Spring RTS Engine

Wow, we got a lot of stickies in here.

I'm not entirely certain if this should go here, could be under feature requests, could be under mapping, could be under lua, could be under development... makes most sense to me here.

I've seen this idea presented in different ways in different mod threads and I think it would be desirable to a number of current mods and mods in the works from what I hear.

The basic idea I present here could be expanded in a number of ways that I'm sure many of you will recognize and what I'm looking for right now if discussion. The implementation of this idea would require the co-operation of many disparate parties.

Also note I'm kinda out of it right now cause I'm sick

The Basic Idea:
LUA Scripts included in maps that present methods and set variables that are referenced by a mod's LUA.

"Points of Interest" is a term I think I've stolen from other RTS campaign editors... basically what I'm talking about is those glowing runes in warcraft II, galactic gates and rocks in TA, gas vents in starcraft, the fruit spot in pacman, the supply depots in C&C, the factories in advance wars, the resource beacons in Dawn or War, the artifacts in Age of empires etc... basically stuff on the map that players interact with in some way that didn't start in their posession.

In most RTS games these types of things are added to the map at the time of creation in the map editor and we currently have the ability to place trees, wrecks and geos into maps at creation time, and the ability to add pretty much anything else during the start script... the problem is that we have a lot of different kinds of mods to support with these maps and for a mapper this means ensuring that the map works with different mods... which limits both the mod designer (to take in to account the features that other mods use) and the mapper to include any special things that a mod needs... okay, brain derailed there.

Anyways, what I am thinking is that we have a LUA script that defines a number of "points of interest"... some kind of array of positions and types, and then modders and use these locations to place items on the map. so for example.

You have a Fantasy mod and it needs lots of mana crystals scattered around the map and a few gold mines at strategic points and then some Fortress in a locations away from the start points that generates the generate neutral mobs.

So the hard and stupid way to do this would be to make a LUA script in the mods that generates mana randomly anywhere that the metal density is about 1%, replace all the geos with gold mines, and generate a fortress randomly on a flat surface thats far enough from the players start positions that it won't interfere with them.

What I'm suggesting is that the map designer included a list of X, Co-ordinates of a number of different type... then the positions of the mana could be used to place tiberium in a C&C mod, the positions of the gold mines could be used for DoW style resource points... now the fortress positions might not ... man... I'm really goiing out of it... could someone explain what I'm thinking here... I know it makes sense.

geos

yeah, but what if a mod needs like 50 of them or 4 different kinds of buildings (maybe it's like Advance Wars and they need a command building generated at the start pos and then cities, factorys, airports and seaports to claim around the map... you could define where these things go on your map as PoI8-PoI12 and then when somone wants a similar distribution of resources for their map they could use them, or just PoI12 (which is going to be a strategic point on the seashore) to place some fish for a Age of Empires Style Mod, which would also use PoI27 which is strategic resourse points in the water where fishing ships could sail to... which in a WarCraft II style mod would be used as Oil deposites.

The problem with that is that noone knows what or how many PoIs should be there. Imagine your mod uses PoI 1 as a resource site and only allows harvesting from those sites. A map author thinks PoI1 should be the king of the hill goal. Whoops. If you don't know what your map gets interpreted as you can't know where points of interest should go and which ids they should have.

I would either use geos or mex spots, use lua to table them and create my poi on those spots depending on how many I would want, I would use either geo or mex.

I think the problem is that there is only one "class" of Geo.

Here's my example: you have a mod where there is an "artifact" that needs to be near every potential base. EG the gold mines from Warcraft. You could use the metal-map and treemap to produce trees, but you're screwed for the goldmines.

Then things get ugly once we start getting into other things, like claimable neutral units. That possibly could be done by swapping out neutral features.

Either way, the problem is that the OTA map features don't correspond to other gametypes, and so modders are limited in using Spring/TA maps. There are no guarantees - most maps will have only a metal map (and even that's not reliable), and any other type of feature would have to be extrapolated from the metal map.

Either way, I think this is a lost cause - expecting mappers to place features that they don't even know how they will be used in game is futile. Mappers map for specific mods - better to design your mod such that it is either compatible with maps from TA (noting that you could do some hefty lua to infer things from the TA map features), or provide your own maps.

The problem happens because some maps MUST have one of X, Y or Z. A naval Warcraft map MUST have some oil deposits. A Starcraft map MUST have a vespene source for each spawn point. A Z map MUST have flags and buildings around them. And even mappers won't necessarily obey this.

Expecting them to puts a tremendous burden on mappers that they won't follow through with. Better approach would be to provide it as a script/engine resource. If you need something near each start point, then what you really want is to tell the engine (or Lua script) "Give me one location per spawn point that is X distance from each spawn point, reachable by path-class Y, and has room for template Z, for the first N spawn points. If not possible, spit a message to the screen to tell the user that this map is inappropriate for this mod".

Sure, it's flawed - but I doubt it would be any worse than the human approach.

Or get someone to make you mod specific maps. We've got those for KP.

Yeah, I was also worried that when you have someone "choose start position in game" it would break a carefully layed out stuffs...

How close are we to supporting multi-player scenario files(lua start scripts)... cause that would be a sufficient replacement.

So when someone releases a Mod that requires special buildings they can make a scenario file for that Mod for each map(or whichever maps they like), in theory this could be trivial to do and thanks to the magic of Lua it could even alter the map slightly to conform to gameplay requirements... And it doesn't always have to be all new, we could adapt a scenario from another mod to the target mod when we get a couple of them.

Odds are if I had to be sure my features were added to a map, I would just spawn them within a distance of my startpoints.

Isn't it possible to have a map LUA that spawns these things?

KDR_11k wrote:The problem with that is that noone knows what or how many PoIs should be there. Imagine your mod uses PoI 1 as a resource site and only allows harvesting from those sites. A map author thinks PoI1 should be the king of the hill goal. Whoops. If you don't know what your map gets interpreted as you can't know where points of interest should go and which ids they should have.

So rank them. A type 1 PoI is found evenly distributed around the map, a type 2 is evenly distributed except there are none in the mapper's intended starting locations, and type 3 is for the king of the hill kind of ultimate goal. Maybe an additional classification for the density of each PoI type distribution, say A for extremely sparse and E for dense, and allow PoIs to have more than one rank and classification. So TA geos might be considered type 2B or 1B, Fibre's doomsday thingamabob would be type 3A, TA metal patches might be 1C or 1D.

+1

I completely agree, this would be the best way to handle it, and mods should be able to change out the PoIs. Let's say if it's BA, or CA, the level 3 is a large metal/energy object. Level 2 is geos, and level 1 is mines.

Working towards PoI heavy mapping, I personally believe it's fruitless to attempt to maintain universal map compatibility, and it's also not beneficial.

It's kind of like trying to make pants that will fit everyone. Even if you succeed, it will be incredibly hard to do in comparison to making normal pants, and chances are for the vast majority of people it they will be uncomfortable. For some things a tailored fit is just by far ideal. Mapping for increasingly divergent RTS games is one of them.

I am tempted to say something petty about looking at the last WIP shot from PURE again a little more closely, but that'd be kicking the shins of the blind, just for the hell of it.

Let's just say that you guys are all going to get what you want. Soon. Maps specific to game designs, that are *also* maps that aren't specific to anything. Game designs that rebuild the maps, to suit their needs. Map code that gets along with game code. Game code that obliterates map code results, because it might break game design.

Everything in between. Stuff I haven't thought of yet. Mobile hordes. Creatures, invaders, animals, plants that want you for lunch. Sound FX, animated waves along a beach, you name it, if it can be coded, it will be possible. Best part is that aside from making killer content, it won't be hard to use.

It's ready, people, the code works. Now it's just content, testing, AIs being trained to be a little less retarded, and it's done.

Argh, if you think arbitrarily placed points of interest can compare to carefully placed points with a regard for the resulting gameplay and map dynamics you're fooling yourself. If I was using control points yet I wouldn't be very happy with them being arbitrarily placed, I'd want them in strategically interesting locations and properly balanced.

Mappers cannot map for all mods, modders cannot mod for all maps.