new New dev directions.

[Aciid]

O'haio long time no see.

 

I got back into A3 and directly hopped into see the status of ACE3, CBA and Alive very much to my surprise they've all grown to stable now.

So after fiddling around with a couple of these old school perpetual mission engines, I thought on how about incorporating Alive into the storyline of IA for example and get into replacing the opfor logistics with Alive.

 

Simply put this would allow the mission to scale as per how many users there are without the need to code each mission's logistics individually. Of course fine-tuning is inevitable, but Alive's components give a good base on staying scalable. 

 

What already is set as preferred non spammy and agreed that theres not more than 1 side-mission and 1 main-mission at given time.

But into the logistics now, they are now more less static. Alive allows parameterization of logistics  so for example a main mission with 10 players on the server we'd figure at-least one team's worth of players are fighting towards to accomplish the main-task.

 

The mission would then keep logistic re-inforcement around at ~40 units, as per before triggering a condition which would stop re-inforcing for example killing at-least x-amount of this mission's opfor units or killing the teamleader's radio comms or vehicle.

 

On a much more grande scale Alive's modules can spawn somewhere up to 80-250 strong companies and even 400-800 strong battalions AI re-inforcement, this of course causes some latency and gpu issues with lower-end boxes. But given more "life-like" scenarios these re-inforcements as a game mechanic are not just that the guy spawns in a loo and opens the door. No they ride out from given starts and routes to attack route and of course try to over come the blufor force.

 

CBA in this equation would mean more flexible ways to code the actual framework to support the missions, IA already has a way good structure in my opinion I haven't checked the latest globals in the code since the last time I was involved with IA and the dev efforts of mysql connectors. CBA adds a ton of macros and eventhandlers which extend the sqf's capabilities in such a way that normal ARMA for some reason don't carry. Like 

 

ACE3 in general is just deepens the game and makes it more interesting in way that it strips the arcade way from Arma and makes each step count, whether it be stepping a feet away from the 50'cal to not get deafened or avoiding shrapnel, actually theres too much to even try to explain ACE. So I'm fairly sure that anyone interested in it can checkout if it's suitable for them. Of course ACE's medic system is at times a little too tedious to work with if you have men falling left an right and a ratio of 1 combat medic per 10 men. Since you need to patch each limb and give injections of bloods and pharms, it can be disabled per mission.

 

But something to think about I'm looking for people to think this forwards with (I literally don't know anyone involved with current IA, would be great to get to know you guys again) , the new style of Arma3 servers that have been popping up are all great but I just can't put my finger on what i'm actually looking for. I've tried Epoch, KOTH, Warfare, COOP, Battleryoale, Life. 

 

I think something what is missing is truly a in-depth COOP mode that is easy enough to jump in but has that much depth that it can keep you interested in a way that IA has kept many of us in the past years.

 

Cheers

[ranarick]

well the idea about the reinforcements riding in and not just popping up seems good to me but the only problem i see for IA is that it runs on the public non modded server so a mod like ACE is a no go. I Would say talk to the other AW devs a bit and maybe you can get started on something . And if you really want ACE into it well we got EU3 wich is running ACE among other mods

[MDCCLXXVI]

Good luck with your 3 FPS goals

 

In reality you have a 100-150 AI performance budget on a good dedicated box before server and client FPS just dies. Cut those numbers by 25% if they are in combat mode.

[Aciid]

Thanks for the rest assurance and optimism ( no need to get upset ) , I don't think resources should even be considered as obstacle when planning good gameplay. Nowadays DC xeon core @3.5ghz goes for 0.10€/hour.

 

Also considering that spawning enemies that are not observed by player's should even be considered at this scale.

And having hundreds of AI enemies was just an example how the logistic spawn ratios work.

 

Somewhere around total 100 AI's but with different logistic mechanics could be how it works. I've tested Alive at with 200 infantry, it seemed pretty runnable and this was with my own 4970k. 

 

Of course if there's nothing to discuss over this I think I'll go on my own separate way. I'm not saying that this should be "IA 2". I'm just trying to gather interest. 

[Nicolai]

Last time I had a look at Alive (0.9.9.1506221), the ammo crates spawning in with RHS (I believe it was RHS 0.3.8) groups were unusable by the RHS units, as the crates were virtual ones, which additionally only contained ammo for vanilla arms. I don't know how you intend to configure your Alive modules with regard to the spawning in of supplies - for me leaving the supplies deactivated made it quite odd, as well. The option of not spawning supplies, circumventing that issue, resulted in numbers of combatant groups without ammo left after a while of battle. Non-virtual crates have been added to RHS in 0.3.9, however, I didn't verify that's working with Alive, yet.

 

As written somewhere before, I ran into the same performance bottlenecks, that everyone runs into, resulting in having to run a dedicated server with headless client managing the AI, as soon as you intend to have a few real human players with you in the field.

 

Even if that setup was working with regard to supplies first and if it was performing comparably to a setup with DAC secondly, you would still have the mentally challenged vanilla AI controlling the units in the Alive.

 

So, thinking things furtherly, by adding some more sophisticated AI mod to that 'large scale scenario', like Vcom_AI, my performance was completely gone.

 

It would be an expensive price to pay for that feature of riding in of the reinforcements instead of them popping up, when implemented that way and understanding everything said correctly, I'd say.

[MDCCLXXVI]

Thanks for the rest assurance and optimism ( no need to get upset ) , I don't think resources should even be considered as obstacle when planning good gameplay. Nowadays DC xeon core @3.5ghz goes for 0.10€/hour.

 

Also considering that spawning enemies that are not observed by player's should even be considered at this scale.

And having hundreds of AI enemies was just an example how the logistic spawn ratios work.

 

Somewhere around total 100 AI's but with different logistic mechanics could be how it works. I've tested Alive at with 200 infantry, it seemed pretty runnable and this was with my own 4970k. 

 

Of course if there's nothing to discuss over this I think I'll go on my own separate way. I'm not saying that this should be "IA 2". I'm just trying to gather interest. 

 

For what its worth, I like the idea.

 

I just think its difficult to keep 50+ players busy (and not bored) in CO-OP, didn't mean to offend.

[Aciid]

Last time I had a look at Alive (0.9.9.1506221), the ammo crates spawning in with RHS (I believe it was RHS 0.3.8) groups were unusable by the RHS units, as the crates were virtual ones, which additionally only contained ammo for vanilla arms. I don't know how you intend to configure your Alive modules with regard to the spawning in of supplies - for me leaving the supplies deactivated made it quite odd, as well. The option of not spawning supplies, circumventing that issue, resulted in numbers of combatant groups without ammo left after a while of battle. Non-virtual crates have been added to RHS in 0.3.9, however, I didn't verify that's working with Alive, yet.

Hey, I think you were in TS when I made a visit one night this is clearly one of the key points to be addressed when thinking about integrating these kind of modules. As the problem with Arma3 mods are they pretty much try to do all the things "better", still share no common ground on extendability or by submodulisation as there are no standards except for BIS functionalities of course.

 

The thing I of course take for granted here is the use of overloadable constructors in code, but I have to say I'm not sure is it feasible in Arma3 without building a few dll's to integrate , but that would take everything all together to a whole different course.

 

 

As written somewhere before, I ran into the same performance bottlenecks, that everyone runs into, resulting in having to run a dedicated server with headless client managing the AI, as soon as you intend to have a few real human players with you in the field.

Definitely interesting to look into would stand-by headless clients work with dynamic spawned AI units.

The last time I was involved with IA's server when we scaled out the first servers was that "oh we just got 250% more performing server, but still we are experiencing lag." Turns out that at the time 2013-2014 it was due to net-code and 100MB/s worth of dedicated line was not enough. But nowadays it seems that ~80-120 players not a problem with 1GB/s lines on King of the hill, Epoch etc. 

Of course there will always be lag due to assets for lower end boxes.

 

Just a couple days ago there was a new binary-release that was stressed with a server that had 129 max-players 

Arma3 more tweaks +FPS PvP,test 129 maxplayers on 94.23.112.9:2102,use v2 performance client 1.52.132709

 

 

Even if that setup was working with regard to supplies first and if it was performing comparably to a setup with DAC secondly, you would still have the mentally challenged vanilla AI controlling the units in the Alive.

From what I've gathered about Alive is that they have kinda backported from Arma2 "custom AI" scripts origin from armaholic and other 3rdparties so that they now have a couple of different AI tactics from military, cqb to insurgency with also different variations of "skill" which I'd hope won't mean just "Aimbot on / Aimbot off"

 

So, thinking things furtherly, by adding some more sophisticated AI mod to that 'large scale scenario', like Vcom_AI, my performance was completely gone.

 

It would be an expensive price to pay for that feature of riding in of the reinforcements instead of them popping up, when implemented that way and understanding everything said correctly, I'd say.

This is just where i'm quite going for is to break down the mission'in even further. Have large scale front's and then small style missions that support the main objectives, such as keeping some pathways open. Keeping hold of key areas for logistics / support. All this may remind you about insurgency style missions in Arma 2, where you had to clear out the AI.

 

But as important would be to keep the gameplay approachable, so you don't need to team up and lay down the road 3km's before every fight. But for example that you could be part of a helicopter squadron as it's still possible to just halo in to the ground with your machine gun and spread out suppresive.

 

 

For what its worth, I like the idea.

 

I just think its difficult to keep 50+ players busy (and not bored) in CO-OP, didn't mean to offend.

Yeah I think I just posted in the wrong forum-area, I tried to get enough attention from prospective people that would find this interesting. I'm not really trying that hard to change IA or anything. This is just something I've been thinking of as maybe a whole different server, not that I don't still love you guys <3