Java games advocacy

General Discussions
#1

After reading through the Indies and Java and Example project discussion threads, I’ve been doing a little thinking. There are several active members in this community who are doing a lot of work in terms of building the tools we need to do what we all love with the language we all love (and occasionally hate as the case may be ;)). I have also seen some posts in other game programming forums by certain community members advocating Java for games (usually met with sneers, jeers, and rotten tomatoes).

There’s a lot that can be done to bring more developers into the community and to get our games out on the market. Little things that come to mind: pumping out some good freeware games or the previously suggested demos (to help get that JRE installed and show off what can be done), members of the community linking to each other’s sites (if my games use Java and your games use Java, our cross customers/downloaders will be more likely to be Java-able). But we need something on a larger scale.

I’m thinking a site that distributes freeware/shareware Java-only games might be in order. Java FAQs, a JRE installation tutorial, links to the sites of all developers who distribute through them, highlights of new Java titles, yadda, yadda. This sort of thing might go a long way to getting the JRE out to the right people (casual gamers or whatever you want to call them) while at the same time raising awareness in the game dev community.

Another large scale project that might be beneficial is a Java game dev site geared toward Indies. What we have here is great for what we use it for, but at the end of the day it’s just a message board and a code repository. What I’m thinking of is a site that is designed to not only provide resources (tutorials, code snippits, tools, reviews), but also to help counter the Myths and the FUD that exist (‘Java’s too slow for games!’), and Java advocacy in general. I know someone started a Wikki, but in my experience Wikkis tend to be incomplete (lots of empty categories), awkward to navigate, and generally only useful for someone who is really interested in the topic (and for that purpose they are great).

Both of these are things I think would be worth pursuing. Comments, thoughts?

P.S - I’d personally not have a problem with starting a java game dev site myself, and might do just that regardless of feedback :wink: Getting a company started to distribute Java only games though is something that is beyond me for the moment, and likely most of the community as well. I’m just wondering if any of you think that sort of thing would be as benedficial as I think it would be.

#2

Well, I’ve basically already offered as much as free hosting and exposure on Puppy Games recently but met with an incredibly underwhelming response :-/ ie. zero. Apart from whatever sneakret plans Onyx has afoot.

Cas :slight_smile:

#3

From the point of view of someone coming to java game development for the first time, right now, what we need the most is a clear and clean java game development kit/platform. Currently we have to really invent our platform from parts and pieces. Of course it is great that a lot of work is being done, but right now i can not go and download a Java Game SDK and start working on making a game. i have to go and download parts from this site… parts form other sites… and put them together to make what i think is an ok platform to start with.

As a developer, i really want a Java Game SDK (what i thought the now canceled JSR-134 would provide), and i think having such an SDK would help attract a lot more people to Java for game development

#4

This is precisely what we have attempted to achieve with LWJGL. It is all that is required to make high-performance games in Java, less the JDK. Half a meg of code loveliness.

Cas :slight_smile:

#5

[quote]From the point of view of someone coming to java game development for the first time, right now, what we need the most is a clear and clean java game development kit/platform. Currently we have to really invent our platform from parts and pieces. Of course it is great that a lot of work is being done, but right now i can not go and download a Java Game SDK and start working on making a game. i have to go and download parts from this site… parts form other sites… and put them together to make what i think is an ok platform to start with.
[/quote]
I’m intrigued - what bits do you need that you have to go to various places for?

I’m wondering what more you need in your “platform” than J2SDK (obviously ;)) + JOGL/JInput/JOAL (all in one place - this site) or an alternative of your choosing. I can think of lots of extra libraries that you might use - e.g. ones I’ve used in games dev, such as JavaMailAPI - but I’m not sure what’s wrong with their current distribution.

I thought that the main problems people had with java game development were nearly all to do with having to learn how to write games (technically), and getting round problems in the current SDK implementation from Sun (dodgy 2D API’s, dodgy Timers, no OpenGL, etc). I can see another big problem on the horizon - the discovery that writing games (non-technical aspects) is actually extremely difficult (e.g. developing gameplay is a far cry from implementing a Specifications document; e.g. the massive costs of soft-content; etc) - but AFAICS there are so few people around here who’ve got that far yet that these problems aren’t cropping up often…

IMHO the biggest thing newcomers need is more education - articles, tutorials, etc (hint, HINT, ChrisM et al!).

#6

The J* libraries aren’t a single download-and-you’re-done SDK though.

Writing games in Java is simply just nowhere near as easy as BlitzBasic. If anyone wants to learn how to write games, I tell them to use BlitzBasic. You learn how to write games - the theory’s the same - but you don’t have all the Java crap to worry about. Java is syntactically nasty, messy, fiddly, and difficult to read, and you need to know so many stupid tricks to get it to do the simplest things (gimme a fullscreen window! draw me a sprite that isn’t slow!). The only thing we did with LWJGL was throw out all the Java complexity and replace most of it with driver-level complexity :confused: The syntax is still nasty.

Cas :slight_smile:

#7

If it’s so nasty, why are you using it? :). The nasty horrid thing is opengl (directx is worse from the little i’ve seen), using a nice scenegraph that wraps it all up is the way to go, oh wait, we have/had one, java3d/xith3d, this removes a lot of the nastyness for you when it comes to writting games (on the graphics front), no it’s not perfect, but what is :), xith certainly come with prebuild binaries for jogl and joal (and I think jinput now), which means xith covers your single download, as for the comparison with BlitzBasic, i dunno, i never used it. The only graphics code I’ve done has been java or c++ (unless ascii characters made to look like a window environment count :))

/me sticks the oar in and stirs things up a little

Endolf

#8

All I can say is - download Blitz and try it out. You will be amazed at how easy it is to just start writing games.

I used Java because, at the end of the day, it’s more powerful and does other clever things. And because I started 3 years ago when Blitz was nowhere.

Now, I find OpenGL to be especially easy - I came from nowhere, with no experience, and had stuff running in GL in minutes, even in C. It’s roughly as complex as Graphics2D is, with another dimension when you want it. The only real difference is that it’s fast as hell and designed to be that way! It’s also an API that’ll grow with you. You’ll never come up against a scenegraph’s limitations.

The real problem is simply the framework in which all this exists, and the overcomplex ways of doing things. Look at this:


if mousebutton(0) and not spritevisible(1) then
sprite 1, x, y, zapanimation
spritevelocity 1, bulletspeed, 0
end if

and this:


Sprite sprite = new Sprite();
...
if (Mouse.isButtonDown(0) && !sprite.isVisible()) {
sprite.setLocation(x, y);
sprite.setAnimation(zapanimation);
sprite.setVelocity(bulletspeed, 0);
}

That’s a tiny trivial snippet but you know it’s only going to get harder to read and understand from then on…

Cas :slight_smile:

#9

Hi
gl itself in terms of the commands are easy, it’s just some things are a PITA to do, take something as simple as a transparent object, oh wait, now I have to do two rendering passes, depth sort my transparent objects myself, enable and disable the zbufferwrite (or what ever it was called) and draw my transparent objects in reverse depth order. Hmm, or I could just throw them in my scenegraph and drool at the output :slight_smile:

I agree that some things are harder than they should be, like creating a full screen window including changing resolution, but it’s a small price to pay :slight_smile:

Endolf

P.S. are you taking up a new career, I just noticed the holiday place link in your siggy :stuck_out_tongue:

#10

Nah that’s just down the road from where I live :smiley:
Actually I went for a job interview the other day to do .NET stuff. Somebody save me!!! But I’m broke as hell thanks to all this game dev biz and I need to get something sorted or there will be Trouble.

Cas :slight_smile:

#11

You think


if mousebutton(0) and not spritevisible(1) then 
sprite 1, x, y, zapanimation 
spritevelocity 1, bulletspeed, 0 
end if

is nicer than this


Sprite sprite = new Sprite(); 
... 
if (Mouse.isButtonDown(0) && !sprite.isVisible()) { 
sprite.setLocation(x, y); 
sprite.setAnimation(zapanimation); 
sprite.setVelocity(bulletspeed, 0); 
}

Are you completely loopy!?

Kev

#12

[quote]The J* libraries aren’t a single download-and-you’re-done SDK though.
[/quote]
I assumed that getting everything from one site, all linked from one index, would be good enough (which is the case with the J* libs). Equally, I thought LWJGL would meet the basic requirements?

Ah. I didn’t appreciate the problem was “ease of programming”; I assumed the poster wanted to write games the hard way, not have someone create a spoon-feeding IDE for games development (Nb I don’t use that term derogatively; spoon-feeding is great - but it’s extremely expensive in terms of resources to build). If that’s what you want, I agree that individuals normally don’t have the resources to make a game that goes beyond the limits of what BB (and others…e.g. Dark? [not looked at DB for a long time, though…]) support, so should probably just skip java.

I think creating such a dev environment is way beyond what any java organizations, or groups of individuals, are capable of, so that you end up in exactly the situation the poster described - lots of different bits offered by lots of different groups, because the “bits” are small enough that groups can redefine / improve them (e.g. Xith).

I don’t think you can even ask for something better; there are few companies (Microsoft, perhaps) who could afford to even try this - but even MS would only get it 60% right, going by past evidence (they’ve put huge amounts of resources into attempting to make things easy to use, but even with 2000-strong programmer teams they routinely fail on basic points :(. Even if you subscribe to the theory that their dev teams are often incompetent, you have to admin what they try is really hard).

Blitz, Dark, etc, manage it by offering a massively cut-down expressiveness. Sure you can make quake in Blitz these days, and quite easily AFAICS - but you probably wouldn’t manage to make HL2 in it (or perhaps you could? I know little about what is hard in Blitz…). Nor could you do even a simple MMOG, just because you need far more expressive power than they provide. Even simplifying relatively small things like gameserver development is massively expensive (it’s something that the company I work for specialises in) and enough to keep a large number of developers, architects, API designers etc busy for many years (e.g. Zona has blown several million dollars on it already, and not got all that far IIRC). IMHO you’re looking at many many millions of dollars to do the whole game-dev environment, without loss of expressive power :(.

#13

hehe, I’m with kev… I prefer the Java code above.

#14

Well, so do I, because I totally understand it. But if I just wanted to bash out a prototype idea… Java makes me type 3x as much in the short term. And indeed, probably in the long term, for all it’s clever OOP goodness.

The OP of course might like to take note that outside of the Java world the situation is pretty much the same…

Cas :slight_smile:

#15

Well Cas i don’t use LWJGL, mostly because after looking at the SUN presentation at JavaOne and seeing that it had pointers exposed it sort of turned me off. I saw some talk in your forums about removing those and i couldn’t find any status on that front. Not trying to get into a debate here… i just chose the technologies from the games.dev.java.net project because it looked like the right way to go at the time.

Now back to what i was saying about a Java Game SDK.

Right now on this site we have a few different binary files that include different technologies for helping you access the hardware on your system to write things like games. That is basically it. In my mind that is not a Game SDK. What i want (and what you get when you want to write games for other platforms) is an SDK with the following:

1 single SDK file/installer with everything you need

/include and /lib : directory with all the binaries we have today (Jogl,Joal,JInput)
/bin : ?? useful tools for game development
/api-docs : with all the javadoc we have today
/documentation : ?? nothing here yet, need to fill this up with useful info
/tutorials : Put the NeHe ports here as well as some of the other tutorials people have/are working on
/examples : ?? need some example games here

Basically i want to be able to go to games.dev.java.net site and on “Java Game SDK” for the OS i’m going to do my development on. Extract the file on my system and i should be able to start running the examples, trying some of the tools, read the documentation, do some tutorials, and basically be ready to start working on a game.

Here i’m not talking about an IDE, or a game “engine” (although both could be added by the community). I’m talking about a single file that, for right now, contains the stuff we currently have. As time goes on, everyone in the community can help add to each of the above locations with more docs, tools, tutorials, and examples. Over time the SDK will turn into a useful kit for people that want to write Java Games. Just like the Java SDK, we need a single SDK for Java Game Development.

Hopefully i have explained this clear enough. If you still need more help understanding, please go download the DirectX SDK or go to Tapwave.com and sign up to download their SDK and you’ll start to understand what i’m talking about.

#16

That’s precisely where we’re headed with LWJGL 1.0. Exactly.

Cas :slight_smile:

ps. There have been no pointers in LWJGL for quite a while now, and on the whole, it’s a cleaner design.

#17

OK Cas, i’ll go take a look at it right now then. It was not clearn to me that things have changed. If LWJGL provides an actual SDK rather then just a few jars/dlls that let you access the metal then i’m all for it.

#18

That’s the raw material I was thinking of - mainly documentation, tutorials, intro material, quickstart guides, etc etc.

Packaging it together is a nice flourish, but IMHO we’ve still not yet achieved the first step of trying to get some kind of official momentum going: backing, encouragement, promotion etc, for the raw material (mainly docs, tutorials, etc - others have already done a great job of getting the libraries created and putting Sun behind them in an official way; I’d like to see JGO come out of the closet and publish docs in addition to binaries/source).

#19

Yeah, uh, we haven’t quite gotten to 1.0 yet :wink: Still in alpha at 0.7…

Cas :slight_smile: