Game Editor

[paul424]

Here you can post your ideas what do you expect to be in the game editor, so be nice monsters and "scrap yourself dimm boy !'

[Skorpio]

I'd like to be able to add and delete creatures, rooms and props in the editor (and also change their stats). An automatic update function for models would be helpful when a model or texture gets changed, so you don't have to restart the game every time. And it's difficult to change the map layout. It should be possible to click on empty space to add a tile and you should also be able to remove tiles from an existing map.

Also, we have to think about lights.

[Danimal]

A random map creation function, the ability to modify tiles, of adding creatures (specifing if they are neutral, hostile or belong to an AI keeper) and adding buildings and then being able to save everything into a ready to load map file.

[paul424]

It should be possible to click on empty space to add a tile and you should also be able to remove tiles from an existing map.

Looks like your current build from the ICE AGE.


Surely some of those are pretty easy to implement , but programming CEGUI is big pain for me , plus intorducing some top level scheme for new object family is a magic, as I am not the main OD designer and not genius of OOP .

Having current scheme in mind, where should funciton GenerateTileLayout(Tile_Container) , GenerateTileLayout( Rectangle coordiantes ) , . ... belong....
Where the container keeping the temporal GameObject not on the GameMap should go ?

[paul424]

For generting levels compositor pattern would be good :
http://www.oodesign.com/composite-pattern.html

In leafs there would be some simple gen methods like Depth-first search or some cellural automate, each compositor level would merge all below prev level , do some simple operation ( like Tile type permutation ) and provide simple interface to the level up .

http://en.wikipedia.org/wiki/Maze_generation_algorithm

[paul424]

This generates all kinds of 2d and higher dim mazes , but it uses dirty , low level C as well :
http://sourceforge.net/projects/daedalus/?source=recommended

[Skorpio]

A maze generator might be a good start, but we rather need a generator that can fill large areas with different tiles, create interesting shapes and maybe also adds creatures, buildings and traps. Then we can generate lots of maps and simply pick the best ones for a campaign or multiplayer games (I guess not every random map will be usable). BTW, we need a mirror function to create fair multiplayer maps (copy, paste and rotate would be useful, too). And a map preview window would be good, so that you can check out how a map actually looks before you load it.

[svenskmand]

For procedural generation of stuff I can recommend Perlin Noise

[Obsdark]

Something than i always wanted to see on a Map Editor (and than can Facilitate a lot the work) is make a "Mirror tool", i mean a tool who let you select a part of the map, copy it and invert it like in a mirror and place it where you want.

And add divitions lines who avoid you to easly copy easly, for example 1/4 of the map on the others 3/4 just coping and inverting.

That way the multiplayers maps can be maded much more balanced and become easier to make for the unskilled/casual mapmakers than any before, it would be very usefull for that kind of maps

[Skorpio]

Yeah, we need to be able to create linked duplicates of map parts in order to design fair multiplayer maps. Making a fair 8 player map by hand is probably rather tedious. How do they do that in other real time strategy games?

[domtron]

Has anyone thought about using a 3rd party application instead of making our own editor? It seems that Tiled is a pretty good FOSS editor. It's 2d but as of right now so is opendoungen maps. We could save a lot of man hours in the short term. However if we ever go with 3D dungeons we would need a different editor that actually supported 3D. I don't know howmuch work it would take to setup Tiled for OD map editing. We would need to make 2D images for tiles, creatures, and buildings which is not that hard since we already have the 3d models and textures to make isometric images. But this is another thing that new people to OD will have to install and configure(though some may already have a popular editor like tiled installed).

I think we need more 3rd party FOSS utility tools that are configurable to specific projects(like blender allows export scripts for different file extensions) so FOSS(and maybe even AAA and indi moder's) projects can save time and have access to powerful tools from the start.

Anyway that is my opinion, any comments?

[Bodsda]

Has anyone thought about using a 3rd party application instead of making our own editor? It seems that Tiled is a pretty good FOSS editor. It's 2d but as of right now so is opendoungen maps. We could save a lot of man hours in the short term. However if we ever go with 3D dungeons we would need a different editor that actually supported 3D. I don't know howmuch work it would take to setup Tiled for OD map editing. We would need to make 2D images for tiles, creatures, and buildings which is not that hard since we already have the 3d models and textures to make isometric images. But this is another thing that new people to OD will have to install and configure(though some may already have a popular editor like tiled installed).

I think we need more 3rd party FOSS utility tools that are configurable to specific projects(like blender allows export scripts for different file extensions) so FOSS(and maybe even AAA and indi moder's) projects can save time and have access to powerful tools from the start.

Anyway that is my opinion, any comments?

Could blender even be a potential design tool for 3d maps? I know tuxcart uses it for their tracks.

[domtron]

My post actually stems from an article I read a while ago about the same topic and it used tuxcart/blender as an example. I haven't actually played tuxcart but I think it has continuous solid modeled track which is easy to do in blender. I very much doubt we could use blender for tiled maps. Not positive though since I'm just somewhere between a high novice and low intermediate user of blender(ignoring the fact that my actual art skills are a bit... poor). Maybe one of our artists that use blender could clear up that point?

[Danimal]

the ideal map editor shouldnt be that complex to find IMO, as far as i know all required level info is stored int an xml file, something that exports in that format, even if its 2d driven should work?
Also someone alredy implemented skorpio tileset on a map editor, maybe its worth a look as well?

Maybe im totally wrong too.

[domtron]

Actually our maps are stored in a custom text file. snip from test.level:

{l Code}: {l Select All Code}

...
400 #MapSizeX
400 #MapSizeY
Media/levels/Test.level # The level to load after this level is complete.


# Seats
5  # The number of seats to load.
# color   faction   startingX   startingY   colorR   colorG   colorB
1   Undead   -5   10   0   0.4   0
2   Heroes   15   0   0.4   0   0
3   Villians   15   0   0   0   0.4
4   Constructs   15   0   0.4   0.4   0.1
5   Animals   15   0   0.4   0.1   0.4

# Goals
5  # The number of goals to load.
# goalName   arguments
KillAllEnemies   NULL
ProtectCreature   King
- 1
ClaimNTiles   1000
ClaimNTiles   200
MineNGold   20
+ 1
MineNGold   5000
+ 1
MineNGold   10000
ProtectDungeonTemple   NULL

# Tiles
10560  # The number of tiles to load.
# posX   posY   type   fullness
-54   -98   3   100
-54   -97   3   100
-54   -96   3   100
...


edit: In fact I don't think we have any xml files in the project other then whats the xml branch which died when the person working on it left...

[Skorpio]

What's the problem with the current editor? Can't it just be improved? I guess creating a usable editor in Blender would be as difficult as improving the old one. Mapping in Supertuxkart seems to be more similar to traditional mapping in FPS games, but for OD we need a different approach. Instead of modeling a map, we have to assemble it, and that seems to be easier to do directly in OD, since the game and editor both have to be tile based. Also, you can quickly test a map with the in-game editor and don't have to export everything.

Something that could be streamlined in the current editor is the export-import process from Blender to OD. At the moment the only way to add a creature is to add it to the level file.

[domtron]

The problem with the current editor is that it doesn't have the most rudimentary map making tools available. Correct me if I'm wrong, but all it really is able to do is change tiles between several hard coded types and fullnesses(how the tile connects to other tiles around it) and place one tile or different sized(and shaped?) groups of tiles. It can't save edited maps or make new maps with a defined x,y with and it can't do anything with creatures, buildings, or various meta data(like starting locations, lighting, ect) It is going to require a lot of work to get to a usable state which will take away from actually building OD. [Just to be clear I'm not trying to be mean or anything I'm just stating my understanding of the current situation.]

As I said I didn't think blender was a viable tile editor which you have more or less confirmed. I'm just saying it may be much easier to make an export script and a set of isometric tiles from our existing models and is worth looking into. Tiled is the most popular and seemingly well developed (I haven't had a chance to use it) 3rd party map editor I've seen. If anyone has better candidates(more configurable, 3D support) please speak up.

I'm not even saying that we do this permanently. If at some point down the road we need a more specific map editor we could opt to make our own and pick up where we left off. We could consolidate the current map editor code (which we probably should do anyway) and continue it if/when it is needed.

Also most games have map editors that are separate binaries from the actual game(at least in my experience). I'm not sure why this is but it seems to be a trend with only a few exceptions (age of empires, widelands[if I remember right]).

However the point is moot if there doesn't exist any viable options other then making our own. I'm not all that familiar with map making, dungeon keeper maps, or future plans for OD maps so I'm probably not the best person to look at alternatives but I can probably make the necessary export script if necessary(especially if it is in python ).

[Danimal]

What Domtron said +1, we need a funcional level editor

[paul424]

Domtron , so try to introduce the simplest change in the current editor, implementing the tiniest feature is hundred times better than endlessly discuss big-picture, or you can study the workflow of Editor ( what happens ) and ask about the details you don't understand or you would like to correct ...

[domtron]

Paul I have been studying tile.*(I thought it was a good place to start) and my first commit(which I finally got to push, yay ) was kinda sorta related to the editor. Anyway the purpose of this topic is to discuss the editor and I just wanted to make a case for a possible course of action. Anyway I'm done now unless anyone wants to poke holes in my idea which is most welcome, helps me learn and the like.

EDIT: And I understand your desire for more action less talk, but we need to do both especially if talk will allow a better use of our action.

[Bodsda]

DK2 had a very simple map editor. It was just a grid with flat icons covering the squares, it was very quick to use and scaled really nicely although it did lack some of the more advanced features like mirroring but did iirc include auto map generation.

The current level file design was created by AndrewBuck and I'm unsure if anyone else got up to speed with it. The problem with using a ready made map editor will be weather we can make it load/save our file format correctly.

@domtron - do you want to take this one on and decide if it is worth using Tiled. I think our options will probably be - 1) Use a FOSS editor and change our level loading code to handle the new format. 2) Make our own fully functional editor, no change required to the main codebase. 3) Fork a FOSS editor and change/enhance it to accept our file format

[Skorpio]

It can't save edited maps or make new maps with a defined x,y with and it can't do anything with creatures, buildings, or various meta data(like starting locations, lighting, ect) It is going to require a lot of work to get to a usable state which will take away from actually building OD.

The console commands to make new maps and save are, newmap x y and save filename.

Are there any editors that do what we want already or would they have to be modified and improved as well? Is it more work to modify another editor or to expand the in-game one? Personally, I'd like to have a 3D editor rather than something like Tiled, because I often use it to check if models, lighting and shaders look correct in-game.

[oln]

Saving maps is broken due to some changes in the map code from what i remember, (unless it has been fixed again).

[domtron]

The console commands to make new maps and save are, newmap x y and save filename.

Didn't realize that thank you. They both seem to be broken though.

Are there any editors that do what we want already or would they have to be modified and improved as well? Is it more work to modify another editor or to expand the in-game one?

As I said I don't know of that many map editors and none of them are 3D.

I would say it would definitely be more work to modify/improve another editor(though it would possibly help the foss community as a whole) then finish up our own. However if it just needs an export script written for it then it would be much less. I looked into Tiled and it supports export plugins though thetre doesn't seem to be much documentation.

Personally, I'd like to have a 3D editor rather than something like Tiled, because I often use it to check if models, lighting and shaders look correct in-game.

Ok that is the feed back I was really looking for, If our artists/content creators would mind trading a 3D environment for better mapping tools/faster short term development. So that answers my question unless anyone else wants to add to it.

[Skorpio]

Actually what's more important for me is the direct impression of a 3D view. I think if you just see a 2D map it's more difficult to get a feeling for the dimensions and problematic layouts. The overview might be better in a 2D editor, though. However, I don't think our maps will be too complex, so it probably won't matter much whether the editor is 2D or 3D.