bad benson

the thing is, nearestobjects is heavy and if i don't check often it's easy to miss someone standing on a hill totally visible eventhough further away. and if i use neartargets i might aswell use the values in there instead of cross checking with gethidefrom. not gonna get extra speed from that. what's also gret about nearTargets is that it let's the engine decide who and what qualifies as a target. still gonna totally play around with it though when i can. @ceeeb: yea. i don't get the talk about weird naming, atleast when it comes to knowsabout. what is knowsabout supposed to represent if not knowledge? i mean i get that someone might still know about you if you just vanished behind a large house. but i never saw the value behave in a way that it would represent those nuances. so i'd be curious about how others actually use the output and what certain values mean to them. all i got from it was 1.6 = spotted. when it decreased in older games, what was it based on? also only large distance or blocked line of sight too? i remember it always being not very reactive or dynamic. atleast not in a way that would represent anything human/realistic. that's why i'm curious about people's interpretation of its values.

not sure what you mean tbh. what other things did i say i used it for? can you elaborate? maybe i have a wrong image about the command. i was talking about how the value increases and how it only (supposedly) decreases based on large distances (viewdistance) and not actual line of sight or sound when you use it for units. i honestly never used it for sides or other stuff. talking about _unit1 knowsabout _unit2 output kind of thing. i'm not advocating to remove knowsabout. i wasn't even reading the last page tbh. everytime i debugged the value with two units it was total shit and useless. not dynamic at all. went up in seemingly fixed steps never really decreasing unless you create extreme conditions like large distance or death lol. i just find it pretty shitty for unit specific stuff. maybe i'm missing something so pls enlighten me. this based on tests and all the similar discussions that popped up over the years about being able to reset the value and the general fact that it behaves suboptimal. that's why i would love that perceived position thing a lot. _unit1 perceptionaccuracy _unit2 or something. oh wow. i never knew about this one. sounds pretty much like what i need. thx! can't test it for quite a while but will do as soon as possible. you are right about getting the enemies in the first place though probably. might require additional nearestobjects which would really suck for large distances. still great to know about it. someone should link those with the knowsabout page on the wiki.

yea AI related EHs would be epic! with the Raptor AI in Jurassic arma i used "nearTargets" and the accuracy values in its array. "knowsabout" was really shit for anything detailed anyways tbh. especially when it comes to reversed stuff like detecting if a unit does not know your location anymore. knowsabout pretty much turned out useless for that (might have changed but i doubt it). but it still has its prupose for simple switches. like using the 1.6, or whatever it was, as a trigger to detect full on combat mode for the first time liek for a stealth mission or something. although handling of the "neartargets" array is probably not so fast. so just having the distance between perceived position and actual postion as one value you could pull or even an EH would be super useful. that's the only thing i ended up drawing out of the array anyways. that value reacts super fast and accurately. it's pretty much what you'd expect from something called "knowsabout" ;)

have you compared FPS before and after these changes? not saying it didn't change, just that they might as well not have had any noticable impact as most performance related things so far and that without a measurement the info is worthless. especially since IF there was a change it was pretty hard to notice probably. i have yet to see mass occurrence of posts about performance increase. i mean i think the biggest thing was 5 fps for server performance or something along those lines when Dwarden made an effort. which was nice. but that still won't fix client performance i think. not saying they aren't trying but we all and them especially know that without major changes there won't any major improvement. with so many moving parts small tweaks can even back fire. i think i even read DNA stating that himself somewhere. it's not so much about End Game performance anyways. more about general MP performance in anything that is more than 5 player coop with next to nothing large scale or script wise happening. so waiting for the End Game devs to fix that is probably pointles anyways.

hah! spot on. made me giggle. thx

as i said. i have no pc at the moment whatsoever. i can't run arma on this old mac laptop i'm writing from. that sounds like it's broken. have you followed any of the advice i gave about general mission stuff? have you tested with like 10-30 raptors placed in an empty mission? does that work? that was a number that always worked in the tests during developement. i personally only tested the raptor AI in my own lightweight survival MP scenario (which i might release) with dynamic spawn and clean up and with simple mass tests in otherwise empty editor missions. the result of the latter was that even large numbers of raptors hunting you at the same time "works". FPS dies when they all attack (claw and bite) but the search routine was still working eventhough delayed by seconds depending on count. this all sounds more like they are just breaking like what you described happens with the modules. so rule of thumb: don't try to use them like normal AI in overloaded scripted missions. even normal AI is hungry. but the raptor AI is not prioritized by the engine like the vanilla AI routines on the engine level are. but who knows how the game itself has been changed since release. no point in speculating until i can test again myself. i had a friend confirm to me that they still work when used carefully so all i can do right now is give general advice. if the task modules cause problems, try to ditch them and see if the raptors work then in the long run. anyways. better to wait because right now i can do literally nothing...

@eggbeast: do you guys have viewblock geometry on your clutter? i remember playing around with that in arma 1 for wheat fields. i think the conclusion was that it worked until you got spotted once. then they just knew where you are even if you moved around prone in the ense field. i woner how that behaves in arma 3

i'm pretty sure those are ports of the Marines Huey from arma 2 (grey one) and the old school Huey from OA (green shark face). you should consider using the sample models as a base. they should be pretty solid, especially in terms of LODs and stuff. so you'd just need to update/change the textures a little (if you see fit) and add whatever you want from your own models to those ones. you should ask the CUP guys if they already ported those babies. they will surely share the source files with you so you won't even have to do any porting work. just adding fancy stuff to a working model. in all honesty i haven't looked at your Hueys ingame for a very long time so i can't say if the arma 2 and OA ones are better. just a reminder that the editable source is available. :)

yea i remember them stating that they wanted to avoid disconnect from your character by slowing down its movement in a forced way. i still think it was an excuse though due to a real change/a new system would've been way more work. as you have hinted towards. sight misalignment =/= inertia. especially since the effect, as the name suggest, is limited to when you look through sights. so imho opinion inertia was never implemented. but that's just me maybe. ;) i mean i'm not hating on the misalignment, it's a nice little touch, but it does not fully addresses a single one of the "issues" (depending on what you wanna see in this game personally ofc) with the animation system. it's like a workaround that addresses like 20%. the fact that playing with crosshairs enabled degrades the whole thing to a mere little sight note makes that pretty obvious. if you look at the game in its entirety, it's just something snuck in there without going all the way, arma style. :p i think the way they just fake it by messing with the camera was probably a clever idea some dev had which lead to us getting something instead of nothing aince it was relatively easy to do. what bugs me most is prone turn rate though. i could totally live with the turning speed when upright. but prone is just ridiculous. the inbetween adjusted stances also have no turning animations which is not s big as a problem as with prone from a gameplay perspective but still looks horrible for no reason. another unfinished thing/construction site on the long list.

you will laugh but our approach was to give them NVGs. we made sure that the NVGs don't show up on their head but it will make them have the sight of a proper predatorial animal that hunts at night. this solution was the best since it uses engine functionality instead of adding more and more detailed conditions into the AI cycle. so forst check if you have taken away their NVGs. some people use global NVG removal code for night missions. maybe you have something like that in your mission. i thought they had NVGs by default but maybe not. i bet BIS messed with AI and now i'll have to adjust everything again :/ hm. very weird indeed. this sounds like something in the game changed with how side and enemy relations are handled. maybe at mission start. your fixes make it seem like that. you could also try it the other way around and delay the raptor spawn a little. maybe that works too. when we released it was totally fine. this is kind of frustrating but i will see if i can do something later. no promises though. i definately don't plan on giving this mod a major overhaul. i'll mostly release the internal progress and see if i can do some hotfixes. yea the waypoint thing is the last thing i did before taking a break. so that is 100% confirmed. it will just require you to switch a variable on the raptor in question. easy to use stuff. i will probably have saved some money for replacement parts for my broken pc late in the summer or something. so until then i can't even run arma at all. so you'll have to be a little patient. hm. this sounds either like you are using that other dino mod alongside ours which will cause skeletons and animations being overwritten since the class names still collide. or there is a new bug caused by BIS making changes to the animation stuff in general. hard to tell right now. overall the set up is pretty fragile and as you might have noticed we had to force animations to be played to make it work properly. if we wouldn't do that the raptors would just get stuck and stop working (they would all simultanously be inside a loop running in place) after working perfectly for a short time. hair pullign stuff...arma modding in a nutshell. we never found the reason why that happened so we had to apply this dirty fix for now sadly. one benefit is that they can get to you much faster. it just looks less fluid and natural. i still hope one day whatever hardcoded stuff causes it will be changed so i can turn off the forced animations.

yea i'm obviously not well informed on the matter, because quite frankly i don't care that much about it, sorry :o. i don't see Valve trying to steal mod content anyways so why act like they want to on principle. so that is non issue for me based on the lack of a precedence and logical motivation alone. once you upload stuff on the net it's out of your hands either way. armaholic didn't keep A3L from taking what they liked. i mean i'm all for keeping the workshop legit but i also don't see that happen without reporting by hand. on that matter though i gotta say that armaholic is superior there because stuff gets checked BEFORE uploading. the main thing that makes reporting on the workshop in the long run look so tedious (because it is) is because everyone can first upload stuff and then it has to be found and checked. armaholic going through actual people (foxhound) FIRST skips all those extra steps. maybe that can be taken as inspiration. having more moderators on the workshop and verifying downloads before they get accessible maybe. you can always flag them later if they slip through. they would catch the obvious offenders that jsut want stuff like RHS on the WS but ofc also means more work. i still think some kind of integration of armaholic into the launcher like someone mentioned would be amazing. best of both worlds. maybe through some kind of link files like torrents.

long story short: let's keep arma clunky and broken for another decade just because it's more work than simply superficially updating graphics and assets. sounds like a plan. sold ok before. it's not like adding more and more ontop the old architecture is gonna make substantial changes easier. it's very easy to over and over explain to people the time and money thing, with the same old "let daddy explain you how the world works" speech, everytime a new arma comes out and has the same old problems, because well, it's released now, whatcha gonna do, right? ofc people judge the current game after they get their hands on it. otherwise you'd have to foresee how arma 4 turns out exactly and give feedback on that? how's that gonna work? i'll be honest, i didn't expect the scroll menu to make a come back...but here it is again. tadaaa! bummer...

there's going to be a minor update in the future. i have no working pc (writing from mac) right now so i can't do any work on the AI nor release what i have so far. but maybe Ruppert will repack the latest public version with a key and re-release. i'll talk to him about it. overall the set up of the raptors is pretty fragile (it's arma after all...) and as Corporal_Lib[bR] mentioned the probability is pretty high that stuff breaks just from BIS updating and messing with animation stuff etc. i can't do any tests myself for the time being but here are some general things to keep in mind. - try not having the raptors in groups (they won't share targets but they will also not be any interference by default squad level AI) - generally don't overload your mission with countless third party scripts, this can definately break the AI since it runs a pretty heavy loop to allow a lot of its features - above goes also for number of raptors operating at the same time. each additional raptor will slow down reactions of them overall. not noticable at certain counts but if you go heavy you will notice. just like with normal AI, more units = less performance + more reaction time. in conclusion: you will get the best out of this by making linear horror missions with dynamic raptor spawn. if you do open worldish stuff you HAVE to be economic with all the scripts you use as explained above and you will HAVE to do dynamic spawning. there are countless scripts out there that handle that (EOS for example). the waypoint speed stuff i already fixed irc and i also added a variable to disable and enable the custom AI routines on the fly which also makes them work better with waypoints. these improvements are on hold though until, as mentioned, i have a pc up and running again. i will also need more detailed reports. the raptors are designed standalone. so if you have a problem, first thing to do is test them in an empty mission. so can someone confirm that they don't work at all anymore?

yea sure. but imho it's unlikely to change since it applies to all games on the workshop, doesn't it? why would they change it for one game and its community? but who knows. maybe a separate EULA for arma would be possible. something you have to agree to when you use the workshop a la apple itunes harassment ;) but it sounds probably more realistic than making a new system to watermark/protect files i guess.

@semiconductor: i also think this sounds a little overly dramatic. you make it sound like the AI currently thinks tanks are infantry and you'd have to write a whole new system ontop of the current one to make simple survival moves happen. squad leaders (atleast) already analyse as far as distinguishing if a tank is around or not. how else would they order their squad AT guy to fire upon it? i think AI overall needs a responsive switch between aggressive and defensive behavior. so what if there are other threats around or a friendly team with AT and ammo for it? that still makes the close tank the biggest threat until someone (who cares if it's even blue on blue/red on red) destroys it. defensive mode would mean only fire when fired upon which should already be very easily detectable considering there is engine side suppression detection in place. considering the AI is robotic and non-intelligent in every way indoors (choo! choo! movement on rails) i'd say making them go there is not a good idea though (atleast until we get something htat qualifies as indoors AI). and probably the most potentially mission breaking thing. at the very least going prone and/or towards cover is needed though. not only in that situation. the AI overall lacks these very simple self preservation maneuvers. it's how they should react in general when confronted with superior firepower (like being suppressed the shit out of) imho. arma 3 AI seem more agressive in genreal to me, could be subjective though. it would make firefights more interesting overall if you'd be able to push back forces. have them reset a few hundred meters back and approach again or flank.

i'm by far no expert on the technicalities but to me this all sounds a bit overkill. like people calling for encypted pbos that one time. i also can't see this working so well. sure you could use signatures i guess but that doesn't sound like it would keep people from repacking data into a new pbo creating a new key, right? anything else would require messing with the pbos themselves resulting in people not being able to even look at other people's work for educational purposes, right? is that really what we want? just because there's a problem with the steam EULA. sounds kind of backwards to me. i'd rather have someone (*cough..BIS..* ;)) team up with foxhound and make an armaholic launcher that draws files directly from the site's servers (maybe BI could sponsor additional servers). not sure how hard it would be but there are already tools that do similar stuff in one or the other way (community tools updater comes to mind). i know people are protective of their work for good reasons and i'm like that too in certain cases but let's not pull a "patriot act" move out of fear. there's always good and bad sides to "solutions" like that. like promoting hiding and not sharing of knowledge and an overall overly competitive focus much like the hated life communities do among themselves. and i don't think it will keep anyone from making the tools to go around any type of encryption. armaholic is already the trusted working "system" that is asked for. it's already there. it just needs to get the benefits of the workshop. which as far as i'm concerned are auto download and update. doesn't sound like rocket science to me (especially looking at how many dayZ mod launchers there are). the only thing would be, as mentioned, the required infrastructure to cope with the traffic. i might be missing something though. just thinking out loud.

i was just wondering if it would make any difference, if you could host your stuff yourself and steam would be able to treat that external link properly. just a thought i had just now, didn't think it through, might regret posting it later ;) like treating the workshop more like moddb where you have your page with all the info and videos and all that goodness and then you provide the link with maybe a standardized format (rar or something with a certain folder structure). atm i'm not really sure if steam does all the legal ass covering because they host the stuff literally themselves or if it is simply the association of showing the content on their pages that already requires that. i guess i'm dreaming of a middle way where creators have still full control over the files and are responsible for their contents (and thus carry the "legal burden" themselves) but still the download would be automated and centralised on steam. like addonsync and all these other things clans use with their own servers but linked with steam and the game. although in reality i don't think it would keep people from reuploading who like to mess with the files. but atleast it might get more creators on board and eliminate the "need" for some people to reupload stuff so their lazy ignorant friends don't have to click through some downlod links and install some files into certain folders :p

nobody ever said otherwise, certainly not me. you are the only person drawing an imaginary line. of course they do other things, i sure hope so, how those other things increased performance so far is another story. lol whatever dude ;) let's hope DX12 has an impact because eventhough i appreciate them trying, i don't see any results worth mentioning so far. i mean the game has been released for what, 2 years now? so that's the reason why people get worked up about DX12. it's called being realistic about what you can actual see in, well, reality. which is no gains without DX12. so yea DX12 could be the thing that finally gives this engine a real (even if small) boost since nothing else seems to work. again. show me the gains. what is that well known past patch version that increased performance? good intention to fix something is one thing but if there are no real results it's not surprising that people focus on DX12 implementation rather than what they've heard countless times before, don't you think? there's only so many times you can say "we're working on it" and it'll satisfy people and make them wait patiently. if you like to focus on their good intentions, do that, by all means, good for you. :p

what do you mean? we already have those and so far i couldn't identify them as a performance issue. don't tell me you haven't seen them yet. i have been running around in the editor drooling over them for an hour when they added them ;)

erm nope. there's no and. there's direct causal connection. and even if your interpretation of what they said was true. you are ignoring also Mareks tweet about DX12 having significance for arma 3 and overall the things that DX12 has in store. https://twitter.com/maruksp/status/558735038335172608 if you think they are adding DX12 for water reflections or something then you're just badly informed. as for continuous effort to fix/improve performance. yea sure. i agree. the question is, what are the gains we have seen so far? so i think you just like to tell people to calm down for no reason or something :p hah. good point.

no idea what you're talking about but all i see here is people arguing if one can "nail down performance" by "putting in DX12" or not. saying "nail down performance" is not saying "we might accidently get a few FPS from adding DX12". if you are trying to somehow imply to keep expectations low, then yes, i totally agree that it would be wise to do that.

you could ask the CUP people for the source of the ported arma 2 civs. that would give you more interesting zombies like police men and all that good stuff and you'd actually have editable models to like say rip an arm off or do the deformations that were mentioned, or whatever you wanna do ;) just an idea. because i gotta agree that those dudes in shorts are annoying. but i get it's not your fault and that it's probably much more appealing to handle everything on the script side.

Edit: offtopic... @MistRonin: hah. i knew i saw it somewhere :)

i love this idea. i think this doesn't even require extra engine features. i'm almost sure someone (modder) already tried this an it worked.

that's a really bad example though since it uses a bunch of different assets from different sources etc. maybe even different material types used on different plants without any central thought behind it (no judgement here). what i was trying to say is that i think, with my low understanding of the technical details, by setting up the assets a certain way you can make dense areas perform better. another thing is more freedom for clever usage of lods due to the main canopy of trees hiding everything below from far away. also sharing data, instancing etc. and these techniques probably advance more and more with API updating. if it was just about bigger numbers we wouldn't need API but only faster and faster hardware. anyways. i think it's wise to reserve judgement for when we get hands on. there's enough to judge that we can already test ;)