UFO:Alien Invasion
Development => Coding => Topic started by: Flying Steel on June 20, 2010, 01:01:53 am
-
Is it only that the pathfinding for 2x2 units is not working or would a lot more coding need to be done in other areas as well to make UGVs an operational feature?
I ask because, if that is the only or primary hold up for UGVs, then I'm curious if 3D content for 1x1 combat UGVs could fill the gap while waiting on the pathfinding code for the larger UGVs?
It seems UCGVs such as this:
http://en.wikipedia.org/wiki/File:SWORDS.jpg
or slightly taller designs could easily fit inside a 1x1 square and thus circumvent the pathfinding issue.
P.S. If content for such units would be desirable, lemme know.
-
Actually, I just wrote up a preliminary TODO list on Duke's talk page (http://ufoai.ninex.info/wiki/index.php/User_talk:Duke), since he basically asked the same question. This is something to be further explored and expanded.
-
skirmish minimum
* automatic assignment of one UGV as the 8th soldier
* some default equipment for the UGV
* NO turning turret, NO correct origin for firing
. . .
full support
* 2x2 spawnpoints on all maps
Expanding on what I said earlier and taking this into account, it is totally realistic and graphically doable to have some combat UGVs that are:
1) 1x1 units, just like soldiers, and utilizing the same spawns.
2) Have no turret (they turn to train their aim left or right).
3) And position weapons the same way as soldier models, supporting both upright and crouched weapon origins, if desirable.
The unit would look like those real life, compact, weapon wielding UGVs I linked a picture of above, plus a folding (or telescoping) arm coming out the top of it, wielding the weapon system (allowing it to shoot from soldier height or lower, depending on what's called for by the current situation).
I don't mean a replacement for the ares and phoenix, those are heavier, though probably less versatile (due to their size) designs that will greatly enhance depth and balance when the engine supports them. But this is a concept for a complement to them and the human soldiers, that we will see, without a doubt, in the near future of real life.
And it sounds like it is something that the engine will support sooner, probably a lot sooner.
-
So according to the TODO (http://ufoai.ninex.info/wiki/index.php/TODO/2.5) it looks like the UGV implementation has been pushed back again (to v2.5). Has progress been made on this feature? And any progress on hovering units?
It would be quite nice to see more variety in the kinds of units you both field and face on the battlefield in the next version. As interesting as battlescape combat is in the early game, after so many missions the campaign begins to feel more like endless room clearing. Combined arms forces on both sides of the fight would do a lot to diversify the experience.
If fully implementing this feature as planned requires too many engine additions to be ready for the next release, I'd like to suggest again that an intermediate solution like the one described above could be desirable.
-
You're a modeller, help us make more aliens! Have you seen the alien bestiary (http://ufoai.ninex.info/wiki/index.php/Proposals/Alien_Bestiary)?
Any 1x1 aliens we have modeled and animated can go in right away. 2x2 pathfinding is still not there, and I think that's the big thing preventing UGVs.
No progress on hovering units, either. I think it is a really difficult problem to solve, and our battlescape doesn't get a lot of attention.
-
AFAIR, the last time I tested, UGVs were already working on the battlefield.
They were spawned on the battlescape, you could even move them, but not fire with them.
I guess technically there is not much missing except for spawnpoints on the various maps and some code to make the guns (which are seperate models mounted on top of UGVs) work.
The code to make the guns work should not really be complicated, but if the guns should animate independent from the UGV (turning a gun will not turn the UGV), which would ofc. look better and more professional, more work regarding code and UI would be necessary.
-
@Flying Steel: We need you ! We currently have no real modeller on board. Join the project ! ;D
-
If UFOAI doesn't mind using open content also used in one or more other projects, then I might be able to supply you with more alien models/textures/animations in the future.
No progress on hovering units, either. I think it is a really difficult problem to solve, and our battlescape doesn't get a lot of attention.
That's too bad. The battlescape is where you spend most of your time and apply most of your attention and skill.
AFAIR, the last time I tested, UGVs were already working on the battlefield.
They were spawned on the battlescape, you could even move them, but not fire with them.
I guess technically there is not much missing except for spawnpoints on the various maps and some code to make the guns (which are seperate models mounted on top of UGVs) work.
Hmm, if they can move I wonder if that means the pathfinding problem has been solved. Unless 'pathfinding' refers to an AI issue with planning rather than executing 2x2 unit moves.
Fully functioning gun turrets aren't immediately essential IMO, many tanks and tankettes from the world wars had swivel guns that could only be trained in a narrow forward facing arc. It seems like the 2x2 spawn points wouldn't take long to place either.
-
If UFOAI doesn't mind using open content also used in one or more other projects, then I might be able to supply you with more alien models/textures/animations in the future.
All UFOAI content is required to be open content. That's our philosophy and I don't think we're too proud to use working aliens similar to other games, though they may be the first to be replaced in that mythical utopian future when we have all the artwork we need. :)
That's too bad. The battlescape is where you spend most of your time and apply most of your attention and skill.
Agreed!
-
I'd only release content under a permissive license if I felt it was 'original' in the legal sense. And by other projects I meant open projects, including at least one of my own but also any other games that felt like using the content that I created and released (probably using CC0).
It's like if Vega Strike or Pioneer (http://pioneerspacesim.net/) started using your open licensed interceptor and UFO models as content for its own spacecraft or vice versa. There could be no legal issues with that but you might still choose to try to avoid using content used by other open games purely on artistic grounds or simply not care. Which is why I mentioned this; if you don't mind sharing content I might be able to supply you with original alien artwork that I will probably need for my own open game projects.
Tangentially, you'd then also want to keep an eye on what content licensing TremZ (a fork of Tremulous) ends up using, as they might be a gold mine for alien artwork.
-
Our current Stiletto interceptor came from and has been given back to (http://opengameart.org/content/future-fighterjet-v2) the open source community at opengameart.org. It'd be great if we could have all original content. But, as your post about the lack of new aliens indicated, it is more important for us to have the content itself. If you're making aliens and able to animate them as we need, it could be very useful! :)
If you're looking for material you might be able to use for your own projects, you can always check our data source. I know of at least one alien there that is modeled but not animated (http://ufoai.ninex.info/forum/index.php/topic,4849.msg39041.html#msg39041).
-
Attempting a simple approximation of UGVs for base defence. Emphasis on gameplay rather than art or models.
I'm operating on the assumption that some or all of this makeshift stuff will be tossed out as UGV work continues. That's fine. I just figure it won't hurt to toss in some bot-like things now :)
For this scenario simply say that ubiquitous alien EW rendered presumed future battlefield 'bots unreliable in the field, but heavily modified bots in direct contact with the base can still work even during alien incursions into that base. As discussed upthread these Base UGVs or BUGVs would be simplified versions of the full UGV concept... the idea being that the simpler the systems on these prototypes the less the alien EW can fritz things up.
(and if any of this concept survives in the game perhaps they could be the base from which battlefield-capable UGVs are redeveloped after combat experience and some alien tech is researched)
So no big rollout of story or art. These experimental bots are simple gun platforms with cameras but without turrets. When a base containing BUGVs is attacked your Phalanx troopers don't magically beam in from a dropship halfway around the world... instead the bug-v's roll out of the building labeled "Base Defense UGV Bay" and into action.
building building_botbay
{
name "_Base Defense UGV Bay"
image base/botbay
fixcosts ?0000
build_time 5
varcosts 1000
map_name "quarters"
pedia rs_building_botbay
type botbay
capacity 8
starting_bugvs: 4
}
The BUGV advantages are real but should be matched by their disadvantages.
Advantages:
are rigged to accept any human handheld weapon. this is more of a desperation measure at first but will turn out to be very handy when alien weapons are acquired and researched.
are dual-wielding - the default configuration holds a machine gun and a flamer and can use either at will... but cannot reload either of them.
are armored better than humans
(this is the reality of future robot combat... robots are strong... and that's why the player only starts out with 4 of them)
built in IR scan (still costs TUs but not as much as humans)
are fast on the straightaway but they slow way down on cornering, are even slower on ramps and are like molasses on stairs... or maybe stair-stuck? (dalek syndrome)
copied living quarters as the bugv hq and presume that 4 units are unpacked when the building is finished. As with current living quarters the "building finished" flag signals that the base is now available for attack.
When an attack happens the player gets a chance to choose bot weapons from available stock and the fray begins. Afterwards the bots either do a victory dance and return to the botbay for reloading and repairs... or the bots are scrap and so is the base.
obviously the option of adding human guards to the mix hangs over this but i'm going for all-bots at first just for the fun of it. It can all be justified later one way or the other if someone wants to take the time to do so :)
attached find simple prototypical bot idea... this concept isn't something to spend 70,000 hours baking normalmaps on...
-
the inverse crouch... slows the bot down and renders it a bit less accurate but enables it to shoot over tables and such.
-
here's some gun clarification:
first is the default machine gun and flamer. these are supposed to be the same weapons the troops use. me no artist so sue me :)
second is the bot sans guns. the idea is that the weapon grips go into the green boxes.
third has the flamer replaced with the grenade launcher. a bit hazardous, no reaction fire, no close-in use and the bot still can't reload anything... but it is a grenade launcher.
yes, that's the grenade launcher... just turn your head sideways and close your eyes... see? my world and welcome to it :)
-
apparently no objections yet :) ... onwards...
Some background and support stuff comes up first....
guns and magazine capacity: is there any objection to having magazine capacities defined by the magazine instead of the weapon? cf:
http://ufoai.ninex.info/forum/index.php/topic,4509.msg34597.html#msg34597
Should not affect balance much. If someone wants to load their assault rifle with a custom 100 round dual-drum magazine then just let them take the TU and accuracy penalties for it.
I'd make all the needed ammo transitions and test it of course but I wanted to check on the concept's general viability before trying to code it.
-
please also check out this wiki page:
http://ufoai.ninex.info/wiki/index.php/Talk:UGVs
-
please also check out this wiki page:
http://ufoai.ninex.info/wiki/index.php/Talk:UGVs
Yes, I've been going over the UGV code available. It seems that by coding for these "protobots" along the guidelines for the fullsize UGV's it would literally bridge the gaps between organic actors and mech units... even if these protobots don't get in or stay in the game.
Also, the setups I'm thinking over would have some applications to the cyborgs even if they can't match the capacities of a fully mechanized unit... with unit capabilities scaling up from human to cyborg to bot and then to UGV.
It'll be still be a jump going from the 1x1 bots to the 2x2 full-sized UGV's envisioned for the game but I can try to do a lot of the groundwork for the big units while working out the small ones.
I'll start putting it together and will check in with ideas and questions... that is, unless you think I'm starting off in the wrong direction entirely :)
... keeping the UGVs as a class of employees should work... minimum wage... mandatory time off... paid maternity leave...
-
I wonder if there is any progress in this? I would really like to see this in game!
-
Alright so now it looks like UGVs are no longer on the 2.5 TODO list either. So they've been pushed back to a later version for three or four major versions in a row at this point. The content for them is there, so what is the hold up?
I believe the hold up is demanding too much at once from development of UGV support. Like the ideas that UGVs must be 2x2 with the associated path-finding issues solved, must have separately rotating turrets, must have separate transport space or separate spawn locations on maps, etc. Things like this are not necessary and can be added later as desired.
All that is needed right now is a minimum feature set for UGVs to work in game, in skirmish and campaign, and then to have this minimum of features added to the 2.5 road map and then implemented before that release. Someone please correct me if there are more but I believe we only need two things:
1. The ability to buy and/or manufacture a special "soldier" that has no skills and a very simplified healing mechanic:
2. A special tag for weapons and equipment that distinguishes between infantry and UGV outfits.
-
1 and 2 are actually done, or nearly done, I think. In the development version you can purchase UGVs and UGV weapons. We've also got a basic 1x1 scout ugv modeled and animated, so 2x2 pathfinding isn't necessarily holding it up. Here are some of the missing components (off the top of my head -- probably lots more behind the scenes work)
3. UI changes to allow the player to select UGVs on the battlescape.
4. Changes to aircraft definitions to specify how many UGVs can be carried in a dropship (the alternative, using UGVs like a "soldier" only adds more difficult work restructing the concept of a squad in the code).
5. Changes to aircraft equipping menu to assign UGVs to aircraft.
6. Special equipping UI to select and equip UGVs (even at the bare minimum, you need a system to prevent non-UGV weapons from being equipped on a UGV -- but really you need a separate interface, because they won't have right hand, left hand, holster, belt and backpack spaces)
UGVs keep getting pushed back because they are a really big feature, because the active devs have other priorities and because it's a feature that requires different devs with different capabilities to coordinate on a single project.
Personally, UGVs are WAY down my priority list. I struggle to really understand what kind of benefit they would bring other than the "cool" factor. I'm not dismissing the "cool" factor, but I think there are plenty of other major issues to be dealt with first (campaign balance, better maps, plugging holes in weaponry/tech trees, reaching a temporary conclusion point in the story, improving game mechanics of battlescape, fixing air combat, improving the UI). That's where I am driven to spend my time. Other devs disagree and want the UGVs implemented quickly, but they have other priorities too.
-
1 and 2 are actually done, or nearly done, I think.
I wouldn't even say "nearly done". When I rebuilt the market UI logic I tried to make it UGV compatible but failed. The UGV datastructure is not well designed, it (and how the code handles it) have conflicted parts. First we should sit down and decide how UGV should work including if the UGV should have replacable turret and which tag should be used for it and other things also.
-geever
-
First we should sit down and decide how UGV should work including if the UGV should have replaceable turret and which tag should be used for it and other things also.
Well this is just as good of a place as any to sit down and flesh this stuff out.
Would it help to put together a preliminary specification and then pass it around to everyone so they can propose changes?
(the alternative, using UGVs like a "soldier" only adds more difficult work restructuring the concept of a squad in the code).
Why is that? Because of the eventual unusual space requirement of 2x2 units? For most of the issues you bring up it seems like piggy-backing the UGVs off the existing functionality for soldiers would be the way to start.
6. Special equipping UI to select and equip UGVs (even at the bare minimum, you need a system to prevent non-UGV weapons from being equipped on a UGV -- but really you need a separate interface, because they won't have right hand, left hand, holster, belt and backpack spaces)
Good point.
I struggle to really understand what kind of benefit they would bring other than the "cool" factor.
What fighting machines offer as far as game play is a way to add more specialized and diverse units to the playable "earth" faction in the battlescape. The alien opponents can come in any shape and size and have most any special abilities and disadvantages, and it all makes sense. But with strictly human infantry you might have some variety with powered armor or jump jet suits, but can't get away with nearly as much.
Vehicle types things are also interesting in that they are generally good at doing just one or two things really well, whereas infantry can more easily shift roles but suffer more trying to do any particular one. For examples, a tank can dominate open ground, but does poorly in tight cluttered areas and can't enter most structures at all. A small scout vehicle on the other hand might be the best at exploring or hiding within a building but is too slow and has too poor of standoff fighting ability to survive long outdoors.
-
Vehicle types things are also interesting in that they are generally good at doing just one or two things really well, whereas infantry can more easily shift roles but suffer more trying to do any particular one. For examples, a tank can dominate open ground, but does poorly in tight cluttered areas and can't enter most structures at all. A small scout vehicle on the other hand might be the best at exploring or hiding within a building but is too slow and has too poor of standoff fighting ability to survive long outdoors.
How would you suggest tanks dominate open ground? How do you envision accomplishing this? By giving them excellent long-range accuracy? How do we do this without imbalancing the battlescape (alien AI sucks at long-range encounters -- can't find cover, fires away at unrealistic distances)? How do we make the tank perform poorly in tightly cluttered areas? I can't think of a game mechanic that can be used to achieve this (2x2 may keep them out of the cluttered areas, but it doesn't effect their performance), but there may be one I'm missing.
More issues to be worked out for the scout vehicle. A slow scout vehicle would be little more than a kamikaze machine for checking indoor areas. It's not adding a very interesting game mechanic and could really slow down and narrow battlescape play. I see an incentive for players to spend their time moving the scout UGV around while the soldiers just cluster behind it.
-
Personally, in the original X-Com I've used early (i.e. non-hovering) tanks as a kamikaze scout machines. Of course, rocket tank packs quite a punch, but in about 20% of missions it just was killed by some plasma blaster alien waiting for my troops to disembark, with my dropship in its sights. Before my tank could even have a chance of firing a shot.
-
How would you suggest tanks dominate open ground? How do you envision accomplishing this? By giving them excellent long-range accuracy?
Yes that, combined with better armor that allows them to survive sometime in the open.
How do we make the tank perform poorly in tightly cluttered areas? I can't think of a game mechanic that can be used to achieve this (2x2 may keep them out of the cluttered areas, but it doesn't effect their performance), but there may be one I'm missing.
There are many possible ways. The 2x2 is of course one of them. Or having a special map tile tag for "rough terrain" that doesn't affect infantry but slows down "vehicles". Those are examples of hindering mobility. Another approach is to make them vulnerable to cluttered environments where they can be ambushed from many directions. Directional armor is used by many games for this purpose, which makes the vehicle more easily penetrated by attacks not against it's frontal armor. Requiring additional TUs for turning the vehicle or turret relative to infantry further limits it when enemies can come from many angles.
But those would need to be down the road implementations, I highly recommend not hinging UGV implementation on the development of further features like those above. That's been a recipe for indefinite postponement.
How do we do this without imbalancing the battlescape (alien AI sucks at long-range encounters -- can't find cover, fires away at unrealistic distances)?
That's a tangential problem. It is countered by giving human technology used at the same stage of the game, poorer stats overall.
A slow scout vehicle would be little more than a kamikaze machine for checking indoor areas. It's not adding a very interesting game mechanic and could really slow down and narrow battlescape play.
If it is slow but armed and with fairly heavy survivability, that is one option. It would be like a modern bomb disposal robot in design, but useful for making the initial break-through into a concentrated alien ambush. Another option is a fast vehicle, lightly armed or armored, that truly is a low cost scout. Or a true kamikaze weapon that runs in and blows itself up / launches grenades in every direction.
It is my humble opinion that a unit like one of these suggestions could actually speed up play by making the "clear the crashed harvester of camping aliens" type missions much less tedious.
I see an incentive for players to spend their time moving the scout UGV around while the soldiers just cluster behind it.
Only when you know all the aliens are hiding in that one room, in front of you. Otherwise you might deploy the scout behind you to make sure you don't get ambushed from behind. Or deploy it near the aliens' backdoor in case they try to make a run for it.
-
I think if we create a UGV with better armor and excellent long-range accuracy, then nerf the human tech weapons, we'll just end up turning the UGV into a super-soldier, making other soldiers less relevant. This is the balancing problem I was asking about.
Your ideas around the scout UGV are interesting -- particularly the remotely piloted bomb. We'd have to get the economics right on that one so it wasn't too easy to just buy 1 or 2 UGVs to blow up each mission. But it could be an interesting piece of equipment to deploy in special circumstances. The main difficulty will be figuring out how to keep it from rendering Close specialists obsolete.
My main consideration when I consider "balancing" issues is: how do I add to the arsenal of options without making any existing options obsolete? Close specialists are a difficult breed to keep useful, because they necessarily run higher risks and have more narrow utility. The return on investment needs to be higher than, say, an explosives expert who can do alright in the open or indoors or even indirectly.
-
... The main difficulty will be figuring out how to keep it from rendering Close specialists obsolete...
Once the humans introduce the UGVs, the aliens will develop some ECM-device to counter this, no?
-
I think if we create a UGV with better armor and excellent long-range accuracy, then nerf the human tech weapons, we'll just end up turning the UGV into a super-soldier, making other soldiers less relevant. This is the balancing problem I was asking about.
Sorry I wasn't very clear. What I mean is in general the rule is, you have to make the player's options less effective to make up for the AI's dumbness. So if you add a human heavy UGV "tank", then you have to give it 75% or the armor and 75% of the firepower it would have if the AI was competent or this was a multiplayer only game. Or give the aliens more powerful weapons, like an early game tank and anti-armor weapon of their own.
Then later on, when the AI becomes more smart and modern, using heat maps and such, you must rebalance the arsenal's stats more in favor of earth weaponry. But whatever you do, UGVs don't make any special difference, you just have to balance them up or down for the AI, like you would any infantry weapon or equipment.
Your ideas around the scout UGV are interesting -- particularly the remotely piloted bomb. We'd have to get the economics right on that one so it wasn't too easy to just buy 1 or 2 UGVs to blow up each mission. But it could be an interesting piece of equipment to deploy in special circumstances. The main difficulty will be figuring out how to keep it from rendering Close specialists obsolete.
Because it takes up an entire soldier slot, can only be used once and has no ranged ability whatsoever (versus CQC specialists who can carry a laser sidearm for example).
My main consideration when I consider "balancing" issues is: how do I add to the arsenal of options without making any existing options obsolete? Close specialists are a difficult breed to keep useful, because they necessarily run higher risks and have more narrow utility. The return on investment needs to be higher than, say, an explosives expert who can do alright in the open or indoors or even indirectly.
Well that's where a "breacher" UGV comes in. It goes on point during room clearing and takes the brunt of the alien ambush, possibly sacrificing it's nonliving self, and then the CQC specialist(s) follow immediately behind and do a lot of lower risk killing. That would help one big problem with using CQC specialists, that they don't survive well.
The other problems they have are not directly related to UGVs. Probably hand grenades and grenade launchers are too effective, and thus effectively replace CQC weapons. Or that CQC weapons cost too much TU and do too little damage to make up for such limited range. Or that there aren't any good long range side arms besides the laser pistol. Having UGVs can't help or hurt any of these potential imbalances.
In general terms UGVs are semi-expendable super-specialists. When they die you only lose money, not experience. And while they generally can have a better something than a human soldier can carry, they are also bound to it. A CQC or fire support guy can have a rifle, a sidearm, a knife, several hand grenades, and can pick up other weapons/ammo from fallen comrades and aliens, all in addition to his staple weapon. He can quickly shift from one role to another on the battlefield. But UGVs are locked into using one weapon, maybe two. They are committed to doing one thing especially well versus a non-veteran human specialist. And so they make you think more carefully about their deployment, both before you launch your dropship and while in the battlescape.
-
First we should sit down and decide how UGV should work including if the UGV should have replacable turret and which tag should be used for it and other things also.
So why not do that? :)
Let me start:
1) Separate turret. More realistic and will not cause players to beleive that UGVs are just soldiers in disguise. "tag_turret". Implement the submodel rotation for turrets (rather easy).
2) Let gun be a separate model from turret. "tag_gun" in the turret model.
3) Anyone to continue the list?
-
Questions:
Do we buy UGV with a turret or turrets should be bought separately and installed in equip UI? (preferred separately)
Maybe we buy them with default turret which can be replaced but not removed? (more difficult to code)
-geever
-
My proposal: separately. For example, start with a machinegun (explaining that it was a quick hack to field vehicle as fast as possible in the writeup), then, with some research, add grenade launcher or rocket launcher or even 30mm autocannon (okay, last one was just because I like 2A42 with it's selectable feed feature).
-
If UGVs have separate rotating turrets, how will the player turn the turret without turning the body?
I'm fine with having UGVs and turrets as separate items. But I'd like the UI to have you buy a pre-configured UGV. Otherwise it will get unnecessarilycomplex figuring out which turrets fit which UGVs. It would be easy to define a UGV as a base and a turret, and still re-use the turret models. Battle Tank (Autocannon); Battle Tank (Rocket); Battle Tank (Grenade Launcher). Unless we're going to have more than a dozen possible UGV configurations, I don't think it's worth making it too complicated for the user.
-
If UGVs have separate rotating turrets, how will the player turn the turret without turning the body?
Add something like Shift-RClick or Alt-RClick?
Dozen of combination you can get very easy. 3 UGV types (scout, medium and heavy) x 6 weapon types (machinegun, plasma, laser, cannon, grenade launcher, particle beam). Add here next-gen UGVs...
So, to make life more easy for user may be it's worth to add filters to market instead? Something like "show only compatible weapons/parts/equipment for..." It can be useful for aircraft too.
Making only predefined configurations make upgrades more expensive or non-obvious and force to buy unnecessary things.
-
If UGVs have separate rotating turrets, how will the player turn the turret without turning the body?
Easiest option: skip it, as the original did. Turn turret only when firing, and that's all.
I'm fine with having UGVs and turrets as separate items. But I'd like the UI to have you buy a pre-configured UGV. Otherwise it will get unnecessarilycomplex figuring out which turrets fit which UGVs. It would be easy to define a UGV as a base and a turret, and still re-use the turret models. Battle Tank (Autocannon); Battle Tank (Rocket); Battle Tank (Grenade Launcher). Unless we're going to have more than a dozen possible UGV configurations, I don't think it's worth making it too complicated for the user.
Possibly, we can make some sort of proxy item at the market which will, when bought, unpack itself into base+turret combo? geever, what is your opinion on that?