Java 7 Roundup (Oct 2nd)
This is a weekly update on new Java 7 information. All current and previous information is maintained on my Java 7 page.
Kelly O’Hair posted the third update on the transition to Mercurial for source control. The plan is that build 22 will be the last build using the old Teamware system and build 23 will be the first on Mercurial. They’ve also looked at code conventions and converted all tabs to spaces. Let the flame war begin… :)
Speaking of builds, Java 7 build 21 was released this week. Build 21 breaks out a few more pieces from the main workspace – corba, jaxp, and jaxws.
Modularity
InfoQ posted an interview with Peter Kriens this week, who is well known in the OSGi community. There were a number of interesting exchanges related to JSR 277.
And we also got the news this week that Glyn Normington (member of both JSR 277 and the OSGi JSR expert groups, and general voice of reason) is leaving IBM and thus also leaving the JSR 277 expert group. This is a big loss to the group but it sounds like Glyn is heading for another great opportunity, so I wish him luck!
I also would be remiss if I didn’t the just released Java SE 6 Update N (beautiful name) which is the formerly known Consumer JRE or Java Kernel. This is an early access version but it pulls together a lot of great work to improve the Java web experience, including an improved download and startup experience (based on the kernel work), new look and feel (Nimbus), better graphics acceleration, and better detection of the user’s Java version. This is nice work from Sun and I’m super glad to see it in Java 6, not waiting for Java 7. Kudos to whoever made that call.
More info: JSR 277
Libraries
Stephen Colebourne spent some time this week reviewing JSR 275, the units and quantities JSR, which might have some overlap with the durations and time units being defined in JSR 310. Sounds like the results weren’t too encouraging but you should read his comments for the details.
There has also been a ton of traffic on the JSR 310 mailing list in the last week, more than I’ve had time to assimilate. But of particular interest (if you wish to give feedback) are these updated pages on JSR 310 progress:
Stephen is actively looking for feedback on these, so if you’re interested in the new Date and Time API, please give them your attention and provide feedback on the mailing list.
Language Proposals
Chaotic Java had a post this week about type inference and whether it would be worth it.
Lots and lots of stuff this week on the good old property debate, summarized here:
- Stephen Colebourne on whether properties should be per-class or per-instance
- Rémi Forax on how JSR 295 beans bindings could interact with property support
- JavaLobby thread on whether property support is needed or wanted in Java 7
- Stephen Colebourne proposes some standard terminology for the property discussion
More info: Properties
Hi! My name is Alex Miller and I live in St. Louis. I write code for a living and currently work for 
Java 7 has no support for parallel processing. What means Java 7 is a dead-born child. Moore’s law tells us that a 4-core chip today will be a 16-core chip in 3 years. Another 18 months later it will be a 32-core chip.
Where is Sun’s answer to that problem? Shall we write single-core applications for a 32-core processor using Java 7? It seems that Sun has no answer to that problem. But it seems that IBM has the answer we are awaiting:
http://x10.sourceforge.net/x10home.shtml
If x10 sees the light of the day this will be the end of Java and it will be the end of Sun’s control over Java-technology. Moore’s law demands fast solutions and strong powerful companies supporting those solutions. Companies like IBM, Intel and Google. Sun will be out of the game in 2 years.