, or one of the many lightweight versions (most likely what	 
I'll end up doing).  I'll probably use a component system in a similar way.

But there is a new revision of the standard underway, and objects may be 
included.


> You might also be interested to look at Jonathon Blow's current column 
> in Game Developer where he is developing a language for AI scripting by 
> borrowing from Prolog, Lisp and Perl. Sounds horrid, but as usual with 
> Jonathon's work the result is quite elegant.

There's some stuff here, but not much on Lerp yet.
http://number-none.com/product/



> I assume that would be a Prolog interpreter without all the syntactic 
> sugaring for doing natural language processing.

heh, last I did NLP with prolog I never used any form of sugar!!  It was 
raw prolog.  I've been missing out :)   I doubt it'd take much effort to 
implement the functionality, Cyc seems to do it quite well.

Alex



-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials
Free Linux tutorial presented by Daniel Robbins, President and CEO of
GenToo technologies. Learn everything from fundamentals to system
administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click