"Open RPG engine" -- any takers?

[Prime Junta]

I've been looking a bit at some OSS game engines like OGRE, Crystal Space, Reality Engine and such. It strikes me that most if not all of them appear to be (a) oriented towards twitch games, and (b) operate at a pretty low level of abstraction (e.g. OGRE isn't much more than a pure rendering engine). OTOH most RPG construction kits I've seen appear to be technologically pretty primitive when it comes to the bling factor (we're talking 2D sprites, square grid, and pretty basic sound).

Now, suppose someone were to start an OSS project with the objective of creating a game engine geared specifically towards CRPG's. The main design goal should be to make creation of game content as easy as possible, while providing as much bling as is feasible and handling the trickier issues of platform compatibility.

Suppose the "Open RPG engine" project would have the following goals and characteristics:

* Non-restrictive license (MIT, Apache, BSD or similar)
* Builds on an existing OSS game engine project
* Defines file structure for game modules and other assets
* Not tied to any individual system of game mechanics. Instead, hooks from the API to the game mechanics are abstracted. (Example: an inventory management function is part of the "core," the concept of "bulk" is defined for objects, and an operation to recompute a stat for a character is permitted when an object is added to or removed from inventory. However, what the stat actually is, is defined in the separate game mechanics module.) It should be flexible enough to accommodate anything from S.P.E.C.I.A.L to d20 and between (but not necessarily twitch-based systems).

* Includes:
** Toolset similar in scope to the NWN or TES construction sets
** Rendering, sound, FX
** UI functions such as:
*** Inventory management
*** Party/character control
*** ...etc.
** Library of models, animations, and objects

Basically, what I have in mind is something like the NWN toolkit, only (a) open-source, (b) not organically tied to the d20 system and (c) better.

Thoughts? In particular, is there a project like this already in existence and I just missed it in my quick look-see?

 

[WouldBeCreator]

Because I expect this would take several years to complete, and because my patience doesn't run that long, it doesn't particularly interest me. If such a product were out *now* it would be nice and probably would interest me.

I suspect, though, that as much as anything else you would need lots and lots of decent, free content (i.e., art, music, and sound) if anyone were ever to use it. The appeal of most RPG Makers seems to lie in the free tilesets, sprite sheets, and MIDIs and whatnot.

 

[Drakron]

There is something called RPG maker, they come out with the XP english version about the end of last year.

I know it looks a lot like FF back in SNES days but ...

 

[onerobot]

That is fairly ambitious, and finding existing projects that won't result in a HUGE fork will probably be problematic. There are several pure engines out there (you did a good job finding some of the good ones) but I can't think of anything that really comes close besides FIFE. It's GPL'ed and is cross platform as it uses OpenGl and (I assume) SDL. It seems that they're using the remnants of the IanOut engine (which has been largely reworked) and are turning it into more of an all-purpose editor. A redimentary UI editor seems to be in at this point, but it's currently all pre-alpha stuff.

I'll keep my eyes open for other possibilities, but this is probably about as good as it gets.

 

[Zomg]

Sure, I'd love to use something like that. There are a few one hour mini-RPGs for which I'd like a low-investment outlet to get out of my head, without being encumbered by another game's mechanics and brand penumbra.

 

[bryce777]

Well, as a sort of independent game developer, I personally am not interested. That is just because I am very interested int he programming part, though.

I am also more interested in making a super flexible, ultra moddable gaming system and not being a hopped up modder. Ogre is probably better int hat it's really low level. They also have a good license.

To me a full 3d engine is just stupid for a top down game. I really think that 3d graphics mostly look like garbage compared to 2d backgrounds and such. A system like in toee is the best way to go, I think.

Basically, what I have made are tools that allow you to mod not just individual items and quests, but tools that allow you to mod the gaming system itself. You can select for example how feats work and are gained, how leveling up works (or choose a usebased system and set usage levels and tie skills to actions and/or stats to determine how skills increase - so you can make a system like darklands or silent storm if silent storm didn't have quirks), feats themselves, how magic works, etc.

You can make nearly any sort of system this way. Any game system I have seen, anyhow. The last thing I am worrying about is graphics for now - first the basic modability and file formats that are not going to get fucked up every time I upgrade (like mount and blade), then the multimedia stuff ie spell effects and characters and adding them in for custom spells, then GUI ability to do any sort of action possible for tactical combat that is complex like toee or xcom (and making this modable), and then the gameplay, then the writing and plot, and then the graphics. After that, it's tying it all together into something coherent and looking for a delivery method and then beta testing then BAM instant fame and bitches! Well, instant if you count 25-30 years of development as instant.

 

[bryce777]

So, I suppose my thing is to try and capture the modding community. I will never be a great artist, but at some point I imagine I will bring someone into the fold for that, but not before I have good tools for them to use.

I also want to make the game as moddable in possible in other big ways. Such as, a variety of paradigms for a travel map, or the ability to not have one at all if you like.

There are so many easy things to do that are nice touches that game companies, especially small ones, don't do. I think on any project you need to look at how to solve the general problem and not just the specific one at hand, especially since the specific problem will always be a moving target until things get more finalized - many of the final decisions usually coming about just because of arbitrary decisions made early on that are too difficult to reverse later if you do them in a hardcoded fashion.

 

[Prime Junta]

To me a full 3d engine is just stupid for a top down game. I really think that 3d graphics mostly look like garbage compared to 2d backgrounds and such. A system like in toee is the best way to go, I think.

2D has a few major drawbacks, though: it scales poorly to different screen resolutions/sizes, and objects get occluded by other objects too easily. Also, it's tough to represent 3D objects like architecture in a 2D engine without running into the occlusion problem or having the scene look like your sprites are marching around inside a painting.

Basically, what I have made are tools that allow you to mod not just individual items and quests, but tools that allow you to mod the gaming system itself. You can select for example how feats work and are gained, how leveling up works (or choose a usebased system and set usage levels and tie skills to actions and/or stats to determine how skills increase - so you can make a system like darklands or silent storm if silent storm didn't have quirks), feats themselves, how magic works, etc.

This is exactly what I had in mind with the abstracted game mechanics. I.e., the engine would have hooks that you can link up to stats in the game mechanics, but would not in itself take a stand on what the game mechanics are.

 

[Prime Junta]

That is fairly ambitious, and finding existing projects that won't result in a HUGE fork will probably be problematic.

Make that "very ambitious."

Obviously I haven't looked at any of the engines in detail yet, but assuming they're basically libraries, I can't see why it'd be necessary to fork the project; what I had in mind was another set of libraries linked to and extending the "pure" game engine. IOW it shouldn't be too tough to keep the projects in sync, assuming the core engine API's are relatively stable.

There are several pure engines out there (you did a good job finding some of the good ones) but I can't think of anything that really comes close besides FIFE. It's GPL'ed and is cross platform as it uses OpenGl and (I assume) SDL. It seems that they're using the remnants of the IanOut engine (which has been largely reworked) and are turning it into more of an all-purpose editor. A redimentary UI editor seems to be in at this point, but it's currently all pre-alpha stuff.

Yup, I'll be following FIFE too. It has the advantage of being considerably less ambitious and therefore more likely to actually result in something workable. However, in my view it's hobbled by the GPL license: this means that it's extremely unlikely to get picked up by a commercial venture that could kick-start the engine development. Apache, BSD, or MIT would be much nicer, since it means the engine could be used and contributed to by a company making a commercial game. (In my experience, the most successful OSS projects are the ones that are most used commercially.)

 

[bryce777]

To me a full 3d engine is just stupid for a top down game. I really think that 3d graphics mostly look like garbage compared to 2d backgrounds and such. A system like in toee is the best way to go, I think.

2D has a few major drawbacks, though: it scales poorly to different screen resolutions/sizes, and objects get occluded by other objects too easily. Also, it's tough to represent 3D objects like architecture in a 2D engine without running into the occlusion problem or having the scene look like your sprites are marching around inside a painting.

Basically, what I have made are tools that allow you to mod not just individual items and quests, but tools that allow you to mod the gaming system itself. You can select for example how feats work and are gained, how leveling up works (or choose a usebased system and set usage levels and tie skills to actions and/or stats to determine how skills increase - so you can make a system like darklands or silent storm if silent storm didn't have quirks), feats themselves, how magic works, etc.

This is exactly what I had in mind with the abstracted game mechanics. I.e., the engine would have hooks that you can link up to stats in the game mechanics, but would not in itself take a stand on what the game mechanics are.

The scaling is not a big deal, it is mostly just a programming challenge and not a particularly hard one.

Occlusion is a pain in the ass, but in a falloutlike system I think it's not a big deal. It just takes time and effort, not any special magic. Once you have the general case done, too, any game should be able to benefit from it. It is a problem I have not gotten to yet, though, to be honest. I am not really very far along. I just have tools right now.

 

[Prime Junta]

The scaling is not a big deal, it is mostly just a programming challenge and not a particularly hard one.

It's also a design problem. It's not at all obvious to create 2D graphics (sprites, textures etc.) that look good when scaled significantly. And I'm not talking about interpolation algorithms here.

 

[bryce777]

The scaling is not a big deal, it is mostly just a programming challenge and not a particularly hard one.

It's also a design problem. It's not at all obvious to create 2D graphics (sprites, textures etc.) that look good when scaled significantly. And I'm not talking about interpolation algorithms here.

Well, as I said I am not the big art guy, but I think it just requires you render them down several times instead of once from the originals. You may need to tweak your models a bit for each rendering. So, it is some work there, too, I imagine.

 

[Prime Junta]

The scaling is not a big deal, it is mostly just a programming challenge and not a particularly hard one.

It's also a design problem. It's not at all obvious to create 2D graphics (sprites, textures etc.) that look good when scaled significantly. And I'm not talking about interpolation algorithms here.

Well, as I said I am not the big art guy, but I think it just requires you render them down several times instead of once from the originals. You may need to tweak your models a bit for each rendering. So, it is some work there, too, I imagine.

Indeed. If you're serious about this, btw, I would recommend you find some art guys pronto -- there's no way you can design tools they want to use unless you're an art guy yourself or are in constant contact with them so they can tell you exactly what works and what doesn't. (Trust me on this one; I have personal experience of both failing due to lack of user involvement, and succeeding largely due to active user involvement all through the design process.)

 

[bryce777]

The scaling is not a big deal, it is mostly just a programming challenge and not a particularly hard one.

It's also a design problem. It's not at all obvious to create 2D graphics (sprites, textures etc.) that look good when scaled significantly. And I'm not talking about interpolation algorithms here.

Well, as I said I am not the big art guy, but I think it just requires you render them down several times instead of once from the originals. You may need to tweak your models a bit for each rendering. So, it is some work there, too, I imagine.

Indeed. If you're serious about this, btw, I would recommend you find some art guys pronto -- there's no way you can design tools they want to use unless you're an art guy yourself or are in constant contact with them so they can tell you exactly what works and what doesn't. (Trust me on this one; I have personal experience of both failing due to lack of user involvement, and succeeding largely due to active user involvement all through the design process.)

I am a long way from being that far along. I usually spend maybe 10 hours a week on this project. When I am further along I will definitely seek out some art people, though.

 

[Wysardry]

How would it differ from any of the other RPG engines available or in progress at Sourceforge and elsewhere?

 

[Oarfish]

The main difference would be it if it made it past version 0.2

It's something I have considered messing around with myself, though I think designing a meta gaming engine may be even more complicated than designing a single game I just don't have the art or writing skills to do solo game design.

As for tools, writing the high level engine and editors in .NET would almost certainly give a productivity advantage over pure C++. Fair performance hit and platform lock in though.

I do like the idea of a declarative RPG rule and modelling system using a domain specific language for calculating probabilities, defining attributes and the effects of actions on actors rather than hard coding rules into the engine. Could also be used for simplifying social stuff as well as the physical with a decent semantic model and querying - determining inventory or attributes could use the same extensible mechanism as faction allegiance or alignment.

Pie in the sky stuff though, I'm a lazy bastard and have about 3 million things to get round to doing that never happen when not on a contract.

 

[Prime Junta]

How would it differ from any of the other RPG engines available or in progress at Sourceforge and elsewhere?

Such as...?

Enlighten me -- as I said, I only did some quick looking up of stuff, and didn't find too many OSS RPG engines other than FIFE.

Edit: let me rephrase that -- I didn't find too many *serious-looking* OSS RPG engines other than FIFE. There are plenty of build-your-own-adventure kits, but all the ones I looked at are tied to a specific game system, extremely limited in features and functionality, and pretty primitive technologically. That said, there were a lot of them; if there are some like what I described already going strong, awesome.

 

[Prime Junta]

I do like the idea of a declarative RPG rule and modelling system using a domain specific language for calculating probabilities, defining attributes and the effects of actions on actors rather than hard coding rules into the engine. Could also be used for simplifying social stuff as well as the physical with a decent semantic model and querying - determining inventory or attributes could use the same extensible mechanism as faction allegiance or alignment.

That was precisely what I had in mind. The engine has a certain set of things it can do, and by abstracting that it's possible to link it to any kind of back-end that feeds it content, resolves calls, and tracks stats and variables. An engine event would simply hand back a container and receive a resolution.

Um, sounds really simple. On paper.

Pie in the sky stuff though, I'm a lazy bastard and have about 3 million things to get round to doing that never happen when not on a contract.

Same here. I'm currently pretty solidly chained to my desk for at least another two years or so -- but I feel like doing something completely different after that, and setting up a small-scale (but commercial) game studio is one possibility. So while these ideas and questions I'm throwing out here are nothing but hot air at this time, there is a small possibility that they might turn into something else; if this still feels like a good idea six months form now I might try to get my hands dirty with some of these engines. It'd be kinda cool to put my 22 years of D&D to work...

 

[DarkSign]

To create a 3d rpg engine would be a gigantic undertaking. The closest thing that exists out there is Realmcrafter (specifically for MMOs but could be used for single player). But its graphics suck and its based on crappy Blitz3d.

OGRE,Crystal Space, Irrlicht et. al. arent game engines (meaning rendering engine + possibly physics + possibly client programming etc.) at all. Torque bills itself as a game engine, and rightly so because it has a level editor, asset management and its own scripting language based around database classes. Still its not a do it yourself RPG wizard.

If you've got a few years, a programming degree, and a team of people to help ya...go for it!

 

[Wysardry]

Such as...?

I'm not sure exactly what you had in mind, but following this link will run a search for "+3d +rpg +engine" at Sourceforge.

Sauerbraten and IrrLichtRPG look most interesting from my point of view, as I tried the core engines in the past and found they ran on relatively modest systems.

You could also try searching Devmaster.Net's 3D Engines Database (you'll need to use the advanced search if you want to specify a price range).

 

[Prime Junta]

Such as...?

I'm not sure exactly what you had in mind, but following this link will run a search for "+3d +rpg +engine" at Sourceforge.

Sauerbraten and IrrLichtRPG look most interesting from my point of view, as I tried the core engines in the past and found they ran on relatively modest systems.

You could also try searching Devmaster.Net's 3D Engines Database (you'll need to use the advanced search if you want to specify a price range).

Perhaps I'm still missing something, but I don't see anything there that would fit the spec I outlined in the post at the top of this thread -- the projects appear to be either (a) dead, (b) extremely primitive/early stages, (c) twitch "action RPG", or (d) not open. Sauerbraten does look interesting but at least they describe it as an FPS. Not what I'm looking for; I want to cut out complexity like physics.

 

[bryce777]

Ok, here is something that I have sort of wondered.

I have a ways to go, but maybe I am off track in some of my thinking. I have never seen the huge benefit of a scripting language, though. It seems to me if you are making tools, everything should be doable through your tools. Including creatng items, characters, and quests using these items and characters. Quests really only do a few actual things. Sort of like how a cathedral is just made of bricks. Once you allow them to do these things then you can make any type of quest. This is what I was working on last time right before I ran into some technical problems that have made me sort of rethink what I am doing which were unrelated to this.

Is it just that the tools have not been fleshed out enough to avoid scripting? It seems like if youa re scripting out then you are not going to be able to truly debug any of your quests fully without extensive playtesting, whereas if you create them through a tool that takes all the cases into consideration then you can easily check to make sure that any quest is completeable and makes sense, though you still obviously have to make the dialog make sense on your own.

 

[obediah]

Ok, here is something that I have sort of wondered.

I have a ways to go, but maybe I am off track in some of my thinking. I have never seen the huge benefit of a scripting language, though. It seems to me if you are making tools, everything should be doable through your tools.

I've never done either with a game, but in general a scripting language is more flexible than a gui tool, and scales better with complexity. Take SQL for example, for a simple query an ACCESS screen or html interface is easier than writing SQL and maybe even faster to use. However, it's very easy to think up useful queries that would destroy ACCESS or any of the *Admin web interfaces. They still aren't easy to write in SQL, but they are possible.

A scripting language with a good syntax and well made API is also very fast to write and very simple to read and debug. There's also no reason you couldn't incorporate verification into the language. Although you're showing your age by mentioning verification at all, that went out the window with OO programming.

 

[bryce777]

For that matter, what would your dream feature list be when it comes to modding?

I can probably do nearly anything that you guys can think up with enough time and effort. Even ui creation and modding to some degree.

For example, paper dolls. I have not worked on this at all, but here is what I would like to do:
You make a box and size it, then you pick out an already defined item type which you can assign to the box. So, for example you make a box and make it of a ring type, then you place it on the template where you like and you choose the place on the paper doll and a line is drawn to it.

Then, for individual characters you just need to create the paper doll graphic.

I would also have the obvous combos as default templates. You could also make templates just for the purpose of having a much different looking model - such as for a drawf or modron, or whatever.


I guess there's a fine line, because I have a lot of ideas and do not want to become overwhelmed with new ones, but at the same time I do not want to go off into weird areas which are not in practice what modders truly need.

 

[bryce777]

Ok, here is something that I have sort of wondered.

I have a ways to go, but maybe I am off track in some of my thinking. I have never seen the huge benefit of a scripting language, though. It seems to me if you are making tools, everything should be doable through your tools.

I've never done either with a game, but in general a scripting language is more flexible than a gui tool, and scales better with complexity. Take SQL for example, for a simple query an ACCESS screen or html interface is easier than writing SQL and maybe even faster to use. However, it's very easy to think up useful queries that would destroy ACCESS or any of the *Admin web interfaces. They still aren't easy to write in SQL, but they are possible.

A scripting language with a good syntax and well made API is also very fast to write and very simple to read and debug. There's also no reason you couldn't incorporate verification into the language. Although you're showing your age by mentioning verification at all, that went out the window with OO programming.

"Although you're showing your age by mentioning verification at all, that went out the window with OO programming." Ha! Well, that would explain a lot.

I think the problem I have is that the complexity simply skyrockets once you allow for really generic game system creation.

For example, I allow npcs/quests to have totally different options based on 'reputation'. There might be any number of reputations in the game - this might be a simple good/evil meter or it might be a reputation ala how friendly the gwelfs are or how friendly the ghibbolines are.

I suppose that the system will not NECESSARILY be using the maximum complexity possible, but I imagine when the possibility is there many people will use it.

This could lead to some real ugliness like a whole bunch of nested switch statements which each need their own arguments.

I suppose it depends on how well I can pin down all the possibilities. Maybe for special stuff I can fill in scriplets sort of like the RAD libs that is the VB form editor. that way you could stick some event into a quest that was 99% taken care of automatically. Well, I guess I would have to get back to this and try it out and then have other people try to use it at some point.

My next goal was to make a tester where all the people and items involved in a quest would each be in a room and you could go from person to person and talk to them and test out not only that it worked but that your dialog was nice and made sense.

I suppose the only part of this is chained quests tha overlap where one person might not give you some clue or whatever at first.

This is sort of the heart of the whole thing, though, so I am sure I will rethink things many times.

My other Ultimate Goal is to somehow make ways to ensure that dialog always equates to GAMEPLAY. Your choices should matter, and not just from a standpoint of 'roleplaying' in the sense of wearing a fancy wizzard lord hat or talking in a gruff dwaren fighter's voice.

I want to figure out some way to incorporate that more in the actual way conversations work, but I can't seem tot hink of anything satisfying.