The selection pool for both students and organizations is a competitive one.  We’re hoping that some slight adjustments to the manner in which we put together and submit our application as well as the huge gains the project have made will help to mark 2012 as the first of many successful years being a part of Google Summer of Code.

What Changes are we Making?

Based on feedback we received from Google after our rejection last year as well as considerations for where the project currently is there are two core changes that will be made to the community’s application effort.

Seriously Awesome Ideas, Seriously.

The biggest of these changes is how ideas will be documented.  Rather than a large page with all of the ideas and a sentence or two plus bullet-points for each, a [hopefully] more complete approach will be taken with individual pages devoted to each topic.  These pages will include: Idea Abstract, Detailed Technical Description, Rationale, Expertise Required for Completion, and Point of Contact (Mentor) Information.  The hope is that by providing a detailed breakdown about each of the ideas that it will be made clear what we are striving for and what would be expected of participating students.

Different Project Introduction

Exactly how we go about this is something still being discussed (to which we offer any and all suggestions), but the goal is to come up with an introduction to jMonkeyEngine that highlights the extraordinary community that we are.  Think of the “little” things you see on the forums every day and what that says about us.  Words of encouragement to newcomers, applaud for successful commercial releases, excitement for cool new contributions, the list goes on.  So the question is What has attracted you to jME?

Important Dates and Deadlines

• 4 February 2012, Google Summer of Code 2012 Announced
• 7 February 2012, News entry posted to jME community site announcing the effort
• 9 March 2012, Application deadline for organizations (that’s us!)

Let’s Do It!

So, let me just start by saying, WOW! I think a lot of us saw great potential in 3079, but I’ll be the first to say, I never expected such an explosive popularity, at least not so soon. What can you attribute it to?

Thank you. I couldn’t have done it without you and jMonkeyEngine’s community help.

What can i attribute the explosive popularity to? Well, it started to ramp up quickly after posting a trailer to reddit’s gamedev community where it was well received. A public relations guy for the game Towns, Alex Poysky, saw my game and he has really helped me get 3079 to more places. Yogscast also did a video of my game quite early in development, which helped quite a bit.

Erlend: hah, no doubt. almost 1 million views on that video now I believe

Jeremy: oh wow!

Do you have a favorite statistic thus far?

I was #1 on Desura and IndieDB for a bit. I was also the most common search term to get to Indie Game Magazine‘s website around that time

Let’s talk more about Desura. What has your experience been like with this up-and-coming payment platform?

It has been overall pretty good. It was a bit of back and forth to get it actually up on Desura. Basically, you have to show that your game will sell, and I was selling 3079 on BMT Micro for about a month or so before trying to get on Desura, and that helped. Alex also was a big help, since he already had a game up on Desura. Having my game up on Desura has brought the game to many gamers, but, if i had to nitpick, I’m not a fan of the update authorization waiting period. I put out an update, and it takes about 32 hours for the Desura team to push it out. People who purchase through my website, BMT Micro, get the latest update immediately.

So 3079 was in development roughly 6 months before launch, correct?

I suppose that long before launch on desura, but I was selling it on BMT Micro at around 4 or 5 months

Erlend: and you’ve been holding a day job alongside your game development all this time?

Jeremy: I have, and raising a 4 month old son — I’ve been pretty busy!

Has the success of your game made you reconsider your employment status, or are you happy as a hobby developer?

Wwell, it depends on how things go in the mid to long term. I’m quite happy at my day job, and I enjoy the people I work with. I’d be pretty lonely without my job. If 3079 really continues to take off, we’ll see

So how long would do you reckon you’ve been actively developing games as a hobby?

Since middle school?.. 10-15 years, somewhere in that range. I started with QBasic.

Erlend: To be honest, as much as it scares me, that seems to be the average amount of time required for an indie developer to finally make a hit game  not meaning 10 years of development on one project, but 10 years of trial and error.

Jeremy: i never really thought of getting into commercial game development until about a year or so ago, when i started making android games. But yes, trial and error has much to do with it, and taking lots of constructive criticism to find out what people like even though you will never satisfy everyone!

Would you say your decision to look at games as more than a hobby, and rather a prospective income, helped you push 3079 to success?

Honestly, not really.., developing games has always been my hobby and even as I develop 3079 today, it still feels like a hobby. But it is one i take very seriously and enjoy. Having it sell and make money is really just icing on the cake

So what are your plans for the future as far as game development goes? judging by your personal website, you’re not a one-epic-project kind of guy. how long do you see yourself fully committed to 3079?

That is really difficult to say. I don’t really have a time or date set for when i will stop working with 3079. All of my previous games were fun to develop, but none of them were anything close to 3079, which still has so much potential.

So if all things 3079 were completely erased from existence today, would you already have one or more ideas you’d consider pursuing?

Yeah, i actually have another 3D game idea that I’d like to work on, but I’m currently on 3079. Alex, the PR guy working with me, also has an idea he’d love me to work on too, hehe.

I’ve made lots of 2D games, and now that I know how to make successful 3D games, I’d like to recreate some of those games in 3D

Erlend: with jMonkeyEngine?

Jeremy: of course. Seriously, jMonkeyEngine is where its at, and that is no lie. Cross platform compatibility out of the box, great performance and it is relatively easy to use. Community has been very helpful. Why would i want to go elsewhere?

Erlend: he he, much appreciated.

If you could improve on one thing in jME3 / jMPThe jMonkeyPlatform, a full-fledged IDE with plugins for editing jME3 content and assets. Integral part of the jME3 SDK., what would it be?

I remember when i first was getting started, I wanted more GUI options, and there was NiftyGUI. But I don’t think there was a way to develop the GUI inside the Java code. You had to make XML files. This might have changed by now, but I remember wishing I could just make NfityGUI elements via code.

Lastly, any word of advice for all the budding game developers in the jMonkeyEngine community and elsewhere?

Keep developing and take feedback from your supporters. And set goals. Like, are you trying to make your current game sell, or is it a learning project? It can be difficult to start a new project, wanting to learn how to make games, while at the same time making it polished and commercially viable although a learning project may turn into a commercial game, I suppose.

Also I’m surprised at how time consuming it is to have a popular project, responding to help requests, feature requests etc.

Alright, extra ninja question: What was the 3079 goal?

I wanted to make a 3D game i’d be proud of. I made 3069, a 3D game i wasn’t really proud of. It was decent, but I really wanted to do something better.

My new end goal for 3079 is to get on steam!

You might be interested in:

• 3079 on Desura
• 3079 Home Page
• 3079 Steam Fan Group

First Place

The Cove – @thetoucher

Score: 86%

Second Place

Monkey Mania – @ndebruyn

Score: 79%

Third Place

Pirate Hell – @ceiphren

Score: 73%

Complete reviews and scores for all of the entries can be found in the forum thread.  Our winners will be receiving the prizes outlined in the prize announcement.  @thetoucher can expect a free copy of the upcoming jMonkeyEngine Book in the mail as well as a free WordPress theme courtesy of our friends at  WooThemes.  @ndebruyn and @ceiphren will be offered substantial discounts on both the book and a WordPress theme.  Thank you to all for making this contest the huge success it was; Truly a community feat!

Now that the documentation is finally complete (thanks to Ian Mayo from PlanetMayo Ltd for funding this article), we’d love to have some more monkeys try it out.

A while back we announced a successful deal made with Packt Publishing, entrusting our very own Ruth Kusterer @zathras to author “jMonkeyEngine 3 Beginner’s Guide”. For the past year Ruth has been hard at work, masterfully balancing a tight schedule of wiki updates and chapter deadlines. Consider for a second how much jMonkeyEngine 3 has changed and improved over the course of this year, and you’ll have an idea of the moving target she’s had to deal with. Hopefully by the time the book is released, it will be referring to a jMonkeyEngine 3 gone stable.

While you wait, and as a means to help you decide whether you should buy this book, please go visit book.jmonkeyengine.com for an early preview of the book’s contents! We’ll preview a new chapter roughly every day until all 13 are online.

Is it on your wishlist? Whatever the answer, we’d appreciate your honest input.

Hello, my name is Andreas and I have been working on a paged grass & tree system for jME applications. The lib is based on another lib named  Paged Geometry, originally written for the Ogre3D SDK.

The system is designed for making tree and grass loading and rendering more efficient. The loading is optimized by dividing the geometry into blocks (pages). The paging engine then loads/unloads those blocks based on their distance from the camera. The rendering is optimized mostly through geometry-batching and a system of multiple detail-levels (reduces the visual quality of distant geometry).

The trees still needs a lot more work but the grass system is now ready for a test run, so I’ve chosen to release the grass only (rather then waiting for all of it to be finished).

Note: This is a brand new lib so there will be bugs, and some things may have to be changed to make it more user-friendly. Anyone who wants to try this out now should consider it beta-testing for some time (and hopefully be willing to submit bug reports and suggestions).

Grass

Pros & cons

This kind of system makes it easy to manage vegetation, both loading it, saving it, and controlling what to show and when to show it.

Adding grass procedurally is not very hard. The lib comes with a procedural grassloader and a standard planting algorithm, and it’s not hard to plug new algorithms in.

Adding grass manually is a bit more complicated. Density maps are good for that, of course, but the density maps has to be generated somehow. There will probably be some sort of editor plugin for placing grass (and later trees) onto terrain objects.

Usage

Here is a tutorial on how to get it running.

Here is a link to the official forum thread: http://jmonkeyengine.org/groups/user-code-projects/forum/topic/the-forester/

Here is a youtube video of the grass loader in action. It is available from the forum thread as well.

Grass System Beta Release Video

Note that this has not yet been tried with the terrain grid. I think there may be problems when trying to load grass onto a terrainquad that isn’t loaded yet. I’m aware of this, however, and will fix any problems.

Features

These are the basic features of the grassloader:

• Paged grass. The grass is divided into patches that are loaded/unloaded (and made visible) based on their distance from the camera.
• Impostoring. Each grass page contains two grass geometries – one for the grass up close, and one for grass in the distance. This makes it possible to use a simpler mesh and different shading (lower res textures etc.) on grass that is further away from the camera. There is fading between the two levels of detail; making the transition somewhat smooth and hard to notice. (kudos to johncl for suggesting this).
• Densitymaps. The grass can be placed using a grid of densitymaps, to make grass appear only in desired areas. The densitymaps are B&W images, making it possible to have 256 different levels of density, which is more then plenty.
• Colormaps. The grass colors can be modified using colormaps.
• Self-shadowing. Each patch of grass is slightly darker near the ground to simulate self shadowing. This is done with vertex colors, and does not require any texture manipulation. (johncl’s idea as well).
• Animation. The grass is animated to simulate wind. The direction and amplitude of the swaying is based on a wind vector that can be manipulated from within the grassloader, or provided from an outside source (such as a weather system). It is a fairly simple animation at this point, but it will be improved.
• Multiple grass layers. Each grass-type uses its own grass layer. A layer contains data specific to the grass type, such as quad-size, materials (textures), whether or not to use animation etc. This makes it possible to use several types of grass with the same grassloader. The densitymap/colormap system also allows for one densitymap and colormap per layer, making them completely independent of one another (except for the pagesize, but more on that later).

** The grassloader also comes with a random generator. It is based on the (fairly) new XORShift algorithm, and is very light and fast. It is well suited for producing the type of random numbers used by this system.

The future

It all depends on how the system is recieved.

Should there be an interest for this system it may be integrated with the jMPThe jMonkeyPlatform, a full-fledged IDE with plugins for editing jME3 content and assets. Integral part of the jME3 SDK.; making it possible to use the scene-composer for placing grass and trees. Hopefully I could merge with the new terrain system somehow, making that whole process a lot easier. Until then, the focus will be on bug-fixing and  of course finishing the trees.

Have fun. //Androlo

A week has passed since the Beta Game Contest was announced. It’s been thrilling to say the least, watching two or three new entries popping up every day. We’re quickly approaching a total of 20 entries, and that’s before we’ve even mentioned any prizes! But there will indeed be something of value to take away from this contest besides the enjoyment of making a game in a friendly competition, so here it is:

1st Place

• One (1) free hard copy of the upcoming “jMonkeyEngine Beginner’s Guide”, authored by our very own Ruth Kusterer.
• One (1) free theme of your choice from WooThemes, one of the most popular WordPress premium theme sellers.

2nd place

• A 40% discount off of the upcoming “jMonkeyEngine Beginner’s Guide”
• One (1) $10 coupon (i.e. a $60 discount) for a theme of your choice from WooThemes.

3rd place

• A 40% discount off of the upcoming “jMonkeyEngine Beginner’s Guide”
• One (1) $20 coupon (i.e. a $50 discount) for a theme of your choice from WooThemes.

About the sponsorships

More details regarding sponsorship of the jME3 Beginner’s book are TBA.

I’ve personally been a long time customer of WooThemes. I used one of their themes to create our showcase site at jmonkeyengine.com. My line of thinking for this “deal” was that WordPress makes for an excellent CMS for indies to effortlessly set up a website for their game, and WooThemes makes it look pretty!

We hope you’ll find these prizes to be worth your while. That said, we’re always looking for opportunities to up the ante: any leads you might have for us are most welcome.

Now get back to your game-making!

Update your jME3 SDK today! Simply click the update button on the lower right side of the jME3 SDK or select “check for updates” in the “Help” menu and install the updates, it is all automatic.

Here’s an overview of the new features:

• Better support of Blender 2.6 models through fixes in Ogre3D animation importing as well as tangent generation (if using normal mapping)
• More stability and better performance in the jME3 SDK
• Custom controls editing directly in the SDK SceneComposerA plugin in the jMonkeyEngine3 SDK, allows editing the content of j3o files which represent a complete scene or model. 
• Texture array support
• MonkeyZone now runs with 3.0 beta out of the box
• Raycasting now supported in native bullet physics
• Templates for NiftyGUI controls in the SDK
• Templates for AppStates and Controls in the SDK
• Play/stop physics right in the SDK
• Fix issues with mouse input in the SDK
• Fixes to AppStates, to make their usage more consistent
• Various fixes to timer, making multiplayer games much smoother
• Fix issue with jME3 native extraction not working in Windows under limited user account

Happy coding! ~jME team

This mini-game contest will have the aim of everyone having fun participating and making a good game to showcase jME’s capabilities. There won’t be any genre or restrictions for the mini-game, only that it shouldn’t be too complex for the player to understand and play. The contest will take place between the 7th of November 2011, and the 31st December 2011 (inclusive dates). Below, you can find additional information, including the rules. Any questions are welcomed. Have fun participating, and good luck!

Judges: @memonick, @namo, @sbook.

Prizes: To be announced.

Please see the official Contest Rules Thread for all of the info!

Good luck to all!
The jME Core Team

I’ve always been into procedurally generated game development. Interestingly enough, my first big procedurally generated game was 3059 (a 2D “ASCII art” roguelike game). I went on to create FreeUniverse (a space action/trader game), Gentrieve (a procedurally generated Super Metroid game) and DroidCraft (Minecraft-like game for Android). Wanting to get into 3D programming (and back to my futuristic “roguelike” roots), I started 3069. I had planned an open, block-like world full of monsters, quests and items well before the Minecraft-era hit. I decided to use the Ogre3D engine, which turned out to be more difficult than expected. In the end, 3069 did get completed, but it wasn’t what I originally hoped for.

3079′s Conception

While taking a break from desktop game development, I played some great games: Fallout 3, Just Cause 2 and Minecraft. I loved Fallout 3′s action and role-playing elements, Just Cause 2′s grappling hook and Minecraft’s open, block-like world. Still wanting to create a solid 3D game, I decided to make another sequel to 3059 and 3069: 3079. Now running Linux, I needed a cross-platform development environment and I wanted to try a new, fresh 3D engine designed for games. I decided on Java and did a simple Google search for “3d engine java” — jMonkeyEngine was right at the top. Little did I know at the time how well jMonkeyEngine would work out!

Just another Minecraft game?

As the developer of Mythruna can attest to, it is a challenge to bypass the knee-jerk reaction of “it’s just another Minecraft clone.” To differentiate my game from Minecraft, I decided to focus on character abilities, artificial intelligence, action and questing (things Minecraft lacks) instead of crafting and mining (which Minecraft excels in). I also wanted the game to start in the sky, showing off the limitless ceiling while at the same time avoiding the “you start in an open field” Minecraft-like start. I do all this because I want players to know 3079 is not just another Minecraft clone — it actually plays much more like Fallout 3. Nevertheless, I’m continuing to look into ways to distinguish my game further.

Development

Performace, performance and performance. I started 3079 with the goal of making the fastest and leanest 3D voxel engine possible. At first, my engine was slow and inefficient. I had a node and separate mesh for every block and I was trying to use dynamic lighting. Once I figured out how to use Texture Atlases, I combined 40x40x40 “blocks” into a single dynamically-generated “chunk” mesh using a single material. I later ditched dynamic lighting and “baked” lighting into my textures. I did this by having a large grid of block textures, each with 16 different shades of brightness. I then set the correct shaded texture (using UV mapping) when dynamically generating the “chunk” mesh. This improved performance quite a bit — leaving more room for artificial intelligence (AI) processing. Figuring out how to connect “chunk” edges, especially edges of chunks that haven’t been generated yet, was quite the challenge; using specifically calculated random number seeds proved to be very important. Lighting also was a challenge — I didn’t get it “quite right” until very late in development. It wasn’t until I got performance greatly under control using the Java profiler before getting soft, ray-traced lighting to work (lots of internal caching going on!). All in all, programming the AI was the most fun — because making characters “act as expected” is always an interesting challenge!

Conclusion

I am very happy with 3079. Of course, it isn’t “finished.” I gave it the “BETA” label, because I still want to explore things like networking, “the look” and more unique tools and locations. I believe I really nailed performance and memory consumption which should give me quite the wiggle room to add more to the game. If you are interested in playing 3079, you can get a free demo at http://3079.phr00t.com/.

My thoughts on jME3

Using jME3 was almost a breeze. Getting started was very easy after installing the jMonkeyEngine 3 SDK and downloading the first tutorial project. I especially liked how SimpleApplication included the flying camera and graphics configuration window — I could immediately get started on my game without having to re-invent the wheel here. The forum community was also very helpful and I greatly appreciate all the prompt help I received. Once I started using the nighly SVN (to get daily jME3 updates with Alpha-4), problems I had with the engine were fixed in only a day or two. Now that jME3 BETA is here, things are stable enough where I don’t need to use the nightly SVN.

jME3 was a whole package. It wasn’t just a library you add to your current NetBeans installation — it was a whole SDK There are a few plugins for the SDK that I used, like the model importer and desktop deployment tool. However, there are many others I didn’t get to — like the vehicle and terrain editors (which I’m sure many would find useful).

jME3 Pros:

+ Small learning curve: jMonkeyEngine 3 SDK + tutorial project (using SimpleApplication) = easiest start to a 3D engine ever!
+ Performance: jME3 takes care of view culling & scene management in an efficient manner
+ Java platform: cross-platform support!

jME3 Cons:

+ Not “stable” yet: jME3 is progressing very well, but isn’t fully mature yet. I had to create my own shader for proper fog effects because they haven’t been implemented yet, and some bugs were difficult to work around until they got fixed in nightly releases