Complaining Of Ai

[Fidcal]

And it somewhat spoils some textures too in T2X. Nothing like washed out flat ambient light to make textures look like stone wallpaper. Pity as the scenery is beautiful.

 

I did manage to tweak my gamma in the config and get max contrast on my monitor to make it look half-reasonable but hard to tell what corner to hide in. One of the first things I look for when entering a new location is somewhere to duck in case of emergency.

[bob_arctor]

Some of T2x had aweful texturing though. I'm thinking that museum. Ghastly, although some liked it.

[Order of the Hammer Bureaucrat]

Personally I'd rather not try to mathematically model the psychology of eye movements for every single AI and every single alert state.

 

I would.

 

That's why they are the clients and we are the developers.

 

wee! i'm a client!

 

Gah I had a great post going, pressed something and it went!

 

"Photopic vision is the vision of the light-adapted eye; in many animals, color vision, mediated by cone cells."

 

"Scotopic vision is the monochromatic vision of the eye in dim light. Since cone cells are nonfunctional in low light, scotopic vision is produced exclusively through rod cells. Vision in normal light with functioning cone cells is photopic vision. Mesopic vision occurs in intermediate lighting conditions and is effectively a combination of scotopic and photopic vision."

 

That's the kind of detail I like to see studied before applying to the game.

 

It is a bit of a leap to assume that realisticly modeled sensory algorithms for AI will so badly unbalance the game as to make it unplayable. It won't at all. What it will do is make the player play in a way that is a bit more real-worldly.

 

Exactly. I have searched for years for realism in games.

 

It's also quite a leap of faith to assume that more realism will make automagically make a better game.

Everybody is free to create his own game. Obviously we are doing this for ourselve. It's not as if we have to appease everybody at all costs. So if somebody doesn't like what we consider as balance, then he can either change it for himself, make his own game or accept it. But we have a very good cut through the Thief playerbase, which is really surprising. If we had tried to gather such a homogen selection on purpose I doubt we would have managed it. As such I think that if our team likes it, most players also will like it, because pretty much every type of player is represented.

 

I have a feeling the legacy will remain for a long time with resources used all over the place. Eventually we will have a holodeck.

 

But most burgalries are not creeping past guards, blackjacking them, or anything. It's nicking stuff while people are asleep, with a knife or gun if they wake up from you making noise.

 

Thief is fiction, it's a game. So it has to have game mechanics. In real life anyone trying to be Garrett would get caught really quickly, and that would be a dull game.

 

Yes, Garrett will have to try it with a knife or gun. Or we should try KOing pieople in real life.

 

Some of T2x had aweful texturing though. I'm thinking that museum. Ghastly, although some liked it.

 

The museum was indeed ghastly. made me not want to play, so i skipped to the next mission.

[Springheel]

Personally I'd rather not try to mathematically model the psychology of eye movements for every single AI and every single alert state.

 

 

I would.

 

Well join up and start programming then.

[Ishtvan]

Re: AI eye rotation algorithm

 

It's not even that great an idea for gameplay. The player's not going to be able to see the AI's pupils from afar and anticipate any eye movement pattern (if there is any definite pattern, which there shouldn't be, realistically), but they are able to see which way the AI's head is facing and anticipate the head turning pattern.

 

From the player's perspective, lumping both rod & cone response into the whole AI FOV means they can observe the AI, how they're turning their head, plan in accordance to that pattern and get rewarded by not getting caught. With random eye movements and different sensitivity of rods & cones, from the player's perspective, the AI visual response is pretty inconsistent, and it turns into a crapshoot whether they get caught or not. Some randomness in AI actions: good. Total inconsistency in AI response to player actions: bad.

[bob_arctor]

Well said.

The movement= visibility is already in there I presume: if you move the lightgem brightens

That's all you need to simulate the eye really. Phototic/scotopic vision wouldn't really add much.

[mr_luc]

First up: hello! I used to be a regular on TEG, and left it when, as a 17-year-old, I got the chance to work for a startup and zip out of my dead-end job at a steakhouse.

 

I've returned to the world of things Thiefy after receiving an email about an old, semi-promising FM of mine that I'd walked away from. In fact, I'm actually having the HD's out of my old computers, that I'd since given away, shipped to me -- because I figure, if nothing else, there is some really nice brushwork in a couple of places including some of the best cliffs/stream I've seen in a DE game, and knowing the effort that goes into making anything in DromEd, I figure someone should benefit from that even if it's not in my mission.

 

I only just became aware of the existence of DarkMod, and I have to say: this is fantastic! I'd experimented with the Serious Engine's editors, which likewise give you real-time feedback, and I have to say that the idea of using a really powerful, real-time, proven csg editor to do thiefy things -- in an engine that screams, like Doom's -- seems almost too good to be true.

 

I'd expect that various movement-related subtleties could pose a challenge, like mantling, but the biggest challenge looks to be AI -- if you need any AI help from someone who hacks Common Lisp day in and day out, give me a yodel. (My dream is actually for some engine to make the scripting language support so clean that I could use cffi to plug in clisp for high-level scripting; I know that in the Torque engine, some people recently added support for Lua scripting).

 

I'll probably ask to be a test mapper at some point, actually.

 

What sort of scripting abilities are there for the doom engine?

[Dram]

I'd expect that various movement-related subtleties could pose a challenge, like mantling, but the biggest challenge looks to be AI

 

Hello and welcome to the TDM forums

 

I'm glad to be able to tell you that mantling is already finished (and is extremely good, much better then in thief 1/2 or 3)

 

AI is in development and so far they're going great

 

I'll probably ask to be a test mapper at some point, actually.

 

We have several Team Mappers and several Beta Mappers atm so if you wish to apply for a position in either please send an application to recruiting@thedarkmod.com and New Horizon should get back to you soon.

 

What sort of scripting abilities are there for the doom engine?

 

Well, the SDK allows many things except access to the renderer, however you can do many many things with scripts. You can make a script for a map to make for example a monster appear after a certain time somewhere etc. So in general, D3's scripting is very good.

 

 

Hope that answers your questions

[mr_luc]

Hello and welcome to the TDM forums

 

I'm glad to be able to tell you that mantling is already finished (and is extremely good, much better then in thief 1/2 or 3)

 

AI is in development and so far they're going great

We have several Team Mappers and several Beta Mappers atm so if you wish to apply for a position in either please send an application to recruiting@thedarkmod.com and New Horizon should get back to you soon.

Well, the SDK allows many things except access to the renderer, however you can do many many things with scripts. You can make a script for a map to make for example a monster appear after a certain time somewhere etc. So in general, D3's scripting is very good.

Hope that answers your questions

just googled -- I detest the use of $ (or %, or anything else) to indicate symbols (ninja edit: I mean, variables), but the syntax looks pretty friendly; I guess I should download the sdk and just start reading.

 

That's quite cool. This is all quite cool. It knocked me on my ass; Thief 1 and Tribes 1 are still my favorite games of all time, lots of love for ttlg and friends of ttlg.

 

I'll probably apply to be a beta mapper in a week or two; I have yet to buy Doom 3, heh, so I need to do that first and hopefully familiarize myself with the api it exposes.

 

So cool, thanks a million for helping us keep our taffin' fun alive!

Edited March 10, 2006 by mr_luc

[Dram]

So cool, thanks a million for helping us keep our taffin' fun alive!

 

Hell, that's why this teams here

[Gildoran]

just googled -- I detest the use of $ (or %, or anything else) to indicate symbols (ninja edit: I mean, variables), but the syntax looks pretty friendly; I guess I should download the sdk and just start reading.

Then you'll be happy to know that $blah isn't used for referencing variables. $entityName is more or less shorthand for sys.getEntity("entityName"). Actual variables are referenced just by name, like in C. BTW, you don't need the SDK to use D3 scripting; D3 can read scripts without the aid of an external compiler. All you need is D3 and notepad.

[Ishtvan]

Yeah the script is kind've modeled on C++, although some things are missing like arrays and structures. We definitely have a lot of work left to do on the AI, so we could definitely use some help in that area. We would have to think about how to divide up the tasks though since a few people have ended up working on various aspects of the AI at the same time.

 

Anyway, if you want to help out with programming, I'd say practice D3 scripting a bit and then apply in the "help wanted" section.

[Ishtvan]

@mr_luc :

 

You still have any interest in helping out with the AI? We haven't heard from our AI coder in a while, he might have disappeared.

[sophisticatedZombie]

I went transdimensional for a while. I focussed my 40khz nanotubes on the 5th pentagon side too long and got taken to the galactic node.

 

Anyway, I'm back... but could certainly use some help. If you want to write AI code talk to sparhawk and see what he thinks of the idea.

[sparhawk]

I wouldn't wait to long.

[Ishtvan]

Sorry SZ, at least I just said "might have" disappeared. For a while it seemed like we had some sort of curse where everyone who worked on the AI disappeared, we went thru something like 3 people before.

[sophisticatedZombie]

I think it is the shear enormity of the transformations required to do anything with FPS AI that frustrates people away. At the moment, I'm having fun, but there was alot of grunt work for a while.

[kohan69]

Re: AI eye rotation algorithm

 

It's not even that great an idea for gameplay. The player's not going to be able to see the AI's pupils from afar and anticipate any eye movement pattern (if there is any definite pattern, which there shouldn't be, realistically), but they are able to see which way the AI's head is facing and anticipate the head turning pattern.

 

From the player's perspective, lumping both rod & cone response into the whole AI FOV means they can observe the AI, how they're turning their head, plan in accordance to that pattern and get rewarded by not getting caught.

 

 

Exactly. One thing that extremely annoys me in some stealth games is that the NPC's FOV/cone is in the same position while the NPC is performing an idle animation of stretching or looking around, but I'm sure this was already addressed in TDM - right?

[Ishtvan]

Yes.

[Domarius]

Yeah, in T2, there was an idle anim that would look directly at you and freak you out, but you soon learned that the actual FOV wasn't really moving.