The Java ecosystem is a news factory these days. Herewith a few words on Java SE 6, JRuby 0.9.2, and Groovy 1.0 RC1.

Groovy! · Guillame Laforge announced Groovy 1.0 RC1. I’ve been among the skeptical, those saying “Yeah, Groovy looks cool, but I have real issues about the language syntax and semantics not having reached 1.0.” Well, looks like they’re going to make those issues go away. Which means we should all have a closer look at things like Grails.

Java! · They sent me tons of slide-shows and white-papers and so on about all the cool stuff in Java SE 6, which should be up on by the time you read this. I’m not going to try to cover that waterfront. Java has become so big and so ubiquitous that if you’re somewhere in the computing biz, something about this release probably touches your life.

The horse’s mouth would in this case be Mark Reinhold’s Flashing past the finish; check it out.

[Update: The GNU/Linux distros are ready, same-day; that’s a first.]

I’ll mention two issues: performance and support.

Performance graph, Java 1.1-6

Actual real people I know doing actual real work that pushes Java to the edge have been making happy noises about the performance; to a degree greater than that graph would suggest. It’s hard to believe that after all these years that there’s much left to squeeze out of the JVM; you have to be impressed. [Update: and check Martin Probst in the comments.]

I’m no expert on this, but apparently the structure by which you can buy support for Java apps has been re-built, for example see Sun Spectrum Java Multi Platform. People have been working terribly hard on this and expect great things.

JRuby! · The JRuby releases come thick & fast these days. JRuby 0.9.2 is another rung in the ladder; they’re closing in on a bunch of their targets notably including Rails. These are some of the guys I mentioned in the paragraph above who have been smiling about Java 6 performance.


Comment feed for ongoing:Comments feed

From: Martin Probst (Dec 11 2006, at 11:24)

The last time I checked our native XML database [ X-Hive] ran about 25% faster on Java 6 than on Java 5. Quite impressive, possibly due to creation of temporary arrays on the stack rather than the heap in Java 6.


From: Casper (Dec 11 2006, at 12:07)

Not sure how interesting it is to compare to the 1.2 some 8 years ago. However, a 10% increase from 1.5 is decent, now the memmory footprint should be improved.


From: Patrick Correia (Dec 11 2006, at 19:48)

As someone who hasn't been following the developments of SE 6, it's disappointing to go to the "New Features and Enhancements" section of the documentation ( and find a bunch of dead links -- and the ones that aren't dead point to pages that are obviously stale (e.g. the "Drag and Drop" link on that page, which points to a page last updated in January 2006).


From: Matthias Ernst (Dec 11 2006, at 23:57)

[Martin Probst]: Quite impressive, possibly due to creation of temporary arrays on the stack rather than the heap in Java 6

I'm glad about the performance increase but it's certainly not due to stack allocation. That's a wide-spread rumour - Java 6 does employ some escape analysis but it's "only" used for lock elision so far. See comment by steved further down in

Actually I wonder how many Java programs would start dying with stack overflow once stack allocation was employed ...



From: Henri Sivonen (Dec 12 2006, at 03:07)

Is there lock removal for io streams? References to InputStream and OutputStream are typically not stack-scoped but in practice a given stream is only read from or written to by one thread. (And, in hindsight, it seems that making the streams synchronized in Java 1.0 was a mistake.)


author · Dad · software · colophon · rights

December 11, 2006
· Technology (85 fragments)
· · Java (123 more)


I am an employee of, but the opinions expressed here are my own, and no other party necessarily agrees with them.

A full disclosure of my professional interests is on the author page.