The Gang Garrison 2 Forum
Off-Topic Discussion => Amateur Gamedev => Topic started by: MedO on June 07, 2009, 04:08:35 pm
-
I've been meaning to start this topic for a while now, and StealthArcher's own thread has finally made me conquer my lazyness. This topic will provide information about my progress with JGang and plans for what to do next. If you have ideas, criticism or want to discuss anything else about JGang, feel free to post here as well.
So what is JGang?
In short, it's a game engine for networked 2d platformer games, written in Java.
JGang is the second incarnation of the Gang Game Engine, which was originally planned and started as a C++ engine by cspotcode and me, as an engine that could be used for GG2 or similar games like Soldat, Liero or Worms. Since cspotcode didn't have much time to work on the project after we had done a some of planning and started the C++ code, I worked on it alone for a while and finally decided to start over in Java, since I have a lot more experience with the language.
Current progress
The most polished system right now is the resource management, which manages sound and image files and also allows to define Entities (game objects) in XML files. It provides facilities for easy and powerful modding and also allows you to include new graphics, sounds and entities in user maps. Some of these features are not fully done yet though.
There are also a few classes for graphics already, and currently I'm working on input handling (e.g. button mapping).
If you want to see something right now, have a look here: http://www.uni-koblenz.de/~smaxein/JGangTest.zip (http://www.uni-koblenz.de/~smaxein/JGangTest.zip)
You will need a recent version of Java to run it, and you might only see a slideshow if you have an integrated graphics chip. Have a look into the resources directory and play around with the xml file a bit if you want to get an idea of what you could change in a mod/map without any programming.
-
So is the game going to become more difficult to run?
-
Is double-clicking on a jar- or jnlp-file more difficult than double-clicking on an exe file? Besides that difference, the only thing that would become more difficult is that you'd have to install Java first.
-
Just out of curiousity, will the C++ engine still be finished?
This engine seems pretty cool. No more need for Game Maker Pro to mod the game. How will the actual game run though, if you see the JGangTest as a slideshow? That's what I'm seeing right now, with the image changing every second or so.
-
The C++ engine was not very far when I stopped, and I think it's unlikely that it will be continued. There is nothing to stop you from doing it yourself, it's open source and hosted on Launchpad (https://launchpad.net/ganggameengine).
If you only see a slideshow then Java2d doesn't use hardware accelleration on your system. You could try updating to the latest Java release and see if that helps, and there are also a few command line options one could test, but if you are using an older or integrated graphics card then chances are that things won't improve.
I'm aware that this is a problem and I might switch to directly using OpenGL, but right now there are a few points that keep me from doing that:
- Java2d hardware accelleration IS improving, so there's a realistic chance that it will work on more machines once JGang is actually usable
- Having to code my own OpenGL graphics drawing stuff would be quite a bit of work
- In the future, more people will have graphics cards which support the required functions
So I'm focusing on other parts of the engine first, and when those are finished I'll re-evaluate the graphics question. Just for comparison, on my PC (GeForce 8500 GT, Intel Core 2 Duo 2,66Ghz) JGangTest runs perfectly fluid and uses ~3% CPU.
-
Before anyone says JAVA = SLOW
its bull,
the only thing that would make java slow is if the persons computer was crap,
even runescape can run in high detail, and if you have a decent computer, that actually can run pretty fast.
Java or C++ would be the best course for gg
Java also being opensource, and availible on almost every OS imaginable.
-
Is double-clicking on a jar- or jnlp-file more difficult than double-clicking on an exe file? Besides that difference, the only thing that would become more difficult is that you'd have to install Java first.
I meant computer wise.
-
To be honest; I was more excited for the original C++ version than I was for this. It's not that I really think that java is unusable, but to me it's a lot less accessible.
.exe Is still easier to just pick up and go. It feels somehow more complete.
... Still, a more competent engine is a more competent engine.
-
Nothing but a slide show; And I updated java.
If it helps, I have integrated graphics, and so do a majority of the people that play this game as a substitute for TF2. Most people don't update java but it's usually by default on most prebuilts, so there is a higher technical curve here that might not be friendly for the userbase.
Additionally, you may want to consider serious gamer for the project. He's somewhat of a odd choice having no proper expertise on gg2 besides modifying it for cheats, but I believe he has some java expertise.
-
I did explain my thoughts about the graphics issue above, and my plan is definitely that it should run in a usable way on anything that's at least as powerful as a second-generation netbook (i.e. Intel Atom N270, Intel integrated graphics). But as I said, the issue is improving without me having to to do anything. I just tried a nightly build of Java 7 and with that the JGangTest runs well even on my EeePC 901.
-
Progress today:
I read up on how AWT and Swing (Java's GUI toolkits) draw their components and decided to make some changes to be able to use Swing's automatic double buffering, and to allow games to use Swing components to build their menus.
Also read an overview of Java's security model. This will become quite important later, because I want to allow mods and maps to ship with custom code, and since those will be auto-downloaded when people join a game on some unknown server, it must be ensured that this code can't do anything bad. Fortunately, Java supports this kind of stuff very well: You can set up a SecurityManager that is used to check whether the code is allowed to perform potentially dangerous operations (opening files or network connections etc.)
-
Make it playable in Nokia 6131 and then we may talk.
-
Make it playable in Nokia 6131 and then we may talk.
I lol'd. Let me clear yout his.
He is using PC java which is J2SE and java enabled phones are using J2ME 8D
-
Make it playable in Nokia 6131 and then we may talk.
I lol'd. Let me clear yout his.
He is using PC java which is J2SE and java enabled phones are using J2ME 8D
I know.
But, what I mean is that I'm avaiable to test if he makes for mobile phone.
-
Is there a way to help you with this?
That test runs slower than tf2 for me D8. You should definitely use OpenGL or another library for the graphics.
-
Is there a way to help you with this?
At the moment I'm not sure yet which license to use for this project. Since I'm putting quite a bit of work into this, I wanted to keep the option to try and make some money off it, and as soon as I allow other people to contribute some awkward problems will pop up. If contributions were made under a license like GPL or CC-BY-NC, I couldn't try to sell it later. Accepting contributions under a license that allows me to sell it later without sharing profit seems unfair. The only other solution is to create an agreement with all contributors about how the profits will be shared, which seems tedious considering that I'm not even sure there's any profit to be made.
Before anyone starts raging at me now, let me make one point clear: This engine will always be completely free (as in beer) and open-source for noncommercial projects. The only thing I plan is asking for a bit of money from people who want to make a profit with their game, because I feel I'd deserve some share.
That test runs slower than tf2 for me D8. You should definitely use OpenGL or another library for the graphics.
Please read the above discusison of the issue.
-
I'm still not quite happy that we're moving over to java.
... Question: How much of our current game will be in this version? Will it be a almost direct port or are we going to see something completely different?
-
I'm still not quite happy that we're moving over to java.
... Question: How much of our current game will be in this version? Will it be a almost direct port or are we going to see something completely different?
Well, if it's any consolation, I'm making a C++ engine >.> I'm assuming that this will be done long before origin, thanks to the fact he started earlier, and is probably more experienced as a programmer in general. We could probably pick after Origin is done, or just use this (I dunno if picking would work as most non-"in the know" people would pick mine solely off the mention of the word "java" given this topic's reactions >.>). I don't mind the Java part anymore since they finally listened to my 34 complaints and fixed it for the W7 RC.
-
Dr. Random: I'm not working on porting GG2 right now, even though the package name of the example game might suggest otherwise :hehe:.
If and when and to which engine it should be ported can be decided later, when we have at least one that's complete. And a little competition might make us both try a bit harder to deliver something worthwile.
-
Doc Random, I suggest you be happy someone's working on this. I agree with Medo, competition may turn out to be good. We'll see which is better for this purpose. However, that would not make the other a waste of time, because it would be useful for other purposes.
-
To be honest; I was more excited for the original C++ version than I was for this. It's not that I really think that java is unusable, but to me it's a lot less accessible.
.exe Is still easier to just pick up and go. It feels somehow more complete.
... Still, a more competent engine is a more competent engine.
You should have Java installed anyway. To me, it's just one of those things that you install, like .NET, DirectX, or the Flash plugin. Exe only "feels" more complete because most people don't know better. Jar is just as usable.
I would point you to Cultris, but their site seems to have disappeared. It was a very fast, online multiplayer tetris clone that had hardware-accelerated graphics (with plenty of eye candy), was coded in Java, and was released as a single EXE (not a JAR). So Java can be extremely accessible.
Java also has something called Java Web Start. With JWS, you click a link on a website, hit yes on a security popup, and the program runs automatically. I don't know if Medo wants to use JWS, but I thought it was worth mentioning anyway.
-
The new engine could get rid of hacks/cheats, but I dunno...it seems to lag on a lot of computers, and, when added to the fact that, in a real situation, there's lots of client-client-client...-server lag, I think it could crash a few computers. On the bright side, the physics are excellent! In fact, they could even rival Source!....in 2D...
-
Max, the only people that can use Hacks in the GM7 format are the server hosts. Otherwise, you get a lot of stuf that doesn't sync with the server, which in turn makes whatever you did useless unless you're hosting.
-
Added parallax scrolling back- and foregrounds to the rooms. See here (http://www.uni-koblenz.de/~smaxein/JGangTest2.zip) for some "huge robot in the sky" fun. Other notable changes
- you can now resize the window and it works properly
- control some random block with your arrow keys (don't go too far in any direction though, because the room is pretty small and your block will freeze if it leaves the room)
Next up, the camera/coordinate transform system will get a change to make it easier to draw pixel graphics exactly 1:1 (or some other integer ratio).
-
Here is the replacement for anyone who feels bored with that weird and terrible ball.
(http://i565.photobucket.com/albums/ss98/gufu1992b/Ball.png)
Just save the image in the resources folder.
-
Today I started to create a Character class for an example game and noticed that Box2D has some drawbacks for platformer development. Some things which are standard platformer fare are actually pretty hard to implement with it, for example: flipping or scaling shapes (To have your character look the other way), or one-sided platforms (those that you can jump through from below).
Since GG2 only uses rectangular collision boxes and doesn't support one-sided platforms either (last I know at least), this is not a problem if we want to port it, but other projects using JGang should not be forced to use that particular physics engine for simulating their entity movement. I'll change the current implementation so that JBox2D is not hardwired in (I'll continue to use it in my example though).
I'm also beginning to plan what the first release should look like. It should be possible to create simple games with it already, but networking and sound will be missing at first.
-
Seriously people, this test is fantastic. The phisics of this game are much more realistic than the GG2 ones, it's much more smooth, and round objects have round collision boxes 8D.
I hope the final test is going well. I'm looking forward to this (and when I do that, I really mean it). :yaymoney:
-
Lots of changes under the hood, but nothing new to see yet. Since the changes are mostly technical, the description below is probably unintelligable for anyone who isn't a Java programmer.
You can now use other physics engines than JBox2D (or other, simpler ways for moving your objects around). The existing JBox2D integration has been put into its own Java package as one possible option.
The XML resource files now work differently. You can now put any object into a resource file as long as its class has a Constructor that takes a DOM Element. In order for this to work, you now have to give the qualified name of every resource object's class in the .xml file.
Before:
<SimpleBox2dEntity name="box" appearance="boxImage">
<shape ref="boxShape"/>
</SimpleBox2dEntity>
Now:
<object class="de.uni_koblenz.smaxein.platonic.box2d.SimpleBox2dEntity$Factory" name="box" appearance="boxImage">
<shape ref="boxShape"/>
</object>
This is a bit more verbose, but also more flexible since tag names don't have to be mapped to resource object types anymore. It also shows a bit more clearly what's actually going on. In this example, you can see now that the described object is actually a "factory" object that can be used to create entity instances, instead of being an entity instance itself.
This change is also the first step for allowing mods and maps to bring their own Entity classes, since those can now be directly referenced from the xml file.
-
Game Maker is slower than Java, and Java can be used on any system, and in web browsers. Not to mention that if you don't compile to bytecode or whatever I'm thinking of, you can search for strings of code in the .jar.
I had to convert GG2's .gmk to .gm6 to view search for a spiece of code :cry:
-
Actually, You could do the SUN thing and make it LGPL
Its what open office is under, and open office is also sold commercially as star office, with proprietary stuff added.
You would have the advantages of open source, but the but the option of a commercial venture,
even though you could make money with it being gpl, it would probably be support, or selling compiled games.
the alternative route would be go the ID software way and not opensource it until you cant make money with it anymore, so you just release the source to the community without any media, so people still have to make tiles and 3d modules, and add the parts of the engine that were removed when it was open sourced.
-
Another option is dual-licensing under the GPL for open-source projects and a commercial license for proprietary projects, like Qt is. People like the KDE or VLC folks use it for free, and companies like Opera Software pay for a license.
But these options don't answer Medo's question: If he accepts code contributions, should he/how does he pay contributors if and when he makes any money.
In the case of Qt, I think contributors just don't get any money. Which makes sense, because I'm betting Qt software does way more work than the contributors do.
-
Thanks for the suggestions. GPL might be a feasible option for dual-licensing, but I have come to dislike it somewhat because of its viral character. I don't feel it necessary that everyone who uses the engine (for free games) should have to open their own source code. The LGPL would be an option for going completely free, but no use for dual-licensing because everyone could use the engine for commercial projects as well with that license. That's why I'd currently favor something like the Creative Commons by-nc license.
A small announcement about development, I have decided to stop until I have finished a project for university. It is crucial for the timing of my exams that I get that done soon, afterward I will resume work on JGang. It should take at most four weeks.
To while away the time, I should probably post a link to the most recent test where you can run around as Quote and push those blocks and balls around. This uses a new generic Character class which is supposed implement generic movement and collision behaviour/information for player characters. It is incomplete so far, for example, there's no way to check whether the character is standing on something at the moment. As a result, you can jump while in the air and catapult yourself out of the room.
http://userpages.uni-koblenz.de/~smaxein/JGangTest3.zip (http://userpages.uni-koblenz.de/~smaxein/JGangTest3.zip)
-
This is actually getting fairly interesting, and now that I have dedicated video, it doesn't move like a dying elephant.
The moving actually seems fairly smooth, but could be toned down for a bit more controllability.
However it seems everything is coming out fine. Especially those physics on the squares and balls. And I simply adore the moving background.
-
You could always just say that contributors will be credited for their help... but nothing more.. and if someone isn't willing to live with that they can just not contribute. No need to pay out for it all. I'd enjoy just being able to point out to someone else that pays a game that "Hey, I helped make that"
-
Any progress on this? Would be really nice to know since it sounds/looks/tastes interesting.
if not then i will hit you :hehe:
-
No, I'm currently using my hobby coding time for GG2 bugfixing. I'll definitely continue with this though.
-
No, I'm currently using my hobby coding time for GG2 bugfixing. I'll definitely continue with this though.
Yes please! I absolutely love the physics engine!
[/possibility of "why don't you marry it" joke]
-
This currently is full of win, it works better too.
-
So far, so good. One question I have though: Will the finished game be as heavy on physics? Having several objects rolling around the screen could result in some lag, and probably desync.
-
Good day sirs, ladies.
I hereby declare this project un-shelved.
The university project I had been talking about is still not finished because I had to stop working on it in turn in order to learn for my exams, but at the moment I have a bit of spare time again and imma use some of it on this. And some on Dragon Age :angel:.
So far, so good. One question I have though: Will the finished game be as heavy on physics? Having several objects rolling around the screen could result in some lag, and probably desync.
JGang is an engine/framework, designed to be useful for different types of games. If we decided port GG2 over it wouldn't automatically become a physics-fest.
-
:wuv:
But... I can't run it :angry:
-
I tried it out and it ran pretty smoothly.
-
You mean really smoothly.
-
Networking support 1: Basic Lobby system
As a first step toward networking support I'm writing some generic lobby server/client code. Much of the server code from GG2 can probably be kept, but the system will be expanded a bit to allow game-specific server properties and server-side filtering options.
This doesn't make for the world's best lobby system (for example, it doesn't have any social component like "friends", and it doesn't feature matchmaking/NAT punchthrough), but it's a first step and can be expanded later. Fortunately, the lobby functionality can be very well separated from the rest of the networking, so switching to a more powerful lobby system later should be a painless process for game developers.
A downside of the current GG2 lobby system is that the server code is a Java application that directly communicates via TCP and UDP, so it's not possible to host it on a cheap webspace with PHP and mysql. I'm currently considering to make the server interface HTTP-based instead, e.g. as a web service, which would allow implementing an alternative PHP solution. However, I'm not sure that the added complexity and communication overhead is worth it.
-
Medo, this is completely unrelated, but what the hell is NAT?
I've heard it everywhere and supposedly it's opened on my computer, but I have no idea what it is.
I suppose on topic note: At least this is alive. Is it possible to ever get a friends sort of service for gang garrison in the future? That would be an interesting feature.
-
http://lmgtfy.com/?q=NAT (http://lmgtfy.com/?q=NAT)
Is it possible to ever get a friends sort of service for gang garrison in the future?
Possible, sure. But I don't think I'll write it.
-
Networking support 1: Basic Lobby system
As a first step toward networking support I'm writing some generic lobby server/client code. Much of the server code from GG2 can probably be kept, but the system will be expanded a bit to allow game-specific server properties and server-side filtering options.
This doesn't make for the world's best lobby system (for example, it doesn't have any social component like "friends", and it doesn't feature matchmaking/NAT punchthrough), but it's a first step and can be expanded later. Fortunately, the lobby functionality can be very well separated from the rest of the networking, so switching to a more powerful lobby system later should be a painless process for game developers.
A downside of the current GG2 lobby system is that the server code is a Java application that directly communicates via TCP and UDP, so it's not possible to host it on a cheap webspace with PHP and mysql. I'm currently considering to make the server interface HTTP-based instead, e.g. as a web service, which would allow implementing an alternative PHP solution. However, I'm not sure that the added complexity and communication overhead is worth it.
(http://knowyourmeme.com/i/21018/original/HL2GordonFreemanfsjal.png?1254091976)
ohboyohboyohboyohboyohboyohboyohboyohboyohboyohbo y
-
This sounds pretty cool.
-
Networking support 1: Basic Lobby system
As a first step toward networking support I'm writing some generic lobby server/client code. Much of the server code from GG2 can probably be kept, but the system will be expanded a bit to allow game-specific server properties and server-side filtering options.
This doesn't make for the world's best lobby system (for example, it doesn't have any social component like "friends", and it doesn't feature matchmaking/NAT punchthrough), but it's a first step and can be expanded later. Fortunately, the lobby functionality can be very well separated from the rest of the networking, so switching to a more powerful lobby system later should be a painless process for game developers.
A downside of the current GG2 lobby system is that the server code is a Java application that directly communicates via TCP and UDP, so it's not possible to host it on a cheap webspace with PHP and mysql. I'm currently considering to make the server interface HTTP-based instead, e.g. as a web service, which would allow implementing an alternative PHP solution. However, I'm not sure that the added complexity and communication overhead is worth it.
(http://knowyourmeme.com/i/21018/original/HL2GordonFreemanfsjal.png?1254091976)
ohboyohboyohboyohboyohboyohboyohboyohboyohboyohbo y
+agree
-
> December 09, 2009
> February 08, 2010
lopl
-
> December 09, 2009
> February 08, 2010
lopl
1. D'OH!
2. D: :cry:
-
If you do decide to go with the "PHP" solution, I'd be more than happy to help. I don't think it would be that complex - in fact, IMO it remain simple and might even be more robust.
-
The problem with a PHP based lobby is that it doesn't fit well into the current GG2 implementation, and since a GG2 port to any other platform still seems a long way off it wouldn't be useful to anyone right now. In the GM version, the server would have to send an HTTP request to the lobby every now and then, which apparently can't be done in a nonblocking way with 39dll. That's why I went for UDP originally, you can send UDP packets without blocking so the server won't get a tiny freeze twice a minute.
Also generally, sorry for the lack of progress. To prevent disappointments I'll stick to posting about stuff I did instead of stuff I plan to do for now. One new thing I made so far is a component that can read a Garrison Builder level and provide access to its background image, instance definitions and walkmask.
-
java? does that mean I can put it on my web site?
-
java? does that mean I can put it on my web site?
If MedO chooses to use the web Java service. That's what I've been able to gather from reading the thread, I'm not entirely sure.
I think he would have to program a bit more to have it compatible, but it's not impossible.
-
Are you talking about a Java Applet?
Yeah, there could be some security complications (applets are restricted from doing certain things for security) but I don't think it would be that much of a hassle.
Your visitors would have to agree to a security warning though (twice) I believe.
-
why is this in playable projects
I know this topic is 120 days old, but still
-
Dunno, maybe because I released a prototype at one point.
-
Dunno, maybe because I released a prototype at one point.
That's odd, because that alone shouldn't qualify it.
Not like it matters.
-
Dunno, maybe because I released a prototype at one point.
That's odd, because that alone shouldn't qualify it.
Not like it matters.
Playable Projects ==> Ability to play project
Prototype ==> Semi-Playable to Playable build ==> Playable Project
-
I've always taken 'playable projects' to mean that the base of the project is done and that a good portion of the features are present.
-
I've always taken 'playable projects' to mean that the base of the project is done and that a good portion of the features are present.
so the subforum is supposed to be empty?
-
No, soviet undead, front line strike, and at least two or three other earned and deserve their spot in Playable Projects.
-
Dunno, maybe because I released a prototype at one point.
That's odd, because that alone shouldn't qualify it.
Not like it matters.
Playable Projects ==> Ability to play project
Prototype ==> Semi-Playable to Playable build ==> Playable Project
Wait, then how come half the stuff in new projects isn't here.
-
Because getting into playable projects is a thing you have to ask for. Bassie doesn't have time to watch every project's development.
-
Because getting into playable projects is a thing you have to ask for. Bassie doesn't have time to watch every project's development.
ooooic
-
Because getting into playable projects is a thing you have to ask for. Bassie doesn't have time to watch every project's development.
This is correct, if you want to get your game in playable projects just pm me! I mostly check recent posts to see if there is any douchebagging going on, however the past months have been douche-free :c1: