 |
|
shawnmmcc
|
|
I've play SMAC since shortly after its release, though I've only been on the forums for a month. There seems to be a consensus that the AI could use some overhaul. Can it be modified directly, rather than indirectly with faction editing/game difficulty? Does anybody know where in the game the code is kept?
If enough people supported it, maybe firaxis would make the SMAC AI code available now that civ3 with a better AI engine has been released. Then the various SMAC/X communities could work on an effective builder AI. Obviously Firaxis would protect their copyright, and possibly even discover some good coders out there. Maybe a new release? Just a thought.
|
|
| |
|
|
|
zsozso
|
 |
of Chiron
Oct 1999
time: 00:25
|
|
I do not know about any way to modify the AI. However, if Firaxis would be willing to give us some information or a script language input or any other way to do it, I would be very interested working on it. Even if they only do as much as release detailed information about the game save format, then I could already write an "external" AI, i.e. an automated PBEM/hot-seat oppenent program, which you could use instead of a human player to play a turn in a hot-seat game. What I mean, is that you play your turn, save it, run the external AI program that reads in the save and produces another save with its moves made, then you load that up and so on...
I could even make a web or email based version of it, where you can send/upload your file and it would sned you back the save after the turn is played by the external AI.
If there are any other programmers out there willing to chime in, then we could make it into serious open-source project. 
|
|
| |
|
|
|
spartak
|
|
I'm all in favour. Unless firaxis are still making money from the game (which I don't think they are) they owe it to their fanbs to allow someone else to make the AI work.
|
|
| |
|
|
|
ellie
|
|
Hmm i bought a legal copy of alien crossfire.
It IS available for free download on a number of sites (i know one site bound to have it), but it feels wrong copying it
Ellie
|
|
| |
|
|
|
zsozso
|
|
of Chiron
Oct 1999
time: 00:25
|
|
quote:
Originally posted by Archaic
As I recall, didn't Firaxis already refuse a similar request quite some time ago? |
Was that request about opening the AI for programming by users, or giving out information on the game save format ? Because the later one is a much smaller request in terms of IP rights or work on their side. I mean, they must have some internal docs about the save format, so it is zero work to release that and they do not give away much IP with that either. Being in the software business myself, I wouldn't refuse the request for the format, but I would have some concerns about giving access to AI code.
So, maybe it is worth approaching them about the save format. Of course, that solution would not be so convenient for the players (requires saving/loading each turn), but at least it is a work-around. I'm sure a lot of players would be willing to take that much trouble for the benefit of playing against a decent AI.
We could even make the web version to learn, i.e. build up a knowledge base from the games it played and use tactics "borrowed" from players who won. This way the more we play it, the smarter it gets...
Last edited by zsozso on 03-02-2003 at 18:56
|
|
| |
|
|
|
shawnmmcc
|
|
EmperorJay, thanks for the post. Actually, it would not be that "hard" to write the AI, just moderately complex. My programming experience predates the Windows paradigm (secret project, increase economy +2, automatic status reports on enemy secret projects, +2 to probe attempts to sabotage enemy patents  )
What would be important is getting enough people involved for the various strategy/faction/world sizes to be covered. You are probably talking megabytes of added programming code, and a necessity to run it on the last generation of computers. Probably P3/Thunderbirds over 500mhz (preferably 1ghz) with 256MB of memory, again preferably DDR or Rambus (the Windows paradigm  ).
It should not be difficult to write an AI that is faction specific, that takes into account all factors through the midgame. String them together, and you have a moderately successful AI. If the programming is done properly, it will be modular and you may have a master routine that calls the following subroutines - Large World, dry terraforming, Morgan, no neighbors. When neighbors are discovered, a threat analysis program could be called, with the options of peaceful, border guard, potential hostile, known hostile or war taking over the "no neighbors" subroutine.
A late game AI might be more difficult, I don't know for sure because my experience has been mostly solo, when the game is almost always decided by midgame, usually about the time Fusion Power and Helicopters are discovered. The AI still will be at a disadvantage compared to a experienced/skilled player, but set this kind of AI to transcend and even they will have a very tough time. Let's just say no more winning as Mirium in a solo Huge World transcend game 
|
|
| |
|
| |
All times are GMT. The time now is 05:25.
Apolyton Time is 00:25. |
top of page
|
| archivepost |
|
Forum Rules:
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts
|
HTML code is ON
vB code is ON
Smilies are ON
[IMG] code is ON
|
|
|
|
|
|
01-02-2003 13:15
Your idea would make a quick customized faction AI much faster, as in small map/Hive, which eventually could be assembled into a larger universal program. Nice idea. 
)
|