Thoughts about java, ruby, agile and other every developer stuff.

Thursday, June 28, 2007

Jazoon - Day 3

Yep this was a social day. We've got just few presentation, we have Die Hard 4.0 pre-presentation and an occasion to see 5D movie, than from 20 to ... we have Sun Microsystems party. It's 10 years jugs (Java User Group Switzerland) and 25 years of Sun Microsystems.

The keynote was about the IBM jazz which I will describe a little bit more latter. There was also some news about Spring Framework version 2.1. Other presentation was about memory management, JavaFX, JBI 2.0 and Groovy.

Today we had two keynotes, and my thoughts are in Poland. It’s time to came back and do some movements to achieve dreams ;)

Wednesday, June 27, 2007

Jazoon - Day 2

Yesterday talks were amazing. The most interesting talk was about glassfish architecture by Jerome Doches from Sun Microsystems. He talks a lot about low level architecture of glassfish v3, and what can be interested glassfish starts in 2-3 seconds, and …. Supports rail, phobos and some others application which are no java application, and it happens automatically. Another talk was about Hibernate Search and I write more about it later because it very advance topic and I want know more about it.


JMaki also made impression on me. It is very nice framework and if I must compare it to something you know, I would choose spring framework. jMaki is something as spring but for javaScript libraries. And of course jMaki topic lands on my to-know list ;)

Others talks was also very interesting about lucene, java db and performance (it faster than MySQL !!!) and the last one was about Java Benchmarking Tool build on JUnit. If I find some time I will look closer to this topic.

The best part was BOF sessions, and I choose a two session with Neal Gafter. He is responsible for java closures specification, we are talking about it nearly two hours, and now I 100% convinced to closures in java, next topic was about java puzzlers and Neal learn us a lot. After four hours (I didn’t recognize that) I take a walk to home, that was raining but Zurich is the beautiful city I ever seen, I make some photos and go sleep. Today social event day ;)
And now I am waiting for Erich Gamma speak.

Pedro

Tuesday, June 26, 2007

After Day One

Aaron was great, he told us about python and groovy and java limitations such as
  • bean creation - why sun don't create setters and getters for us, when we would be know if setter/getter do something or not.
  • closures - it's normal for ruby groovy and others
  • ORM limitation - ORM in python and grails is nearly perfect
  • list - build in list and map support for python and groovy makes programs more readable.
  • null - groovy .? operator is perfect for null pointers, you don't have to do if (a ==null) or (a.b() == null) ...

The next was Nati Shalom with "Scalable as Google simple as Spring" He told us about Space Base Architecture (SBA), and showed us a limitation in scalability standard solutions. The main reasons are

Data Tier - I/O overhead, contention
Bussines Tier - Processing cpu
Messaging - complex workflows increase latencies

There we try divide this it will create another problems with 2PC (2 Phase Commit). One of the solutions is change from tired base to service centric. But services have its own limitation where they must be state full. Loosely coupled services mean more opportunity to improve scalability, but when services are state full it is bad idea. So we can do SOAP on space introducing read, write, take notify operation. Than we can create processing unit in one VM and adding more PU makes scalable linear.

GigaSpaces with Interface21 make for free OpenSpaces which you can try now.

Where was also talk about Rich Client Platform and remoting which suggest that
- we must consider Single Sign On system - client may use many different systems
- Spring supports such remoting as
Hessian, Burlap, Spring HttpInvoker, RMI, RMI-IIOP/CORBA

- we can consider also Jboss remoting
- and of course don't use webservices if you don't have to

Another talk was about Sun SPOT (small programmable object technology)
http://research.sun.com/projects/dashboard.php?id=155. And it was amazing because hardware which Simon Ritter from Sun Microsystems shows us. He introduced Squawk which is a research virtual machine for the Java language.

Next talk from Bruno Bossola Java Chamion(https://java-champions.dev.java.net/) and Torino JUG member was about
Enterprise Application Testing using open source tools
He told us about Fitness and the selenium.

Fitness - it's wiki that allows you write a test in a table from customer
input and expected result. Because it's a wiki so you can write with it documentation. We can run test on business logic, application services, internal or external, stored procedures, it can be run by maven.
Time goes and Bruno told about selenium and selenium RC for just few minutes/

And the last was about "Dualcore, Niagara, Rock -
Why Java Developers should care about Hardware"
by Peter Gassmann from (Sun Microsystem)
He speaks about new processors and that we as java developers should care about it, and try to write software as multithreaded. Today we have more then one processor in our laptops.

Suggested read : "JAVA concurrency in practice" by Brian Goetz

So write application as multithreaded for as many as you can imagine processor and check your hardware where you tuning.

Monday, June 25, 2007

The First Day - Jazoon

I’ve got some problems with my trips. I’ve got a hotel in Basel, and Jazoon is in Zurich, maybe it’s not so big problem, but I don’t like waste money and time. Unfortunately there was some conferences in Zurich other than Jazoon so every hotel in town was fully occupied. When I was accepted my Basel hotel, I gave one chance, very small hotel without WIFI and others stuff. So I’ve found it. It’s in center of Zurich downtown, great place to stay and sightseeing Zurich.

I am in love in Switzerland and Zurich, and of course I was informed my wife that we move to Switzerland as soon as possible.

The keynote by Ted Neward was about languages build on top of JVM. I was very surprised because on December 2006 I’ve got similar seminary in my company. There are so many such languages and what Ted suggested to everyone is to try build your own language because it’s now very easy to do. The key features are
• Virtualization which make we can worry about memory management and others low-level stuff.
• Tools which helps us build parsers, AST, provide us opportunity to modify byte code.
I certainly try write my own language.

Now waiting for Aaron Digulla speech about what wrong with java.

Pedro