Maybe a good idea... about trafic and pedestrians !!!!

[Guest]

A lot of MTA Player want pedestrians and trafic in game... maybe possible if something exists : i make a list of what we need and how to manage that.

1 ) We need a pathfile ( with vehicle trafic and pedestrians ) with marked sequences.

2 ) When client connect, server have just to send the marked sequence to client.

3 ) Client have to manage their own pedestrian wich work with the pathfile at a precise sequence sended by the server.

About the choc with vehicle, pedestrian etc... the client have just to get other players's positions and manage their own choc and their own pedestrians.

With a marked sequences on the pathfile, server will run without more work. Just a little bit of upload.. maybe 1 or 2kB/s hummm...

Give me your opinion !!!!!

[[->>CURLY<<-]]

i love new people on the firs post making good sugestions but then u have to tell them to get a nice hot worm cosy mug of stfu and use the search butten so that they can find out not 2 post crap again

[ijsf]

stfu curly..

I think it might be a good idea, although i don't know how the path thing exactly works, for example with abnormal positions (when you park your car in front of a ped, so he has to walk around your car)

[Boss]

Okay,tell me if I'm saying crap,but I think those exceptions are hard coded in the exe.

[Oyst]

stfu curly..

Well put, dont be so god damn rude, hes trying to make a suggestion.

Anyway, that sounds intresting. Wouldnt there still be a problem of "server lag" if they were to be added?

[Boss]

Yes,but I'm sure it wouldn't happen with the peds,but...

[eAi]

Sorry this idea won't work. The peds don't follow a path strictly. They don't exist when they aren't near the player and are created at random (not seeded)...

eAi

[Cerbera]

Sorry this idea won't work. The peds don't follow a path strictly. They don't exist when they aren't near the player and are created at random (not seeded)...

eAi

Yep.

I'm totally with Curly on this one and am totally bemused why IJs isn't as well. This has been discussed to death and has been shown unfeasable every time. By running a simple forum search folks like eAi and myself wouldn't have to keep posting the same response because of people making the threads.

I'm sure you meant well but running a search would have been much more helpful.

GTA1 and 2 (which had civilian peds and traffic) use a TOTALLY different technique to...just about every other multiplayer game that exists! Game engines are causal, of course, and GTA1 and 2 exploit this by constantly and precisely synchronising the game times. This is why you have to wait for everyone to have joined the chat screen in GTA2 before you can actually begin the level; it means everyone starts the level at the exact same game time.

GTA1 and 2 both ran with a fixed frames/second (25 to be precise) although they could skip chunks of frames out if load was become too great. Because of this, essentially everything is set in place before the first frame is ever drawn, whilst the level first loads. Once the first frame is drawn, user input is detected, the result processed and the second frame drawn. Whilst user input is being listened to, the engine is creating/moving pedestrians according to the player position and game time.

This is why the civilian traffic and peds sync up in GTA1 and 2 fairly well, although you do get sync failures occasionally. This method is impossible to implement in MTA because it would prevent people from joining games that were already running. Additionally, the response times between user input and the action of the actor apparent to everyone else in the game is so large that the additional strain of trying to manually synchronise SCM-created traffic makes it unworkable as well.

I've made posts like this several times, so can we always have topics like this locked on sight in the future?

[Mr.Bill]

hmm well it hasnt been posted for a while and the way mta was progressing there might of been some light on the situation.. maybe not...

what we should do is just take pics of peds, print em out, cut the outline and use sticky stuff to put it on your screan!

[Brophy]

i dont like this idea, it would interfear with gameplay like when trying to get away from a mad posse or racing

[dystopia]

even if impossible, at least it's logical thinking, a lot better input than say we get from blaze eh? lol

[Guest]

i dont think the people that are flaming this guy are reading exactly what he said.

he didnt say lets make the peds sync with both computers, he said make a time file with paths that they must follow so that nothing needs to be syncd period.

[eAi]

I read. IJs discussed it with me (after he posted) and we came to the conclusion that I posted previously.

eAi

[eAi]

I thought again about this. And i thought of an idea. And it works. Mostly.

We'll see

eAi

[dystopia]

seems to me

in theory: car sits slightly left to npc, npc adjusts to go right around the car, radjusts once past car, as in game, every players game has the same AI

in practice: car doesn't exist due to lag, car is stuck in ground due to lag, car blows up on one screen and not another, npcs will act differently based upon something like that.

I see at least some reasons why you say mostly, seems like it'd have to wait for the more efficient new netcode y'all are working on.

[eAi]

The problem with your "in theory" bit is that the pedestrians/actors aren't all in the same place on each client and they randomly interact. The key to the problem is to remove the randomness or at least seed it.

eAi

[Guest]

Humm, i found paths.ipl in \Data\maps\

maybe is that file wich generate paths for trafic.

[Cerbera]

The problem with your "in theory" bit is that the pedestrians/actors aren't all in the same place on each client and they randomly interact. The key to the problem is to remove the randomness or at least seed it.

eAi

I'm pretty sure that my analysis of the civilian actor problem in GTA3 and Vice was correct. Computer games are not (and actually cannot) contain any truely random function. The only chaotic factor is user input but the results it causes are decided by the game engine and therefore follow rigid systems. Nothing random about it whatsoever.

The problem is making sure the game engines are making thier next-frame calculations based on the same environment. Meaning that all vehicles MUST be in the precise same positions, all actors in the right positions facing the right direction with thier shooting precisely matching. If any slight deviation from perfect synchronisation occurs, the magical effect of the serendipity of the game engines is lost.

Because Vice and GTA3 do not run thier calculations based on the frame rate, ensuring this snynchronisation is not broken is all but impossible. If it was easy to make playable multiplayer R* would have done it, just like they did with GTA1, GTA: London '69 (and '61) and GTA2.

It's a bad and unworkable idea, so trash it along with the others I say. Creating a street racing mode and/or distributing an SCM kit with which people can add multiplayer support to thier own map mods (as is so desperatly needed with Dexxland, Majera and as would benefit Myriad Islands so well) should be the number 1 priority for the MTA:VC team. I've harped on about this in other threads and nudge the MTA team every month or two with great positive response, but no actual work comes from it.

[eAi]

Yes, you're right it can't be truly random, but as good as really... But anyway, your theory relies on the pedestrians being always moving, but in gta3 and vc the peds are only loaded when they're near you. Otherwise, they don't move.

My method that i'm developing works by removing control of peds from the game tottally, and "out sourcing" it to my program. This has been mostly successful... Just need to work on the more complex stuff (such as actors following paths in reverse and interacting)...

eAi

[Cerbera]

Yes, you're right it can't be truly random, but as good as really... But anyway, your theory relies on the pedestrians being always moving, but in gta3 and vc the peds are only loaded when they're near you. Otherwise, they don't move.

My method that i'm developing works by removing control of peds from the game tottally, and "out sourcing" it to my program. This has been mostly successful... Just need to work on the more complex stuff (such as actors following paths in reverse and interacting)...

eAi

Ah, a memory hack program *could* use the causal nature of computing to it's advantage to synchronise much as GTA1/2 do. However, you are incorrect about pedestrians always moving in the early GTA's. They were drawn precisely the same as GTA3 and Vice - think about it; having a square kilometre of animated pedestrian sprites would require more processing power than most machines in the mid nineties could muster.

(EDIT) The worry with memory hack type programs is that they drain system resources quite drastically on many machines. As most players don't have the sense to turn down thier resolution to make Vice run correctly in NORMAL play and few turn it turn further to allow full speed in MTA:VC, adding yet another rapidly updating program to memory will surely cripple many players' computers and create unplayable levels of lag for those in the game with them?

[eAi]

Yes. The option would be to use a "walk to point" method. Haven't really played gta/gta2 etc much so i can't really comment on it propperly

eAi

[Guest]

(EDIT) The worry with memory hack type programs is that they drain system resources quite drastically on many machines. As most players don't have the sense to turn down thier resolution to make Vice run correctly in NORMAL play and few turn it turn further to allow full speed in MTA:VC, adding yet another rapidly updating program to memory will surely cripple many players' computers and create unplayable levels of lag for those in the game with them?

i beg to differ cerb, i run the game at 1280 and play for 5 hours strait at a time and i never notice any slowdown, it depends on ur system specs, thats the only thing that tells u what rez u can run the game at.

[Cerbera]

(EDIT) The worry with memory hack type programs is that they drain system resources quite drastically on many machines. As most players don't have the sense to turn down thier resolution to make Vice run correctly in NORMAL play and few turn it turn further to allow full speed in MTA:VC, adding yet another rapidly updating program to memory will surely cripple many players' computers and create unplayable levels of lag for those in the game with them?

i beg to differ cerb, i run the game at 1280 and play for 5 hours strait at a time and i never notice any slowdown, it depends on ur system specs, thats the only thing that tells u what rez u can run the game at.

How is that different to what I said? Most people simply don't bother to find out the limit they can run the game at without slowdown, they just put up with it. The minority like you, eAi and myself will no doubt have tried different settings to see which works best for thier system. The keyword is "Minority" though, and I think you'd agree with me on that.

[Guest]

i had actually thought of this idea this morning and came on here to see if somebody else had suggested it. and he has suggested it just as i thought of it however i think many people are not interpreting it properly.

basically what would happen is everybody will have a pathfile (this means everytime you play the peds will do the same thing but that could be fixed later on with multiple pathfiles). When you connect to a server the server will tell the client which pathfile the game is using and the time it is currently up to.

Now what will happen is the client will know where every pedestrian is at once because they will all be following a set route. This does not mean it has to render all of them, it just renders the ones in its range based on the pathfile, so the CPU is always calling on the pathfile instead of randomly creating pedestrians.

when something random happens such as a car blocking the route of pedestrian cars or something then the server will send this info to the clients that are within range of the change in path and their paths will be altered to suit the new pedestrian flow.

This pathfile could also be implimented in single player and it would make the game much better (for example when a car you like drives past but it goes out of range, when you steal another car to catch up its gone, the pathfile will fix that)

i hope this clears things up/adds new ideas to the mod crew

[Posty]

Why did you bump this topic? If you read through it, you probably saw reasons on why they wont/cant do it... Dont bump topics months old just to say you thought of it too. If its been suggested, and is in an old topic, just leave it, okay?