Complete.Org: Mailing Lists: Archives: freeciv-dev: April 2001:
[Freeciv-Dev] Re: suggestion
Home

[Freeciv-Dev] Re: suggestion

[Top] [All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index] [Thread Index]
To: 'Grzegorz Borowiak' <grzes@xxxxxxxxxxxxxxx>
Cc: "'freeciv-dev@xxxxxxxxxxx'" <freeciv-dev@xxxxxxxxxxx>
Subject: [Freeciv-Dev] Re: suggestion
From: Michele Slocovich <m.s@xxxxxx>
Date: Tue, 24 Apr 2001 00:17:27 +0200

Ok, your description of future possibilities clearly matches what were my
unspoken thoughts.
But for one thing, which has also been pointed at in previous replies to yr
msg, the C,C++ choice.

First, i got to admit i am not (anymore) a "real" programmer, therefore i
was thinking of a "scripting language".

What could point in favour of C++ is the fact that inventing a specific
language could be too much an effort 
to build up such a thing from scratch without possibly seeing results before
a while.

I wouldn't be too much concerned of performance issues, at least unless the
bots' tournaments you depict have
also to be run within computational limits :), (ie all running 25Mhz 386
with 20 seconds turns..)

Anyway, what in intermediate solution could be is to "expose" or "export"
methods, properties etc of the client,
so that some poor programmers (say using perl or something "easier" than
c++, or whatever someone would choose)
could pilot the thing and get some feedback. 

I don't know whether this is possible and how it could be implemented, but
it could be easy enough to "test the audience" 
for some interest in further development in this direction...

-----------------------------------------------
Michele Slocovich
-----------------------------------------------

> -----Original Message-----
> From: Grzegorz Borowiak [SMTP:grzes@xxxxxxxxxxxxxxx]
> Sent: Monday, April 23, 2001 12:51 PM
> To:   Michele Slocovich
> Cc:   'freeciv-dev@xxxxxxxxxxx'
> Subject:      Re: [Freeciv-Dev] suggestion
> 
> On Sun, 22 Apr 2001, Michele Slocovich wrote:
> 
> > Idea
> > scripting language for client side ai implementation
> 
> My proposals: C, C++.
> 
> All other are too slow to provide good ai.
> 
> > Description
> > implementing the possibility of -initially- scripting some macros and
> > further to 
> > have bots playing.
> 
> I see it as common skeleton, whose purpose is:
> 
> - fetch/update current state of game from server
> - call a user-defined function, which does ai action
> - respond to server.
> 
> So it would be a skeleton for such client agents. What you put in user
> function is what you want to; neural networks, case-based reasoning or
> just imperative algorithms.
> 
> It would consist some modules:
> - actual skeleton, which does what above and is a kind of software
> interface;
> - some useful functions for data processing;
> - user-defined module.
> 
> Future:
> 
> 1. International contest of bots (bot vs bot etc.). Something like core
> wars.
> 
> 2. Gallery of bots. I see it as a web page on www.freeciv.org containig
> many bots to download and short characteristics for each. E.g.: "good on
> land, but poor across the ocean, smartly uses mountains and hills for war
> operations; winner of international bot contest on Aug 2001 in category
> 'troubled land'". For built-in ai of freeciv server corresponding
> characteristics could be: "very good on oceans, average on land, extremely
> poor in air; so far the best known freeciv ai in the world".
> 
>                                          _____________
> Grzes                  \_______________________|_______________________/
> grzes@xxxxxxxxxxxxxxx                         (_)
> http://gnu.univ.gda.pl/~grzes/
> 'Nam chodzi o odglos paszczowo.'  'Patataj patataj patata...'



[Prev in Thread] Current Thread [Next in Thread]