Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Eagleon]

I'm curious as to what type of AI will eventually be used? A want/need/fear weighing system ALA the Sims seems the most organic, easy to develop solution to me, for basic automation. It also probably would have the advantage of being easily customizable, and of low system requirements.

What's all planned?

[Toady One]

There's a bit outlined throughout the Mental and Civilization future pages, but maybe not enough, and it is not organized well.

I'd been playing with this a bit in the dwarf game -- I guess it's like what you describe.  The dwarves currently default to doing whatever job they can to help out (or mill around otherwise) and when able they acquire wealth based on the kinds of materials they personally like.  On the other hand, if they need to eat or drink or sleep, they'll cancel their current activity to take care of this (unless they are involved in a high risk activity like fighting or leading a dangerous pet).  If they are presented with an opponent, they will fight or flee based on an assessment of the risk, and when the danger is over, they'll go back to their default activities.

I imagine something similar will work in Armok.  However, they have a bit more juggling to do, since they'll have a list of allegiances to various "entities".  So, for instance, a vassal to a lord might have that allegiance but also be a member of a cult and be loyal to their own family.  Each decision that comes down the line would be filtered against these allegiances (which could, say, be measure by a percentage and a few saved important events and status ratings), and if one of them is stretched to the breaking point by a conflict, the break would be made (mentally if not officially).  So if the cult leader asked this person to sacrifice their child, the person would need to weigh the gain and harm to arise to each of their allied entities (based on their importance) and to themselves through this action and via repercussions from the entities.  This would have to be done numerically on the computer, so I'll just have to restrict plot devices to those which I can comprehend in this way.

I suppose actions should be greatly influenced by the personal ambitions of the creature independent of allegiances -- for speed's sake, most creatures probably wouldn't have any or they wouldn't matter.  On the smaller scale of day to day activities, Armok will probably work just as in the dwarf game described above.  The dwarf game has some minor implementations of the larger, sweeping decisions, but not on this scale, since the dwarf game is more of a Tolkie-good-evil setting.  Good for practice before the real thing, anyway.

Once I get done with the item release, I'm going to go back to the dwarves until they are done.  This should inform the decision making that people will have to do to make the Quaint Thorpe Armok release functional.  Assuming I still have a job...

[X]

quote:Originally posted by Toady One:
<STRONG>Tolkie-good-evil setting.
</STRONG>

Bad Toad. I hate it when people get all 'EPIC BATTLE BETWEEN GOOD AND EVIL' over Tolkien. It belittles the subject matter. Not that I'm questioning your scholarship, just your representation.  

Sounds good stuff on the AI otherwise though.

X

[Toady One]

I wasn't referring to an epic battle between good and evil, but rather that people in Tolk typically didn't have that many conflicting allegiances to balance.  Although Tolk has a number of good examples of the decision making I'm talking about (where Aragorn went after the breaking of the fellowship for example), most people seemed to have one overarching motivation.  This is not a criticism, but rather the distinction I was attempting to draw.  Maybe it's because I'm reading Silm -- Morgoth tends to make things stark for decision makers, although there is still a bit of jostling.

[Handshakes]

quote:So if the cult leader asked this person to sacrifice their child...  

This caught my eye. I am curius, when would an event like this arise? I mean, this sounds like a very specific task that could possibly have no effect on the player assuming he is not the person being asked for the sac or he is not the cult leader. Could you elaborate perhaps on what kind of actions npcs have?

Actions to that level of detail would be really cool, but I would imagine very taxing to one's system.

Forgive me if I misinterpreted, but every time you talk about design aspects of your game, Toady, the game sounds even greater and greater!

[Toady One]

Well, since it's an perpetually ongoing project, the design will continue to outpace the programming.  I'm not making any promises!

So here, the idea is to have randomly generated plots.  You've probably played a game or read a book where you have to save somebody from being sacrificed -- this is an example of when it would come up.  It would be taxing on the system to keep track of all the plot events, but the unimportant ones can be abstracted away like everything else.  It doesn't need to keep track of everything.  If you visit a city, it will generate a number of plot lines and so on to keep things interesting (using the civ and creature structures, and a random seed).  If you leave, it won't try to advance those plots while you are away, but more likely save the number of days/years you've been gone, and recreate and advance the plots when you come back.  If you involve yourself heavily in a plot line, it will be saved and can be used to create plot lines in new places you visit.  It won't be that hard to set up, but it will be a task to make it not seem "cookie-cuttery".

[Handshakes]

Genius I say!