Tropical Software Observations

20 January 2009

Posted by Unknown

at 9:24 PM



Sun Tech Days 2009, Singapore 20 January

Just finished the first day of Sun Tech Days 2009 -- in summary I give it 4 stars out of 5.

Arrived early and saw a huge crowd of people queuing up for registration. The registration process was efficient -- I got my badge, orange goodie bag, and was quickly shown the way to the foyer. I'll be frank, I was expecting a light breakfast. No breakfast, only coffee and tea. Well as long as they serve coffee I am happy.

The welcoming ceremony started at around 9:10 am. Man, it was hilarious to see the big guys dressed as some Chinese gods. Check out some pictures and video in Arun's blog.

We were given a small Chinese rattle drum on the way into the hall. It's cool to use those small drums instead of clapping. This drum rattling event was supposed to be in Singapore's Book of Records for largest number of developers drumming away. Well whatever floats their boat.

Next thing I remember is seeing Mr Jeff Jackson taking on the stage to coordinate a few demo sessions. Between the sessions I remember lots of mentions of Glassfish, social networking, Google, YouTube, Ruby, OpenSolaris, etc.

Notable demos in the technology showcase:

  1. Mr Arun Gupta demoing Glassfish Prelude -- nice to see Felix container there. Which reminds me, I need to check with the folks at Qi4J. It's been a while since I've dabbled with OSGi stuff.
  2. jMaki demo -- yawn yawn.
  3. Simon Ritter's Wii controller hack -- very cool stuff. I hope he will put the projects and code that he used for the Wii hack in his blog soon.
  4. Somebody got lucky and went home with a Toshiba laptop. I am still jealous.

Next I went on to the JavaFX: The New Platform for Rich Internet Applications track session. Well I had my doubts about JavaFX before but not after this session. Man I'm going to pull out my favorite book on animation and graphics -- Foundation Actionscript 3.0 Animation:Making Things Move and try some exercises with JavaFX. The Wow factor is definitively there but I'm also wondering aloud if this could be another one of those J2ME feel-good-sessions: you know everyone feels really excited to see the rallying around J2ME in McNealy's days as the next platform to bring computing power to everyone's palm, and people end up jumping in droves to Android and iPhone development. Anyway -- a big applause to the demo, it was smooth and very entertaining. Especially when the sharks swim away!

After that I went on to the DTrace: Bridging the Observability Gap for Java and Scripting Applications. DTrace is one of those technologies that I follow closely but could not grasp fully at the same time. It's like a myth that -- that until I really see it as really really cool, I have to keep my mouth shut in telling others about it.

Peter Karlsson was the speaker and though he had some hiccups with his demo, I really wished he didn't give up. He should've probably stuck his fingers on the console, fixed the problems and continued calmly. But I guess keeping calm is easier said and than done when you are running out of time. I enjoyed it anyway but since I will be attending the 9400 – Exposing the Depths of Java Applications with DTrace I thought I will have my chance to know DTrace better later.

Then we are let out for a 15 minute break. I had another cup of coffee and some cakes. Nice. Next to my lab session on DTrace.

Here is a summary of the LAB-9400: Exposing the Depths of Java Applications with DTrace session:
  1. Peter Karlsson and Joey Shen are conducting this session
  2. Many didn't come with their laptop and I'm not sure what they got out from this since the session is pretty code intensive
  3. Of those with laptops I'm the only one with Mac so the bootable USB of OpenSolaris won't work on me. Oh well at least I have DTrace built in on my Mac OSX 10.5.6. heh
  4. The lab slides are about 200+ pages -- we were told that it would take about 8 hours to cover them all. We have like about 2 hours. :-O
  5. The depth of DTrace is overwhelming -- I barely understand the D language: probe, predicate, and action scripting.
  6. While Peter was talking about the introduction of DTrace, I did a quick search on "Mac OSX dtrace dtracetoolkit", I realize that my mac already have most of the DTraceToolkit available. Try "ls /usr/bin/*.d"
  7. Later I got a good grasp at how DTrace and D language works -- I began to open up the D scripts on my Mac to see the gory details of the scripts.
  8. Too bad Java 6 on Mac doesn't support DTrace yet. (Apple tends to bump up Java guys)
  9. DTrace is more than a super steroid debugger -- you can turn it into all kinds of monitoring instrumentation. (Need to google for an anti-keylogger D script)
  10. This session is way too short and too brief. If Peter and Joey are reading this, may I suggest to you guys this: pick a topic short enough to fit into 1.5 hours. Stop telling me about all the cool stuff I can do with DTrace. I know that already, or else I wouldn't be sitting there in the first place. Show me how to "X-Ray" my Java apps!
  11. At the end of this session, I am reminded of the saying, "the teacher only shows you the door, you have to go in yourself". Well thanks, I will be writing about D scripts shortly.

Next, maybe I had too much coffee, was palpitating. By now it's already 4.30pm, I had another coffee in this very short break.

On my next session track, I have the choice of going to:
  1. MySQL: The Database for Web 2.0
  2. Java SE 6 Update 10: Simplified Java Deployment
  3. Thriving on OpenSolaris
I chose the OpenSolaris session because I wanted to see some OpenSolaris 2008.11 in action. Here are my thoughts of that session:
  1. I was late a bit and didn't get the speaker's name. I got the overall impression that he is floating the on clouds with OpenSolaris here.
  2. Way too much time playing with Compiz on the Gnome desktop. C'mon guys, it's Compiz -- lets move on already!
  3. And seriously, somebody should tell the crowd that OpenSolaris is using the Gnome desktop -- we're developers, we will get it even if you need to explain such things as X11, desktop client and Unix stuff like that.
  4. The ZFS part was not impressive. I have seen better webminar videos on this. yawn
  5. The DTrace part was a waste of time. He should've probably skipped it altogether or summed it up in a one paragraph or politely pointed to some online resource.
  6. The speaker fumbled way too often on the command line. Stop using "clear", use "ctrl+l" to clear the screen. (hold on, is OpenSolaris using Bash by default?)
5.45pm, rushed to the Groovy and Grails session track. Boy, I have something serious to say about this.
  1. I was expecting to see some experts sum up to us the latest JRuby progress in this session.
  2. Why the heck is the description of the track about JRuby on Rails yet titled as Groovy and Grails? If I knew it was really on Groovy and Grails I would have skipped it and gone to the Porting Open Source Packages to OpenSolaris session.
  3. Mr Chuk Munn Lee, Sir, I don't think you did your homework on Grails. I sort of tuned out when I heard Mr Chuk said that Grails doesn't support many-to-many relationship. I thought I heard wrong, but he followed by explaining how you need to tweak some mappings to get that feature.
  4. The example on Groovy was too shallow and the Grails part was not good. I think the audience would appreciate what Grails is all about by starting the introduction to domain modeling and GORM instead of starting with "How many of you have programmed taglibs in JSP before?"
  5. Mr Arun Gupta was 3 seats away on my left -- he looked tuned out too. (Lol)
On my way out, I handed in my feedback form and was surprised to get a nice flap-over-bag. It's basically a small soft messenger bag. I'm happy to find the latest OpenSolaris 2008.11 inside. See the other goodies I got today below.


Chuk Lee said...

Hi this is Chuk. Guilty as charged. I've not seen the m:n relationship. Will try to update that. Will try to do better next time. Thanks for the feedback.

Unknown said...

Glad to hear that. :-)

Unknown said...


Peter her, we know the HOL is way to big. We are re working it in to two HOLs. One basic and one problem oriented, which I guess that is what you want. I'll try to figure out a way to do that HOL in Singapore when we are done :)

And it was Alex not Joey that was helping me with the HOL

As for the demo in my session, it looks like I found a bug in the current JVM in the dev IPS repository

Unknown said...

"Qi4j Folk" here...

Not that I understand the connection between Qi4j and Glassfish, but perhaps you just recall the wrong aspect of similar "folks", and meant the "OPS4J Folks"

- Niclas Hedhman