Reinhold Already Talking About Java 9
Java 9. That's not a mistake. In John Duimovich and Mark Reinhold's "Java Renaissance" keynote today at EclipseCon 2011 they talked about Java 7, which finally has a release date (July 28, 2011), Java 8, which will bring modularity that is compatible with OSGi, and... wait for it.... Java NINE.
For those of us poor souls who are not at the conference, which has been really awesome so far (according to attendees), all we have are a few key tweets from Justin Kestelyn of Oracle:


And a separate tweet on JDK 9 that credits Arun Gupta for also mentioning the ideas:

Crazy that they're already thinking that far ahead when Java 8 is still more than a year away. So what are your thoughts? Are these features everything we ever wanted? Or are they just more clutter in an already complex JDK?
Another point of major interest to supporters of OSGi is Java 8, which will implement its own brand of modularity through Project Jigsaw. Many parties, including the Eclipse Foundation, are very interested in OSGi succeeding. This is why Reinhold wanted to assure audience members that there would be an OSGi migration path from Java 8. Some tweets about that:


Java 9, folks. What do you think?
For those of us poor souls who are not at the conference, which has been really awesome so far (according to attendees), all we have are a few key tweets from Justin Kestelyn of Oracle:


And a separate tweet on JDK 9 that credits Arun Gupta for also mentioning the ideas:

Crazy that they're already thinking that far ahead when Java 8 is still more than a year away. So what are your thoughts? Are these features everything we ever wanted? Or are they just more clutter in an already complex JDK?
Another point of major interest to supporters of OSGi is Java 8, which will implement its own brand of modularity through Project Jigsaw. Many parties, including the Eclipse Foundation, are very interested in OSGi succeeding. This is why Reinhold wanted to assure audience members that there would be an OSGi migration path from Java 8. Some tweets about that:


Java 9, folks. What do you think?






Comments
Jochen Bedersdorfer replied on Wed, 2011/03/23 - 9:36pm
I say, forget about modularity and OSGi and bring on multicore-scalability, hypervisor FIRST.
The time is now, not in 3 years.
Artur Biesiadowski replied on Thu, 2011/03/24 - 7:11am
I'm feeling a bit out of the loop here:
Hypervisor thing is about implementing some kind of java Isolates to support hosting of multiple apps in shared environment?
Multicores - I suppose we are talking here about massive multi-cores, like 1000+. I think that java fares quite well on 8-32 core systems, doesn't it ?
How they plan to help multitenancy in JDK? I thought it is more application-level, rather than virtual machine-level feature?
@Jochen
What kind of systems you are running java which would require massive multi-core scalability? I suppose Azul is doing it's own work in the area, but Sun JDK??
Andy Leung replied on Thu, 2011/03/24 - 9:55am
Jochen Bedersdorfer replied on Thu, 2011/03/24 - 2:04pm
in response to:
Andy Leung
Java is not only a language and as a platform hosting applications, it needs to make sure that it can actually use all the modern facilities provided by the OS.
This is not the case right now.
Jochen Bedersdorfer replied on Thu, 2011/03/24 - 2:10pm
in response to:
Artur Biesiadowski
Despite all the advances in java.util.concurrent, concurrent programming is still very hard and error prone.
(A need scala claims to address. With a language that is more complex than Java, hmmmm)
While the JVM does a fairly decent job working with multiple cores, any sort of non-trivial parallel programming problem is still hard to implement and to maintain if you want to maximize use of the CPU.
While vendors exist to ease the pain here, proper support on the language and VM level is highly appreciated.
Alex(JAlexoid) ... replied on Thu, 2011/03/24 - 4:33pm
in response to:
Jochen Bedersdorfer
Unfortunately we are left with either hard/unnatural programming models(functional programming), error prone concurrency or transactional memory. Since most issues hover around shared and mutable data.
But concurrency was hard for decades and I don't expect it to become easier any time soon.
Even will all of that syntactic sugar, unless you dispose of shared and mutable data you still have the same issues.
Liezel Jane Jandayan replied on Mon, 2011/08/22 - 9:12pm
James Kear replied on Tue, 2011/09/06 - 2:43pm
Carla Brian replied on Sun, 2012/04/15 - 10:21am