Discuss? 
Jigsaw, long arrays and hypervisor awareness… Don’t go blowing me away with your wild ambitions or anything, Oracle. I know that’s not the totality of the roadmap, but what’s even the point of mentioning a roadmap if you don’t draw any roads on it?
Looking forward to Project Jigsaw in Java 8, hopefully it’ll deliver and not end up being broken in some serious way. If they get it right, it should be brilliant for creating and bundling small stripped down JVM’s with your games. Should also help decrease JVM startup time and footprint.
Jigsaw will be pretty awesome, and if they can get JavaFX integrated with a user-experience focused bullet proof cross platform browser plug-in I’ll be more than happy.
Actually I guess there is one thing that does look noteworthy, and that’s the hint that they might finally acknowledge reality and close the coffin lid on J2ME, at least as it pertains to smartphone-level devices. Oracle being what it is means they’ll almost certainly still strangle the baby in the crib as licensing goes, making Android the palatable option for the forseeable future, but at least there’s some hope of credible competition to drive Android forward at the very least.
That extra info mentions better support for structs, native code, and multidimensional arrays. Sounds pretty promising to me, even though it’s slated for JDK10
Hm yes, I reckon there’s a 5 year wait right there.
Cas
[quote]JDK7u6 Improved OS Integration, auto-update
[/quote]
Finally, I hope that means silent updates like Chrome and no more Java nagging messages for end users.
Why do they need that long?!?!? >:(
Err…changing to pure OO seems crazy. That’s a pretty much a new language. Couple that with structs = huh??
I can’t quite see the real point of making everything object oriented - strikes me as being the “if the only tool you have is a hammer…” proverb being realised. I also can’t see how it’s going to help performance or even reduce complexity, and therefore, how it will help Java remain a mainstream general purpose language. Still, all we can do is yell at them.
Cas
Is that so bad really? Everyone keeps complaining about the old one. I wouldn’t mind some radical backwards-compatibility breaking changes.
That one is a bit mystifying, yes…
I wouldn’t mind a pure OO java-like language as well as java…esp if it had some interesting features like default methods, maybe becomes, double or multiple dispatch, etc. etc. But it really should be a “new” language. Going that half-way step between class-based OO and prototyped based could be very interesting (open classes). But changing “default” java…a very strictly typed language? That’s not a backward or forward step, that’s big paradigm shift.
I like jigsaw
I don’t like auto-update
I thought java had been OO since begin.
pure OO…everything is logically an object. So no more primitive types. Apparently JDK 8 will include support of unsigned int/longs…finally (static methods in Integer & Long, compare, divide & mod).
Naw, just a standard API like you probably have already written yourself once or twice. If I read this article correctly:
https://blogs.oracle.com/darcy/entry/unsigned_api
it is more aimed at being a foundation for other language services surrounding unsigned types apparently.
I wish they’d just ditch the whole idea of a system JVM, forever, and concentrate on making Java available as an embeddable library.
Cas
Sure, but the difference is like with lead zero, trailing zero & population count like calls being added to the API, these will be promoted to single opcodes at the backend. When we write this stuff ourselves, it’s much less likely (like approaching zero percent chance in most cases).
But… that’d ruin all applets! :persecutioncomplex:
On a serious note, that would remove both the annoyance of the JavaUpdater and the drive-by infections we’re dealing with, which eventually will be better for Java’s reputation.