Java4K 2013!

[davedes]

Wouldn't the boilerplate be pretty significant with OpenAL/OpenGL? Unless you are using immediate mode, I guess...

[princec]

Can't say that there really is a vast amount of boilerplate... and of course the good thing about boilerplate is just how well it compresses with Pack200 / gzip.

Cas

[pjt33]

I vote for Java 7 too, solely because I've gotten used to some of the new syntax, like generics type inference:

1	Map<String, List<String>> myMap = new HashMap<>();

You can use those and still produce Java 6 class files with the -target option.

Games published by our own members! Check 'em out!

[bainit]

cool:)
so many ideas, so little time..., so little space...!

[appel]

Yes, LWJGL has quite a few useful thingies that would make life for 4K developers nice, even without using OpenGL. For example OpenAL = reliable sound with very few API calls (and hence, probably less bytes, provided you generate your sound algorithmically). And the game controller stuff and other IO APIs, again, nice and simple. Not to mention stuff like main loops and display setup are all trivial in LWJGL. Food for thought. Maybe could convince appel to try it as an experiment for this anniversary year?

Cas

Could be possible.

There are always crybabies that won't have it and prefer pure java 1.4 still (you know who you are  ).

Only problem is it might be a slippery slope, next year we'll allow slick, jme, jogl, and then we'll lift the external dependencies limit altogether as long as the application code stays 4kb or less...and then we'll have people making 4k libraries just for the 4k contest that helps them to cheat in implementing their specific game.

But if there's enough support to allow lwjgl to compete on equal footing as java2d/jre-only then I'm happy to add it.

Or we could split it up into two side by side contests, java2d 4k and lwjgl 4k.


One practical thing to consider is hosting lwjgl libraries on a cdn/cloud, not sure if my datalink allows for thousands of lwjgl downloads per day.

[princec]

Only problem is it might be a slippery slope, next year we'll allow slick, jme, jogl, and then we'll lift the external dependencies limit altogether as long as the application code stays 4kb or less...and then we'll have people making 4k libraries just for the 4k contest that helps them to cheat in implementing their specific game.

On the other hand, lwjgl.jar contains purely wrapper functionality and minimalism - no actual utility classes. Slick and jme are game frameworks. I assume JogAmp has a jar with just the bindings in and not the utilities somewhere and there'd be no reason not to allow it. Even SWT should be allowed if you're some sort of masochist. The idea is that libraries which contain code other than APIs to existing OS functions could be generally disallowed.

Cas

[moogie]

I am more for leaving it as the pure java spec as that means that one order of complexity is removed and any incompatibilities from third party libraries too.

If we would allow other "API" libraries then in order for entrants to compete they will need to be across all libraries available. They will need to be able to know that api A can perform function F with less byte code than api B etc.

However I am not vehemently against the idea as long as the core of the competition is kept... 4k jar, no access to remote resources.

[princec]

It'd be purely optional... only reason I suggested it is because I know absolutely nothing about Java2D and lots about OpenGL.

Cas

[davedes]

It'd be purely optional... only reason I suggested it is because I know absolutely nothing about Java2D and lots about OpenGL.

Cas

Also... doing 3D would be a lot easier than making your own rasterizer.

Maybe a Java4K and Java4K-GL would be suitable...?

[moogie]

The only problem that I can see being an issue is that having two streams (java4k and java4k-gl) is that the entries would be split between the two so there may not be a large variety, turn out for each stream individually... but maybe that is not a big deal...

And of course the bitching about what GL library would be authorized for use in the java4k-gl stream

Games published by our own members! Check 'em out!

[namrog84]

Maybe we should have a vote onto allow LWJGL as a permitable (3rd party that doesn't count against the 4k)??


I personally would be okay to having others use it, but probably not utilize it though. Except maybe for a more smoother and consistent timer/render loop.

[Groboclown]

I don't have any experience with LWJGL, so I'm curious as to how much overhead it would add to an Applet versus using the straight Java2D libs.  I'm thinking about the constant pool size additional requirements for the class names and method signatures.  I know that's a subjective question, as LWJGL offers features not in the Java2D libs.

My question is really about whether users of LWJGL would have an "unfair advantage" over those who don't use it.  That itself is also a subjective question, because just by using a library doesn't make a game fun.

All in all, though, I'm in favor of expanding the number of games as long as it doesn't have an impact on those who don't use the library, or on the work load for the judges.

[ra4king]

It won't have much of an impact. You're trading code size for power if you use LWJGL. I'm in favor of including LWJGL as an optional lib for Java4K.

[Jimmt]

I'm all for adding LWJGL. At under 4K, the difference between LWJGL and Java2D is not too extreme.

[princec]

Well, no harm in somebody* just trying it is there?

Cas

* Not me, I'm too busy working for a living etc

[Roquen]

Since RNG related topics are semi-faddish ATM, here's a small one:

1	rng = rng*0xB5262C85;  // just don't have initial seed of zero :)

[Alan_W]

Last year I warmed over my Sharpshooter game for LWJGL16k (which didn't happen again), and personally I think 16k is a better size for doing OpenGL as it allows more room for data.  If Prince ever gets the LWJGL16k going (maybe in another 5 years), I'll finish my second entry .

It might be possible to do a basic 3D game in 4k, but there wouldn't be room for fonts, so no text.  For 2D, I'd rather use Java2D.  Does LWJGL applet performance still suck on OSX since Oracle took over the JVM and ditched support for [insert whatever the interface was called - I'm too lazy to look it up]?  That would definitely swing it to Java2D for me.

Hopefully (No LWJGL16k distraction this year), I'll get a 4k on the blocks.

Incidentally, if anyone wants to do a Java1K, I have done Pong in 1K just for laughs.

Edit: Having had a look over at LWJGL.com, I think JDK 7 on OSX is a work in progress. I.e. currently not working.

[zeroone]

For me, the appeal of the Java 4K contest is attempting to accomplish a lot within a ridiculous, artificial constraint.  If anything, take some libraries away. 

[dapy]

My preference would be to keep the existing J4K contest in its current format, and have a separate OpenGL nK contest - perhaps run at a different time of the year, to ensure that the two contests don't compete for entrants.

[Apo]

My preference would be to keep the existing J4K contest in its current format, and have a separate OpenGL nK contest - perhaps run at a different time of the year, to ensure that the two contests don't compete for entrants.

+1

[Damocles]

I also think this should be 2 seperate contests.

LWJGL requires someone with a standard Java Runtime to download a big library (hundreds of times larger than the 4k game) first.

This kind of looks odd when claiming that the game is just 4kb.

[HeroesGraveDev]

As much as I am a supporter of LWJGL, I would prefer Java2D for a 4k competition, but I don't think that should stop other people from using it.

I reckon it should be either standard Java or LWJGL, no other libraries.

Anyway, I'm going to have a shot at a 4K game to see what I can do...

(Btw, I'm assuming that it is 4KB of RAM. Is this right?)

[Alan_W]

(Btw, I'm assuming that it is 4KB of RAM. Is this right?)

The compressed jar (or nowadays pack.gz) must be 4096 bytes or smaller.
An obfusticator such as proguard and a packing utility (e.g. kzip or 7-zip) will be needed to achieve this.

Alan

[HeroesGraveDev]

Okay, that's a bit easier

[sproingie]

With java7 allowed, javafx becomes a valid target as well.

I also agree that if a big extra library like lwjgl is allowed, it should be a separate contest like LWJGL16K.  I also wouldn't mind seeing LibGDX added to the mix for another such "big library, small program" contest.

[SimonH]

I don't want libraries (much as I love them) I think a 'do you want to install a big untrusted file?' prompt would kinda kill the 4K mood! Anyway, what's the point? A good game will shine through the simplest of graphics.

[princec]

I don't want libraries (much as I love them) I think a 'do you want to install a big untrusted file?' prompt would kinda kill the 4K mood! Anyway, what's the point? A good game will shine through the simplest of graphics.

I just wondered if using an alternative rendering API would be fun for some people is all. I won't touch Java2D with a bargepole myself, it's completely pointless dead tech as far as I'm concerned.
Anyway... if the recent security woes haven't scared everyone off of using Java applets anyway (even I've disabled Java on my system, permanently, now), I doubt a signed dialog from LWJGL is going to scare anyone else off.

Cas

[zeroone]

According to Wikipedia, Java 4K ended long ago:

http://en.wikipedia.org/wiki/Java_4k

[jimeowan]

My preference would be to keep the existing J4K contest in its current format, and have a separate OpenGL nK contest - perhaps run at a different time of the year, to ensure that the two contests don't compete for entrants.

Agreed. Plus it would make the community alive throughout the year and not only from december to march.

[steveyO]

+1 for no external libraries.  The current competition format is fine, so why change it!