After half an hour of walking around trying to determine where exactly the convention centre IS in the Olympic complex (signage would be really nice), I finally managed to make it to the Melbourne satellite event of the Australian leg of Sun Tech Days. I love events like this; the interchange of ideas and pointers in new unexplored directions really get the mental juices flowing. I was a bit late, but still managed to catch most of James Gosling’s keynote.
The highlights for me were many, and they’re going to take some digesting.
James destroyed the “Java is slow” myth (anyone who still believes this hasn’t fired up a new JVM lately). Java runtimes are incredibly optimized with test results showing performance equalling or beating C/C++ equivalents.
-
Linpack -2%
-
Scimark + 4%
with GC being a lot faster than malloc/free.
The reason why dynamic compilation beats its static equivalent as the JVM is able to tweak performance depending on the processor type being used, even in the same type of architecture. This enables the JVM to take advantage of the strengths of AMD chips over Intel and vice versa.
There are lots of good things coming up in Java 7, both in the core and on the mobile. I will be detailing them a bit more once I get around to working out what all of the JSR numbers I scribbled down meant :P. Too fast with the old Powerpoint.
The question of RAD tools came up. James quantified it with a question – what exactly do you mean by rapid? Is it time to demo or time to production deployment? I had never really thought about this, but it does make sense. Java is being focussed on time to production. The reason for this blew my mind. Venture capital provides funding in 3 month lots only. In that time you need to turn an idea into a production grade system.
3 months from idea to production.
As developers we need to scout out the enabling technologies behind this kind of turnaround and work it into the toolbox. And enterprises need to have a good hard think about why they are not achieving similar results (and no, the answer is not to kill your programmers with 120 hour weeks to do it). This only highlights the discrepancy between small startups and large organizations. I have yet to see anything get put in production in less than a year in a large institution.
The other highlight for me was not what was presented, but rather what could be gleaned from the feedback questionnaire. This was one of those basic “who are you, what do you do and what are you using to do it?” numbers. It listed a whole bunch of technologies that Sun are presumably keeping their eyes on. The stand-outs? RoR, Groovy, Grails and Wicket. If you want to skill up on what’s going to be big on the job boards within the near future, these would be a very good start.
The things I’m looking forward to playing with as a result of this morning? JMaki – a super-framework and Netbeans plugin that glues of all the best AJAX frameworks together (mash-ups faster than you can say Dojo), and of course JavaFX – super sweet user interfaces done as simply as a web page.
Oh yeah, and watch out for an announcement in the very, very near future about PHP and its relationship to Java.
Tomorrow in Sydney!