Engine

[Prosper]

Now I'd like to know which features the FIFE is lacking in your opinion.

1. It is a 2D engine. It is better to do emulate 2D in a 3D engine (like New Super Mario Bros) than try to fake a 3D feel in a 2D engine.
2. It doesn't support audio other than Ogg Vorbis
3. No physics engine
4. Exceedingly basic pathfinding. I want clever enemy AI.
5. Exceedingly basic lighting
6. Only supports a handful of resolutions
7. Isn't ported to many platforms (no Mac, no iOS, no Android, no consoles)
8. Little to no shader support
9. Archaic UI tools

The UI tools are fine. Shaders aren't a problem. FIFE is FREE , no consoles is not a loss of kickstarter money. More money means more game. What game? This game , W2, not W3 the FPS.

Pathfinding isn't complicated to implement in any 2D engine. LOL BASIC LIGHTING. I think you are trolling , but I am not sure. Your physics engine complaint is legitimate however. But why would you want an entire engine for physics? The effects the user would take notice of are all small bodies of code.

[adomolis]

You guys serious here? I know Wasteland is not about graphics here, but for the love of... you really want a game to look like from 1995? And this FIFE engine... looks plain terrible. Its 2012 guys!!! There's loads of tools you can create awesome looking environments with.

I liked how the new x-com looks like, so if Wasteland 2 turns at least half of what it looks like here, I'll be happy enough.
(its using Unreal Engine 3)

[valcik]

Or 25USD each day for pizza.

Rock on, Prosper!
Just beware of art-I-choke.

[Prosper]

You guys serious here? I know Wasteland is not about graphics here, but for the love of... you really want a game to look like from 1995? And this FIFE engine... looks plain terrible. Its 2012 guys!!! There's loads of tools you can create awesome looking environments with.

I liked how the new x-com looks like, so if Wasteland 2 turns at least half of what it looks like here, I'll be happy enough.

Um. The FIFE Engine is perfect in every way. Don't be so harsh. Download the engine and the demos. Copy and paste quality game sprites over the demo sprites. Rerun the demo. the important thing is it can load high quality textures. Just pixels yo!

[adomolis]

Um. The FIFE Engine is perfect in every way. Don't be so harsh. Download the engine and the demos. Copy and paste quality game sprites over the demo sprites. Rerun the demo. the important thing is it can load high quality textures. Just pixels yo!

I would have to stick to Unreal3 engine anyways. Its a modern engine. Full 3D. It has a huge library of API's and stuff. Its very very customisable and very well developed plus modern (been updated to dx11 recently). Building a game on the engine costs only 99$ plus 25% royalties from sales, but I think Fargo could negotiate for 20% or even 15% royalties. It may look much but it would save hell lots of time and investments when you think of it how much time and money a new 3d engine form scratch would cost.
I think U3 is a choise very much to be considered.

[dmazz]

Artists can spend months on a single 3D model in a 1st person game and that would make a lower budget title impossible

Brian with that sentence admits that 3D has it's downsides, primaily that textures need to look good, or the 3D art looks like crap, this is especially true for 3D humanoid characters. And so by saying that, he would favor 2D art for it's ability to look really good with a small outlay in time. Especially when it comes to humanoid characters.

Going 2D but doing it in a 3D engine, gives one flexibility in the art department. Some things can be 3D others 2D. It's all good. The only important thing is that it will be top down/isometric in perspective, this is to save time and only needing to make one version of 2D art. So those who were hoping for a movable camera in 3D flying around will be sorely disappointed. But the 3D camera and 3D animated humanoid characters are possible in the sequel given the 3D engine.

We do have the advantage of this being a top down game which saves tremendously on the art creation which in turn allows us to script out numerous outcomes without the concern of creating graphics for every possible situation.

It's not that confusing if you read between the lines.The first sentence claims the advantage of 'top down' is less art creation because there are fewer "situations". I take to mean he means by 'fewer situations' fewer camera angles and unique movement. Unique movement of 2D art assets would mean animation, that requires alot of work if you want it to look good. And different camera angles require different versions of 2D art too, since it isn't in 3D. Brian is talking about the advantages of 2D, which are erased if you start moving the camera around or need your 2D art to move/change in some way.

The question about the game engine was open ended and finished with a 2D or 3D. So it stands to reason Brian would have understood that question to be '2D engine or 3D engine'. Which is a little non sensical given that you can quite easily show 2D graphics on a 3D engine. So he interprets the question to be asking 'will the game be in 2D or 3D', and he says "Clearly this will be a top down game for the reasons I stated above." So he confirms the game will be at least primarily in 2D. What is unclear is whether the engine is 2D or 3D. From what he says just previously to that sentence, it suggests that he will be using a cheap engine, they won't be spending their money on technology. This suggests the engine may be in 2D.

Brian also says "We need as much of the budget as possible to be spent on assets and design and not technology." So here he explicity states he's budget concious when purchasing technology for the game. And for this reason I think the 15 and 30K Unity and Gamebyro engines are too expensive. Use of middleware is also under doubt due to the expense. The use of the Unreal 3 engine is still possible though given it's little to no up front costs. Price wise I think he's looking for something really cheap, he doesn't mention speed, so it seems he plans on building his own tools for the most part, which will be planned out in the design phase also. So given that, in regards to an engine we are talking as cheap as possible

Some caveats though would be sound. Due to the advanced sound system planned for the game he might be preferential to advanced sound support. Brians clear dedication to turn based combat will require a very complex A.I to be developed, something middleware A.I software can help him do. So an engine that supports both A.I and sound middleware, might be strongly preferenced.

Torque 3D appears to be the cheapest bare minimium candidate available. It supports FMOD sound software ($3000) and has a A.I addons and kits you can purchase. I'll add a seperate A.I software licence price though. ($8000) Studio licences for a small team start at $6000 So Total approx price of setup would be 17K+

I'm really hoping they get the Unity engine. It supports the xaitment A.I suite, for which a 'Brain Pack' full license for one year will cost them $8000. And the Miles Sound System $4000 approx. Engine cost with source is 15K, Total cost 27K

These setups are the only decent ones I could find. Brian is dedicating as much of the money as possible to art/content creation, but because the game art is in 2D, they will always have more art than they have time to implement. This art can be saved for future expansion packs. There might not be enough time to develop the complex A.I needed so the A.I software purchase is necessary. Apart from their art, sound will be the second most powerful tool they have for creating immersion, and so great sound software should be used. Brian in an interview mentions he has something special in mind with the music in the game.

I gotta agree with the price for a 3D engine with linux/mac support they seem quite high. For an engine with support, that decent and supports some essential middleware products it starts at 17-27-42 (Gamebyro), then it just goes way up in price, due to revenue pricing. I don't really know the details but it would seems the 2D engines would be significantly cheaper. But the approx 14K for the A.I and sound middleware is a pricetag that won't change. And I doubt many if any 2D engines support such middleware out of the box.

I also read that no commerical developer would ever use an open source engine. The support network just isn't there, if you have a problem putting a post up in a public user forum isn't good enough. This is even more important for this project where there is only 12 months to develop the game. You can't expect them to just wait for days for their questions to be answered. That's why open source engines are a no go. And any engine that hasn't got a robust support network, so no proprietary inhouse engines, or fairly new engines with few game to their credit. (like unigine)

Saying that, perhaps Inxile isn't doing this for profit. In this case they may use a open source engine, but that doesn't mean it won't be a developmental nightmare experience. A better approach to satisfy the fans would be to just make a good mod kit, then later they could free up all the art assets in the game. No way in hell it will be FIFE, no offense.

[Landslide]

Sure, it's older but it had a couple neat features. The biggest of those was structural durability in most of the 3D map assets. Meaning if the damage amount applied to an asset exceeded its durability, it got broked. This was fun, as you could have machine gun fire chew through walls, you could blow out floors with TNT and watch enemy squads fall to their death. You could access buildings through inventive means (bazooka through the kitchen wall). It made the tactical aspect really appealing and replayable. It felt like a sit-down PnP situation, rather than a system with hard rules.

Embracing the sandbox aspect, I think, if going to be an important part of this project - and allowing players to drive he gameplay will be the biggest contribution to that end goal. Whatever engine or visual display method is decided upon, I think it's important to stay as far away from system-induced rigidity as possible.

/2cents

[dmazz]

The engine + tech prices don't look unreasonable. Let's say 2 million dollars are raised. 27K would be 1.35% of the game budget dedicated to technology. That's a very small number.

In response to a previous question, I think the 6 month preproduction phase will be used to produce tons of 2D art (drawings, paintings, stuff like that). They'll be given the text and then told to draw something. Which later in the development phase will be scanned, touched up and finished in the computer and used as the primary art assets. This approach will speed up development tremendously and it's only possible due to going 2D for the art.

[GROUNDZERO]

What about the Neverwinter Nights 1 engine? It´s old, but it runs great on old computers and the graphics still look good. Just a thought.

[Bryce777]

The engine + tech prices don't look unreasonable. Let's say 2 million dollars are raised. 27K would be 1.35% of the game budget dedicated to technology. That's a very small number.

In response to a previous question, I think the 6 month preproduction phase will be used to produce tons of 2D art (drawings, paintings, stuff like that). They'll be given the text and then told to draw something. Which later in the development phase will be scanned, touched up and finished in the computer and used as the primary art assets. This approach will speed up development tremendously and it's only possible due to going 2D for the art.

Unity has never had a real game release on it. It's an engine that focuses on cell phone games. It's come along a lot from what it was but it doesn't have multithreaded skinning meaning it can't do this game, period. It just can't have enough characters onscreen at once, half a dozen max.

It also doesn't let you draw on the screen directly in 2D. That means you are stuck with their completely inadequate GUI system. That means you can't do this game it's just not possible.

You can make a corridor shooter with it, or a sports game like golf, or a first person perspective RPG without many characters onscreen at once, that's all it can do.

[geezer]

Dmazz, how would you go about creating a game character with 2D art and only one bitmap? I'm really curious because if you do have a way to do that it would really save a lot of time for 2D development. I'm genuinely all ears. Are you thinking in terms of some kind of image recognition algorithm which will redraw your single bitmap on-the-fly for each change in perspective?

Also, please explain how a change in perspective from isometric (axonometric) to first person increases the number of bitmaps required. Even a genuine top down game like Driftmoon would seem to require more than one bitmap if implemented in 2D. Although maybe there is a way to rotate the position of the pixels around the z axis in that case.

[dmazz]

Your right it, the Unity engine is graphically very limited. But my prediction is this game will have limited graphics in 2D, something inbetween Advance Wars and Commando 3. This turn based rpg game uses Unity http://300ad.com/styrateg.php

Regarding the FPS view, I was talking about the perspective only in 2D, not 3D. So you only need one bitmap for one character. (think character view in Advance Wars) Dumb way of expressing myself but I'm a layman.

Unity's GUI deficiency is a problem, but it will get Scaleform support around April. It also seems to be a ambitious, growing company, so I'm optimistic of their support quality.

Regarding Driftmoon, this is what I read on it "when making the Driftmoon engine, the intention has all along been to make a sort of RPG Maker for top-down games. Comes with the same tools used to make the whole game." So this definitely might qualify as a total conversion candidate. Quite a few unknowns though, it's doc quality, whether they would agree to some money for support arrangement, if the source code license is even available. A negative is the combat looks real time, and not turn based, requiring the combat system to be created from scratch. Very time consuming, and something that would need to be done without A.I middleware support of any kind.

Also it's unclear whether the Unity 15K and Gamebyro 30-45K (30K for binary 45K for source) engine licences, are for more than one person or for a period of time. So the tech cost to outfit a team could be significantly higher depending on the licensing arrangement. Also given this email I found from Gamebyro, "For full priced AAA boxed titles we’re talking in the realm of $180,000USD for a PC source license. For Casual titles [...]. The source code version is $45,000 USD and the binary (as found in the evaluation) for $30,000USD. " they seem to take alot into account when giving a price, so Inxile might get it for cheaper given it's kickstarter origins. Gamebyro may be graphically overkill now though.

Giving things more thought, the Uneal 3 engines looks unlikely. For a game type like this largely extinct for almost a decade, there will be a strong desire to see Wasteland 2 grow and not just be a one off. So I think Brian will want to reinvest the money made with Wasteland 2 back into it with expansion packs, DLC,a mod kit and maybe even a sequel. Unreal will be taking a good chunk of that money given it's revenue pricing scheme, and that's not good.

[Gizmo]

What about Dark Basic Pro? The engine is commercial, royalty free; and costs under $1,000
(Actually it costs about $50, but there are a lot of commercial addons).

http://www.youtube.com/watch?v=7Rn89gHPxEM
http://www.youtube.com/watch?v=MyP1pyFh6eM
http://www.youtube.com/watch?v=PW1tvDc9cd8

http://www.youtube.com/watch?v=Vh1BApxIUxo

http://www.youtube.com/watch?v=aGw4JoYzPT4
http://www.youtube.com/watch?v=19bzR3cq520
http://www.youtube.com/watch?v=UzYUuxwtM8A

[L1cH]

Anybody hear about Ethanon Engine? 2d engine with pseudo 3d effects.

Ethanon Engine is a free 2D game development tool focused on the recent video cards for hardware acceleration and shader effects. It may be used to develop any style of game, such as top-down view RPG's, side scrollers and shoot'em up games. The main goal of this engine is to provide high-quality lighting, shadow and particle effects with an easy to use visual interface and a simple and direct scripting system.

EE is entirely based on 2D sprites. It uses pixel shaders to perform the normal map based lighting and supports any kind of map construction (e.g.: isometric, orthogonal or hexagonal) because it's based on arbitrarily positioned entities instead of fixed tiles.

To control the game flow and handle its events, Ethanon Engine uses AngelScript for scripts, which brings a familiar C++ syntax to the code. The integration with the scripting language is pretty simple and easy to understand. Many features are included in the scripting system, such as full entity manipulation, save and load games with one instruction, video playback support, and more.
http://www.asantee.net/ethanon/

Video demonstration
http://www.youtube.com/watch?v=FMFIaxrINVU

[dmazz]

DarkBASIC Professional would work, but from the looks of it has no support for A.I, sound or GUI middleware. It has addons but they don't offer nearly the level of functionality or development speed. I also can't find any commercial or even finished games made with it and that's not good.

The Ethanon engine is unsupported and incomplete, and so can't be used. In this regard it's similiar to the XNA RPG maker http://www.youtube.com/watch?v=-Q6ISVaM5Ww
http://www.youtube.com/watch?v=vtYvNEmmHXE
They look impressive though.

In regards to the engine. Here's a nice quote from a survey on choosing an engine.
"What factors do they use in deciding whether to license a game engine, and which one to choose?
The overwhelming favorites were of course cost and time. If we purchase a game engine, we want to spend less money and time than we would have creating the engine on our own. Otherwise why would we bother? In third position to money and time came genre relevancy, or making sure that the game engine works properly for the type of game you are creating. This emphasizes the importance of doing a thorough evaluation of any game engine you’re looking at, and knowing if a particular engine was already used for a game technologically similar to what you are planning to create (a first-person shooter, for example.) Also highly important to the responders was support and thorough documentation, a good-quality content pipeline, and the ability to easily integrate the game engine into one’s existing technology as well as with other middleware libraries."

Apart from a minimium graphical level ability which at this point in time is unknown. The engine also needs to have some linux or macosx support. A decent support network needs to exist in some form also. Those are the minimium requirements imo. Software support for GUI, A.I and sound middleware would speed up development but add to cost.

A bit more information about the Silent Storm engine. It's being used to make Jagged Alliance 3 now being worked on by bitcomposer a german developer, who has no major pc games to their name. But they did publish STALKER Call of Pripyat a post apocalyptic RPG/FPS shooter and were co-publishers of Jagged Alliance online, a web browser tactical MMORPG, currently now in beta stage. Strategy First is the canadian owner of the Jagged Alliance rights, and developer of many turn based strategy games. Everyone involved here would love to see Wasteland 2 get made, and be a success, so they may be receptive, and agree to let InXile piggyback on the development of Jagged Alliance 3, creating the required support network that's missing for the silent storm engine.

A open source possibility is the Ogre engine (+CEGUI), probably the most popular linux engine, and used in more than a few commercial games. Recently accepted again in Google Summer of Code, so still active as ever. Torchlight, a game was made with it in 11 months by a at least 14 member team. There's also the NeoAxis 3D Game Engine, which is based on Ogre and has a commercial team license at 1K. So this looks like a doable semi-open source alternative. Cube 2: Sauerbraten engine is also a possiblity, but less active but it has an official commercial version which provides support. In addition it has some inbuilt A.I functionality which may be useful. Though I can't find any commerical games released with it.

UFO Extraterrestrials is a turn based combat game and a total conversion may be possible. Not sure how much code can be salvaged, but the sequel for the game is due to be released this year. So the engine has support available. The developer is located in Czechovslovakia though. http://ufo.ufo-extraterrestrials.com/

Coming back to the Onyx engine, used in Dungeon Siege III. I remember Brian mentioning Feargus Urquhart CEO of Obsidian Entertainment accompanying him in his publisher pitches for Wasteland. So we can be sure him and his team will be asked about what engine they should use. And they happen to have built and own the Onyx engine (and recently announced they would be using it to make more games). This makes me think the Onyx engine is a prime contender as they will get a special deal and probably be able to hire a Onyx engine master programmer. http://www.youtube.com/watch?v=1Y4pOUQ8 ... r_embedded The engines major problem is it probably doesn't support any A.I middleware, which if used will speed up A.I development for the turn based combat system.

If not Onyx my bet is on Gamebyro. It seems to have a faster content pipeline than Unity and better overall tools.

Going back to the graphics question. Fallout took at least three years and 4.2 million dollars to make. (adjusted for inflation) Wasteland 2 will have approx half the budget and half the development time. And don't forget the simplistic combat system in Fallout, Wasteland 2 is planning to have a far more complex combat system. So it will be a challenge to make a game that's half as long and half as pretty (in 2D), but just as good. So corners will need to be cut most definitely, and the only place where that's possible is in the art detail department. No talking heads and voicework like in Fallout is a good start, but not nearly enough. (voicework doesn't take much time too)

So if it's in 2D, background detail and 2D character detail and animations will need to be reduced compared to Fallout. I'm thinking the 2D art will start to look like something inbetween Command & Conquer: Red Alert released in 1996 and Fallout released in 1998.

But if it's done in 3D but the camera is always in a fixed isometric perspective, shortcuts can be taken, and the overall graphical quality would be at the Tomb Raider: Chronicles level http://www.youtube.com/watch?v=lNbFQJf3JTk (the game engine was developed in 1996, this game released in 2000). But will look better due to the superior tools we have now for vegetation, lighting, shadow, particle effects and 'off the shelf' textures. And cause 3D objects just interact better with these new features.

[Gizmo]

DarkBASIC Professional would work, but from the looks of it has no support for A.I, sound or GUI middleware. It has addons but they don't offer nearly the level of functionality or development speed. I also can't find any commercial or even finished games made with it and that's not good.

My guess is that any team that could make a professional title in DBpro, could make it C/C++. Though it would be a cheap way to prototype. I dunno if it's got some hidden flaw (stability, or lacking something crucial) that would make it a poor choice.

* It's bad that I could only find a couple retail games made with it, but since it's royalty free, why would a developer bother to list the development environment? We can't really know what's out there that was made with it.

[hildiri]

The Jagged Alliance 2 engine is close to ideal.

As for gaming dynamics JA2 was the best at its time...I have a very strong belief that W2 will do the same for our time now...or better?!

On the other hand I expect a bit better graphics & engine than the one I saw in the videos of OP's post.

[dmazz]

In regards to Xenonauts lackluster graphics. I believe the art is all in 2D and the engine itself probably doesn't even support 3D. I will say the 2D art lacks 'style', appears too utilitarian. That it has fully destructible terrain is impressive though.

I've heard the expression there are more game engines than games for linux. There are alot of unproven open source and commercial game engines. Developers don't want to be the first user of a game engine lest they get a nasty surprise. They want a well proven game engine used by tons of developers used to make a variety of games. The choice is even more crucial for smaller developers who have less time and money available to make a game with.

Regarding prototyping, Brian mentions that the (lengthy for a game with a 12 month production cycle) 6 month preproduction will be crucial, something I think indicates that this stage will serve as the prototype stage for all the game mechanics and visual design, done with pen and paper. The 12 month development stage will be used for implementation and filling the game with content. So given that development strategy I'm thinking they'll want a game engine with the fastest content pipeline and ease of programming possible.

Also just completing the game in 12 months isn't good enough. They'll need a significant amount of time for play testing, balancing and polishing. So they will need to finish the game in less than 12 months, say 10 months.

If they use the Onyx engine they'll need to hire an experienced Onyx engine programmer from Obsidian entertainment, and get a crash course in programming for it. I may be underestimating the support obsidian can provide to them though. The gamebyro engine probably a better choice, because apart from it's superior content pipeline and scripting it's also been used by Obsidian entertainment programmers to make Fallout 3 New Vegas. Obsidian entertainment added a significant amount of features to the base Fallout 3 code, so given this they would have some skilled gamebyro engine programmers too.

ps
Saw a great recent Tim Cain interview (obsidian entertainment), where he talks about his involvement with Fallout. http://uk.gamespot.com/shows/gamespot-l ... on20120308
The game took 3 1/2 years to make, no specs or plans initially, just one guy working on a engine for the first six months., 1 artist and 1 scripter for the next six months, 15 people in year 2, 30 people in year 3, 4 man months per talking head.

Given fallouts development history, this is good news for Wasteland 2, whose disciplined and strategic approach to get as much done as possible in 6+12 months is the complete opposite of how Fallout was made.

This helps me better understand Brian's strong desire to not spend alot of money on technology and to maximize the amount of money spent on content for the game. And understanding this mindset better, I realize he's against using the Unreal engine due to how expensive it is, this also disallows other expensive engines like Cryengine. Brian doesn't think expensive engines like Unreal's are suitable for fan funded games like this, he used the Unreal tech for his two previous games, but that was money entrusted to him by a corporation, and they would understandably feel comfortable with the use of the unreal engine given it's popularity, fast development time and no technical surprises. The fan community for this game though would strongly object to the Unreal engine being used, primarily due to it's expense.

And moving along this line of thought which takes into account fan desires. We need to take into account future mods. A mod kit created for the game would have vastly increased powers if it was integrated with a commerical engines toolset. But for this to occur the game engine would need to be free to use for non commercial purposes. That narrows our choice of engines significantly. The Unity engine recently now offers a free license version of their software. The Gamebyro engine offers a 'free evaluation', but only if you fill out a lengthy registration form. Torque 3D doesn't offer a free version of their software. And as for the Onyx engine it is unclear how far Obsidian Entertainment will take it. They announced their thankfulness for being able to use it to develop future games in a twitter post. Currently it's being used to make South Park: The Game, an RPG and in the future a The Wheel of Time game, both potentially major game franchises. So it's quite possible a a toolset for the Dungeon Siege III engine similar to the one for Neverwinter Nights 2 will be released in the future for one of their RPG's. Wasteland 2 would make an ideal candidate for such a powerful tool. A powerful tool like that would make future user texture packs and even special effects very easy for modders. Somewhat alleviating the pressure for Wasteland 2 to have awesome graphics out of the box, given it's limited 12 month development cycle.

So hopefully the Onyx or the Unity engines are used. (Unity also supports C#, Python/boo and Javascript for scripting, so modder friendly) Gamestudio another engine http://www.3dgamestudio.com/ also has a free version, and so does Shiva3D, http://www.youtube.com/watch?v=UCgMCupq_x0 but both don't have any middleware support.

Also check out this interesting piece of info. On March 13th, 2012, Obsidian cancelled a "future next gen project", resulting in the layoff of approximately 20-30 employees both from that project and the upcoming South Park: The Game team. That's 20-30 possible Onyx engine programmers available for hire.

[rogerdv]

Unity doesnt support Linux, and they wont support it unless at least 170 licenses asking explicit linux suport are sold. According to them, supporting a linux port would cost around $270k per year. dont know about Onyx. And think you mentioned Gamebryo in a previous post, AFAIK Gamebryo is dead, Emergent went to bankruptcy last year. Bethesda Creation Engine was built for that reason.

[TheEmissary]

Unity doesnt support Linux, and they wont support it unless at least 170 licenses asking explicit linux suport are sold. According to them, supporting a linux port would cost around $270k per year. dont know about Onyx. And think you mentioned Gamebryo in a previous post, AFAIK Gamebryo is dead, Emergent went to bankruptcy last year. Bethesda Creation Engine was built for that reason.

Emergent isn't the current owner of GameBryo hasn't been for while now. Bethesda Creation Engine is still very much GameBryo just highly customized. Alot of the same engine bugs from the old Elder Scrolls games are in Skyrim some more obvious than others. It is also no different than Activision pretending that Call of Duty isn't using a idTech engine.

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

Unity can technically be ran on Linux through Google's Native Client plugin system. That or the unofficial method of using wine.
http://unity3d.com/support/documentatio ... arted.html