Hmm.. everyone is so worried about AI eating up processing power. This would only work in the case of mmorpgs, but in theory if you have a section of the servers (in our case an individual network) devoted to doing nothing but controlling AI all you need to do is send the actions or results of the AI descision to whomever is close enough to see it or be affected by it.

Wow! That''s pretty cool! =D A couple things...

I think even simulated emotions, even though, we as humans are attuned to finding flaws in as he says are a step up from no emotions. Catz, Dogz, and Creatures use VERY sophisticated AI for their emotional simulations. As far as for RPGs, and MMORPGs, there are many calculations to be done. Maybe its possible to simplify the calculations a little bit. (Maybe it also takes away from the immersion(which in the first place was the reason you''d implement it.) But I think it''d work better than no emotional AI at all. Or maybe it''d end up being a hack job, and annoying,I guess it depends. For one character (yourself) it might not be so difficult. As far as herds of cows and sheep, oof! Although maybe it would be possible to simplify the herd behavior to one entity, herd, which controls the general action of all cows and or sheep, and each has like a FSM. So if an outsider came near, maybe they all think: RUN! And all run. Or if their owner comes to feed them they all crowd around. Although thats not really emotion anymore thats kinda rote, but it''s a little bit more realistic. I don''t know if it should be implemented for everything, as the AI gets more and more complex. The basic challenge of being a game designer is designing something that will be fun and exciting all the time. It was my experience that I was bored to death playing creatures. (1) In RPGS, it is a little difficult for NPCs to have complex enough AI to always be intriguing, and set up new and exciting adventures on their own, and at the same time worry about emotions. Maybe it''s all in the implementation.

-=Lohrno

quote: Original post by astrum
Hmm.. everyone is so worried about AI eating up processing power. This would only work in the case of mmorpgs, but in theory if you have a section of the servers (in our case an individual network) devoted to doing nothing but controlling AI all you need to do is send the actions or results of the AI descision to whomever is close enough to see it or be affected by it.

Yeah...you COULD but cost/benefit...If you have 10 servers on an mmorpg, which each need 5 computers to control it AI, at 500$ each, you''re talking about $25,000 less in profit. While thats not TOO much, my estimate of 500$ for a computer is conservative at best. Increase it to $1000 for pretty good computers, you''re talking about 50k, $2000, 100k. To do AI REALLY well, you need a lot of power. Whether that just comes from future technology, or just buying more computers, it''s definitely something to consider. I''ve thought about this idea too, it''s an okay iddea, just as long as the money is there. Of course since we''re kinda pie in the skying, Okay, we have the money to do this hehe. The other problem with this is what if you are running them from Windows. (I''m not a linux guy but in my experince they DO crash eventually, you cannot keep a windows machine up 24/7 indefinately.) So, what happens when the AI servers go down? Do all the creatures stop moving, like a strange episode of the Twilight zone? Do they all dissappear so as to prevent exploits, this rendering the world suddenly a very barren and lonely place? Both of those have their abilities to be exploited if there are any kind of set spawn points. Or even you could have the server just go down when one of the AI servers goes down, but thats not very nice, and you will get griped at. Just some things to consider! =D

-=Lohrno

quote: Original post by astrum
Hmm.. everyone is so worried about AI eating up processing power. This would only work in the case of mmorpgs, but in theory if you have a section of the servers (in our case an individual network) devoted to doing nothing but controlling AI all you need to do is send the actions or results of the AI descision to whomever is close enough to see it or be affected by it.

I''m not so much concerned with the AI processing power as the bandwidth that it would take to send the position/state of all the animals. If you were ever downtown by the bank in UO, you know how much things lag when there are multiple entities on the screen.


Dave Mark
Intrinsic Algorithm Development
"Reducing the world to mathematical equations!"

hi all, great discussion (i confess i haven''t read it all yet) i was wondering if a lot of the problems inherent in mmorpgs couldn''t be handled in a different way. starting with why mmorpgs are attractive, that is what mmorpgs offer that other online game experiences dont, im wondering if a totally different gaming model can provide the same experience, except maybe without some of the bad parts. i know this isn''t strictly the thread topic, but the two are closely related i think. rather than beat around the bush, here''s the idea:

play is based on the traditional, pen and paper rpg model. if you know it, skip the next paragraph.

in pnps, you have a group of players (ideally 3-6) and a referee. the players take on the role of characters living in an imaginary world. the referee takes on the role of storyteller, describing that world and what the characters, see, smell, touch, etc., to provide an interesting experience for the players. the players describe their character''s actions in response to the referee''s storytelling, and the referee attempts to balance things out and make an interesting adventure.

moving this to the pc, you have 1-10 players, and a referee, who takes on the task of breathing life into npcs (bots) like allies, enemies, people on the street, etc. as long as the players are only interacting with one npc at a time, the referee can breathe life into a character, bypassing the dull AI problem altogether. theres also the possibility of multiple referees, or no ref, but thats another page the ref could also change things about the adventure he didnt like on the fly, create new monsters, etc.

any thoughts?

oh, i know NWN is doing this, my point is, cant this model provide everything a MMORPG provides, and more?

Part of the problem with that is it''s not exactly massive anymore. The cool thing about PnP RPGs is that the GM can make up stuff as he goes along, or can prepare a very long adventure. With a computerized version, you either remove the DM, or it takes longer for the DM to prepare stuff. Really, to plan out an entire level, or to write it down on paper, its easier and quicker to write it out on paper. And if the player does something unexpected in the computer version, there will be some kinda barrier to him, or it will just fall out of the realm of interesting anymore (IE going to just the normal part of the world where there are no quests.) So its quicker for DMs to prepare for PnP RPGs, they can last for hours, whereas something that a computer RPGs DM prepared for 2 hours might last 20 minutes. Yeah, theres premade stuff, but where''s the fun in that?!


Innoc- Yea, that''s a pretty big consideration. I mean not to offend your Ultima sensibilities, but also I dont think their network code is optomized. Is that beside the point? Yep, thats beside the point, 10 entities that require a little chunk is already eating out of the poor people who have modem''s bandwidth. I''m excited about the future where we''ll all have fiber, and live in almost real virtual worlds too, but right now..uh, there are some serious technical issues! But then, where''s the benefit? I mean as much as I''d like to see a herd of cows, I would not like to see herd of cows. Who cares about damn cows? =D It''s cool for 10 seconds that you can scare them all, but it wouldnt necessarily make a selling point for the game. "Buy this game, it has cows that react realistically!" Swarms of insects on the other hand...=D Just as long as the swarm itself is one entity. (Why make 200 entities for something that acts as one anyway?) It''s just the little things you always have to look at! Make sure that what you''re implementing is not only very immersive, but that it really adds to the experience. That of course goes for emotions as well as long as it can be implemented correctly. I think in the future we''d be more likely to see characters that try to have some soul than herd animals! =D But really? are herd animals interesting for MMORPGS? Maybe like pack animals that work in groups of 5, but I think no more than that. Unless of course your MMORPG is a Cowboy thing. I mean, If you see a house with a fence around it, 2 cows, a horse, and a pig running around, you get the idea, its a farm. If you see the same house with 20 cows, it''s also a farm. So yeah, doing herd animals right now is kinda unfeasible, and generally not so desirable.

"Part of the problem with that is it''s not exactly massive anymore. "

that was part of my question, whats good about massive? you get to interact with a society, right? you want wide open spaces, right? you get that social effect here to a great extent, without the "britneys kewl" effect, or lag, etc. if you look at nwn, you''ll see you can make a persistent world without needing it to go mmorpg. thats the idea here, and i very much like the portal system theyve got with NWN too.

"The cool thing about PnP RPGs is that the GM can make up stuff as he goes along, or can prepare a very long adventure."

the editor will be capable of makeing looong quests, as well as adding/ editing on the fly. for big tasks (say you want to add a vampires castle where there wasnt one a second ago) you''ll probably have to call "intermission", load the castle if youre lucky enough to have built it already, -or spend ten minutes setting a simple one up in the toolset if not- and start the adventure up where the game left off.

"With a computerized version, you either remove the DM, or it takes longer for the DM to prepare stuff."

huh?

Really, to plan out an entire level, or to write it down on paper, its easier and quicker to write it out on paper.

not by much. tile sets and a GOOD level editor can make things go pretty quick.

"And if the player does something unexpected in the computer version, there will be some kinda barrier to him, "

like what? the levels will be pretty expansive and there are plenty of options for dealing with the edges of maps (portals being a good one)

"or it will just fall out of the realm of interesting anymore (IE going to just the normal part of the world where there are no quests.)"

the whole point is, the players schedule a game, the ref has the adventure ready, and you get a storyline that unfolds over one or more sessions (who wants to play more than six hours straight anyway?) you dont wander around with nothing to do, ever. the refs job is to keep you busy (which is fun in and of itself btw)

"o its quicker for DMs to prepare for PnP RPGs (not by much), they can last for hours (so will these), whereas something that a computer RPGs DM prepared for 2 hours might last 20 minutes (only if done by bad referees)"

"Yeah, theres premade stuff, but where''s the fun in that?!"

?

thanks for the feedback, btw

ok, i reread and think i understand you, you''re saying it sounds a lot like deathmatch or something. there will be coop levels with the game and available as addons, designed by professionals, in addition to whatever the community comes up with. and campaign worlds, characters, creatures, etc. all the persistency will be there, just without the massive part.

quote: Original post by Lohrno I mean, If you see a house with a fence around it, 2 cows, a horse, and a pig running around, you get the idea, its a farm. If you see the same house with 20 cows, it''s also a farm. So yeah, doing herd animals right now is kinda unfeasible, and generally not so desirable.

Yup. That''s the way I view things. Since it is all representative anyway, as long as I get the point and they save the cycles for stuff that actually MEANS something. I guarantee you this sort of things happens in design sessions all the time. "Uh... yeah that would be neat... but what do we have to give up for it?"

That''s the sort of thing I was getting at in responding to DrMol in the first place. It''s not that he''s such hot shit for thinking up ways that games suck. We all do that. Hell... dev teams probably look at their own games that way quite often. The point is - WHY do they not have the features we all pine for? Certainly it isn''t because no one has thought of many of the ideas. Now if DrMol wants to start trading emails (or posts) with people on how to SOLVE various TECHNICAL and LOGISTICAL problems (and even political ones) that go into actually making a game rather than bitching about one, I am sure he will get a far favorable response on here.



Dave Mark
Intrinsic Algorithm Development
"Reducing the world to mathematical equations!"