[quote]Maybe it’s time to dust off the Android SDK and try it out again!
[/quote]
Still some time before its released and then a lot of time before the phone companies start updating. But hopefully since it has a few nice things from what I hear, the update will get out quicker or there’s going to be a lot more demands(I saw a ton of uncontrolled demanding for 2.1 for droid/milestone and that update was crap).
So no one here still believes in java applet games?
I think in the short term future java applets are looking great as they are finally starting to work really well, especially for 3d stuff with LWJGL and JOGL.
However in the long term doesn’t look too good for java applets considering the competition.
I hope Google does get a dalek web plugin out. More competition helps improve.
I don’t really get what having Dalvik in a browser would bring, other than slower performance. Unless it’s simply having more faith in Google?
I think that Java applets are by far the best technology for heavy applications in the browser, esp. with the recent improvements to the plugin, but it seems like some big players out there just won’t have it. Apple won’t allow any client middleware, and while Google is adding Flash to Android, it seems like they’re only doing it to look nice compared to Apple, as they know Flash’s days are numbered. Microsoft has its own technologies and its own problems. Which leaves Java without a strong consumer-facing backer.
Java developers should find it even more frustrating to see Google’s “Native Client” (released yesterday as a dev preview - http://code.google.com/p/nativeclient/), which allows running applications compiled for LLVM (rather than the JVM) in the browser. I doubt it will have better performance than Java applets, especially once the JRE is modularized, and it is certain to deliver a far worse development experience and perhaps problematic security. Like I said before, the big players won’t put their weight behind an Oracle (or Sun for that matter) controlled technology. So perhaps kapta is right, and in the near future Java is facing some bright improvements, but it probably won’t last. This is another case of the best technology losing to politics. That is, unless Java were to be truly freed of its corporate ownership…
But Java will continue to grow strong on the server, where it seems most of the exciting stuff will be happening anyway.
[quote]So no one here still believes in java applet games?
[/quote]
I think they’re great! For now…
The key question is what will Oracle do to/with Java in the next year. This next year is critical as new markets continue to emerge (like mobile and pseudo-mobile devices). And so far, Oracle has done close to nothing on the evangelist side of things, leaving most of us pondering if this is the end of Java (it’s not). Though the sad reality is, Java is losing market pull. More and more groups are moving away from the “unsure” thing and going with something that seems more stable and more forward-thinking. This is the whole reason why Mono exists, because Java was closed and Sun didn’t have open enough communication to calm people’s fears about it’s direction. Microsoft decided to go it’s own way (which is what it wanted to do with Java as well), and people actually had more “faith” in Microsoft than they did Sun.
This last year, with the Oracle buyout, Java started slipping again (the uncertainty of corporate buyouts almost always causes staggered growth). Then, key Java figures (in the popular opinion anyway) left the company. Oracle is infamously quiet (almost silent) with their communication, focused on the bottom line, and charging/over-charging anyone in anyway possible. So, what are they going to do now? No one knows, and that’s the big problem, as we’re left to make up stories about what the possible outcomes will be (the vast majority being highly negative because of Oracle’s past practices).
SO… if you really want to know what the future holds, and if you should be programming in Applets or something else… wait a year and see. Although I personally, very much doubt that Applets will be a viable future platform, even though they work just great for now.
Well, will see what Oracle wants to do. They announced many months ago that they want to invest in JavaFX among VM performance.
I second this. If you want to make a bespoke, big app that can be deployed via the browser then Java is an excellent choice. If you want something light, cheap and cheerful which 99% of people can play in the browser then there are plenty of better alternatives to Java. Worse of all they are under far more advertising and active development.
Ah, but that was before the deal had been signed and approved. Now Oracle own Sun it’s a very different world. I personally think JavaFX will continue, but it will be aimed almost entirely at the enterprise world. For building the front-ends to Oracles many enterprise Java products, and for various companies own internal apps. Oracle specialise in enterprise products; I don’t think they really care about what happens outside of that domain.
Great news. But it’s a bit unfortunate that they only benchmarked it with the rather meaningless Linpack benchmark. Linpack requires less than 400 lines of code and it only consists of arrays, floating point math, and loops.
So, if you manage to improve array access or floating point calculations (e.g. by having an FPU at your disposal - which of course also is a huge advantage for JavaScript), you’ll see a drastic jump in that benchmark.
But still… 37.5 MFLOPS is pretty impressive. I just wish they’d used a somewhat more complete test. E.g. the Shootout benchmarks would have been nice. While they are all micro benchmarks, they all got a slightly different bias, which in turn means you’d end up with a range which would actually translate fairly well to real applications.
I’m thinking the canvas element with some other interesting tools that are trying to make their way into HTML 5.
Canvas element looks nice. A quick search on the net though has me slightly concerned about possible licensing issues further down the road though.
its a big boat…
Like what? Apple? That has been sorted.
Yeah.
I wasn’t aware this had been resolved.
To be honest, so far I found Canvas/javascript to be completely underwhelming. It’s like going back in time 10 years, but worse.
Seems to be a solution in search of a problem. That problem was essentially solved a long time ago but for no particularly well explained reason some engineers have decided they want to solve it again.
Cas 
After working with engineers in several jobs (I almost was one once), I assure you, they don’t need a reason.
Yep! Everyone thinks they can make a better wheel…
Ye well, it’s currently pretty slow. Especially in Firefox. Drawing one (translated) 640x768 and one (translucent) 480x480 image on top makes it already drop down to 32fps. (Resident Evil 5 runs in high quality at full HD resolution by the way… makes it somewhat embarrassing, really.)
Over at my 10 year old machine I could do the same @ >200fps with OpenGL accelerated Java.
Currently there are two big bottlenecks: a) the compositing and b) the drawing itself is very slow (the per draw call overhead is microscopic in comparison).
The drawing in Chrome is more than 4 times faster, which is enough to make it usable. Opera’s drawing is only about 2 times faster, which is a bit meh.
But, as I already mentioned, this is a problem that will be sorted out in the future.
With WebGL the drawing itself is of course pretty fast, but the compositing overhead doesn’t allow you to crank the resolution all the way up.