Gwadanina Rss » JRoller

JRoller (25 unread)

Skip to page: 1 2 3 ... 10

Indonesia on Java and Mobile Investment

Posted: March 31, 2008, 10:59pm CEST by Frans Thamura

I just try J2MEPolish, this is the interesting project, and have a consolidaton interest with Blueberry, DoJa, and they are the market that Sun never touch, because we know that Sun want to strick with JavaME, and push this standard. event for me.. JavaMe, is bad. bad costly stuff....

This mobile idea arrive because I have a new mission for developing Java community, I will be the first guy that incharge helping JUGs esp in this country to grow, and I am helping JUG Joglosemar, and JavaME Indonesia, of course the JUG that I founded, the JUG Indonesia.

There are several area scope that I feel, this country is to big (25 millions Internet user, growing 25% yearly, 47 millions students)...and I can see JUG Indonesia is to far behind to control this country. Thx Bruno Souza, for your 5 years effort to me, and I can see, the mission that I held now.

The good thing of Indonesia, we have integrated Java curicullu, and more complete than anyone in this country, FREE, Collaboration model, and have OpenSource Exam solution that can be use to make certification not only for our Java program (JENI), but for anyone that related with exam and ceritifcation. Let's bring any standard in your desk as cheap as possible, or may be FREE but sustainable.

I believed that mobile market is getting interesting, and esp after my friend, Wendy, show too much iPhone to me, make me interesting to own one, but we can see iPhone is not 8 millions users yet, and the SDK still not in the market. I registered, and still cannot download the SDK, how an annoying marketing from Steve Job. He still need a new model to develop community, must be market model than Steve's ego marketing model. Remember Steve, KTouch is 340 millions devices vs your 8 millions, but of course you are more more full of press release like another american promotion model, bigger marketing than the sales. beware of the china movement...

I can see China can become another center to be monitored as the leader in the future global leader, and can become a new next big thing, esp if China invest his time in community bulding that cult.

Here, hopefully i can catch up, and I am interesting to invest my time in Mobile Development market, even personally i dont believe in this market, esp the telcos that dont have commitment in this area, and they only interest to invest in SMS Premium and Call only biz. I still seeking Content Marketshare, but I see that SMS Premiums is the only prospectus.

Man, me need something cool in the mobile market.. and I can see the costly movement there, because Nokia with his Nokia Forum, dont have movement, and the rest of mobile devices dont have development penetration, they are devices reseller, and this is time for China player to get this market..

Android, where r u? will u be in the market???
Japplis Toolbox NetBeans plug-in released

Posted: March 31, 2008, 7:55pm CEST by Anthony Goubard
Yet another release.
This time I released a new NetBeans plug-in. It works with NetBeans 6.1 Beta.

[toolbox.japplis.com]

Features:
- URL encoding and decoding
- Base64 encoding and decoding
- Hex encoding and decoding
- SoundEx and refined SoundEx encoding and decoding
- Metaphone and double metaphone encoding and decoding
- Convertion of numbers in hexadecimal, octal and binary
- Get Java system properties, Swing UI defaults, Environment properties
- Text info: character count, word count, MD5, SHA
- XML Validate
- List the characters with the defined font and charset encoding
- Convert long to date, long to date
- Verify that a regular expression matches the text or is included in the text
messaging in Resin

Posted: March 31, 2008, 7:09pm CEST by Scott Ferguson
We updated Resin's message-driven bean configuration and jms configuration and handling, simplifying the configuration, and also making a java.util.concurrent.BlockingQueue API available.
The main documentation is at Resin messaging.

The sending side of JMS has always been annoying. At its core, sending a message to a queue isn't all that complicated. If you use the BlockingQueue API, it should just look like:
```
@Named("my_queue") BlockingQueue _queue;

void doStuff()
{
  _queue.offer("my message");
}
```
But JMS has always overcomplicated that operation, adding the following requirements:
1. configuring and injecting a ConnectionFactory
2. configuring and injecting the actual Queue
3. creating a Session
4. creating a MessageProducer
5. creating a Message object from the Session to wrap the data
6. And, finally, sending the actual message
Which just seems overkill for the basic use case. Yes, for more complicated scenarios, all those steps do make sense, but it's just an unnecessary barrier in the simple case.

So we added a simplified configuration, injection and used based on the BlockingQueue API. The code looks like the previous example. The configuration looks like:
```
<web-app xmlns="http://caucho.com/ns/resin">

  <jms-queue name="my_queue" uri="memory:">
    <init/>
  </jms-queue>

</web-app>
```
The "uri" attribute selects and configures the queue implementation, so you can choose from "memory", "file", "cluster", etc. Any configuration goes into an <init> block. Message Driven Bean Configuration

The receiving end of the message queue is based on the EJB message-driven-bean API. An application's implementation will be a class that implements the MessageListener interface. Since the listeners are pooled, Resin will spawn a number of listeners and set them listening for new messages on the queue. The code looks something like:
```
import javax.jms.*;

public class MyListener implements MessageListener {
  public void onMessage(Message msg)
  {
     ...
  }
}
```
A real application will use WebBeans injection to get access to the application's resources, of course.

The configuration for the message driven beans follows the same pattern as other WebBeans-style beans, and just adds a <destination> tag to configure which queue to read from:
```
<web-app xmlns="http://caucho.com/ns/resin">

  <jms-connection-factory uri="resin:"/>
  <jms-queue name="my-queue" uri="memory:"/>

  <ejb-message-bean class="qa.MyListener">
    <destination>#{my_queue}</destination>
  </ejb-message-bean>
</web-app>
```
A nice bit to the configuration is the EL expression #{my_queue}, which provides a nice, logical way of attaching the queue to the message bean. The great thing about WebBeans, EJB-lite, etc, is how everything is starting to fit together, and build off each other.
How I write Blog entries

Posted: March 31, 2008, 4:00pm CEST by Oliver Burn

How I write Blog entriesFor a long I have been writing Blog entries using Emacs and the nxml-mode to ensure XHTML compliance. More recently I have started using the Emacs Muse publishing system, which I also use for personal notes and to manage a personal Wiki. Today I thought I would try out the facility to publish a document from Google Docs, which I found out about from here.

Alas it did not work. The problem was that I was using the URL http://www.jroller.com/xmlrpc which is suggested by the Google documentation. This guy has the same problem. Reading the JRoller manual revealed that I should be using the URL http://jroller.com/roller-services/xmlrpc instead. As this post demonstrates, it works!
How to Make a JAR executable in Windows?

Posted: March 31, 2008, 3:10pm CEST by Vijay Kumar

In the manifest file of a JAR, it is possible to specify the class to be used when the JVM is lauched with the JAR as parameter. The class must have a main().

Here is the simple example:

import java.awt.*;
import java.awt.event.*;
public class MyClass {
public static void main(String[] args) {
Frame f = new Frame();
f.addWindowListener
(new WindowAdapter() {
public void windowClosing(WindowEvent e) {
System.exit(0);
}
}
);
f.add(new Label("Hello world"));
f.setSize(200,200);
f.setVisible(true);
}
}

Then create this manifest file (manifest.mft) with any text editor.

Manifest-Version: 1.0
Main-Class: MyClass
Class-path: .

Main-Class specifies the entry point with the main(String args[]) method. The Class-path is used to specify the dependency of this jar (if any). You add the directories and jars separated by a space. It is important because when running a jar , the CLASSPATH definition (as defined by the environnement variable) is overridden.
Next, you include the manifest file in the JAR (MyJar.jar) with the MyClass class.
jar cvfm myjar.jar manifest.mft *.class
Then you are able to start the MyClass.class by double-clicking on the MyJar.jar file (if the JRE is correctly installed) or by typing
java -jar myjar.jar
If you need to pass parameters, use the -D switch before the -jar switch and use the getProperty() method to get the value. The file association mechanism is made during the installation of the JRE. You can verify that everything is ok with the following command from a DOS Shell

>assoc .jar .jar=jarfile >ftype jarfile jarfile="C:Program FilesJavajre1.5.0_10binjavaw.exe" -jar "%1" %*

If the association is broken or you need to change the JRE used then by using the assoc/ftype utilities, you can modify the association easily (use /? to display the syntax of the assoc and ftype utilities).

NOTE: On WinXp (or better), your user account needs to be at the Admin level.

Regards,
Vijay Kumar,
Mail Me @:vijay@netbusiness.com
Execute a class / Java application from Windows Explorer with out DOS Console.

Posted: March 31, 2008, 2:51pm CEST by Vijay Kumar

This HowTo will show you how to start a class / Java application (with a main() method) without opening a DOS Shell or using a JAR file directly from Windows Explorer.

We can use the below VB script for running a java application with out DOS Console.

Dim WSHShell, FSO, javaclass, javacompletepath, cmdline
Set WSHShell = WScript.CreateObject("WScript.Shell")
Set FSO = WScript.CreateObject("Scripting.FileSystemObject")
cmdline = "javaw -classpath .;ABCDApp.jar;swingx.jar;........xyz.jar.. lucene-1.4.3.jar;looks-2.0.2.jar;jython-2.2a1.jar -Djava.library.path=. org.mycompany.app.LoginMain
WSHShell.Run cmdline, 1, false
Set WSHShell = Nothing
Set FSO = Nothing
WScript.Quit(0)

Regards,
Vijay Kumar,
Mail Me @:vijay@netbusiness.com
SpringOne 2008

Posted: March 31, 2008, 2:19pm CEST by Rinaldo

E' in linea il programma di SpringOne08

http://www.springone.com/display/SpringOne08/Home
Distributed Data Structures for Java (dds4j)

Posted: March 31, 2008, 2:17pm CEST by Talip Ozturk

Distributed queue implementation is just a beginning for Hazelcast. The goal is to make Hazelcast a data distribution platform for Java. To achieve this goal, the platform should cover the distributed implementations of commonly used data structures such as queues, sets, arrays, maps, tuples and hopefully btrees. This is all about empowering architects and developers by giving them such a powerful tool so that they can easily design and develop faster, highly scalable and reliable applications for their businesses. Here is the plan.

Distributed implementation of {

        java.util.{Queue, Set, List, Map}

        java.util.concurrent.locks.Lock

        java.util.concurrent.{BlockingQueue, Executor, ExecutorService}

        tuples

        btrees

        other requested APIs.

}

plus transaction support.

All of these features will be sharing the same guarantees: highly available, scalable, fast and easy. Cool enough?
Theory and DataPoint in JUnit 4.4

Posted: March 31, 2008, 12:11am CEST by Andre Dietisheim
JUnit 4.4 brings a very interesting possibility to test various datasets:

Theory and DataPoint

A Theory is a test method that shall be executed with different datas. Those datas may be declared as DataPoint(s). The test runner applies all compatible DataPoint to a theory. In the following example (copied from the release notes), filenameIncludesUsername (annotated with @Theory) is run twice with both *USERNAME* strings (tagged with @DataPoint).
```
@RunWith(Theories.class)

public class UserTest {

  @DataPoint public static String GOOD_USERNAME = "optimus";

  @DataPoint public static String USERNAME_WITH_SLASH = "optimus/prime";



  @Theory public void filenameIncludesUsername(String username) {

    assertThat(new User(username).configFileName(), containsString(username));

  }

}
```
The data to be tested may be restricted by using assumptions. An assumption that is not met makes the test runner skip the current datapoint(-test).
```
@Theory public void filenameIncludesUsername(String username) {

    assumeThat(username, not(containsString("/")));

    assertThat(new User(username).configFileName(), containsString(username));

  }
```
The assumption above ensures only dataPoints without '/' are tested. This way separate test methods (theories) may be used to declare the different behaviors (if a username contains a slash and another one if it does not). Scenarios (and their rules) are declared in a more explicit and concise way, your test do document your implementation in a more explicit way.

JUnit 4.4 release notes

Detailled discussion on usage and purpose of Theories and DataPoints
From zero to a fully deployed and functionally tested web application in 5Kb

Posted: March 30, 2008, 9:24pm CEST by Daniel Lopez
It’s not something you would normally do for a production application, but in my last project, I have been able to start with an Ant file of ~5KB and end up with a fully deployed application with some functional tests run against it.

How did I end up doing that? Well, here comes the story:
- A while ago I started experimenting with functional testing so I could run some tests on my pet framework: WebLEAF. The idea was to create a small demo application that people could run and at the same time, that I could also use to run some functional tests on the application/framework.
- For that purpose I created the WebLEAFTest project and created a simple HTML interface and a fancier interface using ExtJS so I could also test some AJAX/JSON requests.
- I wanted to be able to run the tests automatically so I could also experiment with continuous integration, so I added some classes to start the web application using Jetty, the light servlet container, in embedded mode and run the tests using HTMLUnit/JUnit. This paragraph might seem simple, but it involved learning lots of new tricks and using tons of different libraries.
- Then, when I started to write the installation instructions for the application & tests, I realized that the number of different libraries was overwhelming (more than a hundred .jar files from around 28 different libraries) so if I really wanted someone to have a look it, I had to lower the entry barrier. Packing so many libraries with so many different and incompatible licenses was not an option, so I opted for using Ivy, the Apache dependency manager, to download all the required libraries automatically. Easier said than done, but I managed somehow to untangle all the dependencies and duplicated libraries and then find alternative ways of getting the dependencies not directly available in the maven repositories.
- One step further and thanks to Ivy’s examples, I also got Ivy to install automatically if not already present. Great!
- And then I just got an itch I had to scratch… if my project is open source and source can be downloaded directly… why not get the source from the repository, including the build file that does all the magic? And while we are at it, we can ask the user the 3 settings needed and we can download/build/run everything from zero, right? Hence a mini-build file was born that basically asks for the user input, performs a “check out” of the project sources and then calls the appropriate targets to build everything from the sources…
That was a long trip, but in the end, you start with an 5KB build file and you end up with this:
- A self contained web application, expanded as I don’t favour war files in my projects, that shows some basic CRUD operations using different flavours for the business logic (Groovy, compiled Groovy, Java/ plain JDBC, Hibernate, JPA, PHP-Quercus) and with two flavours also for the interface (plain simple HTML and a fancier JavaScript/AJAX interface).
  The application itself uses the embedded database HSQLDB so it requires zero-installation for the DB part.
  The JPA provider used in the default installation is Hibernate’s, but I’ve also made it work with TopLink and Caucho’s Amber implementations just by changing 2 descriptor values.
  Here are some screenshots of the simple demo application, when deployed using a regular servlet container:
  The very simplistic HTML interface: The fancier AJAX interface:
- You also have the application’s source and everything that’s needed to re-create it, including the build scripts to deploy/test it.
- A set of tests that start up the application using Jetty embedded (no external server required) and then perform some functional tests using HTMLUnit, including tests upon the JSON requests using JSON-lib.
- A directory with all the required dependencies ( around ~37MB of artifacts and dependency descriptors ).
Starting from scratch, it takes around 9min 55secs… but it spends around 9min 40secs to download the dependencies, so the real update/build/process usually takes less than 15secs. If you include the tests, the process takes around 30secs, so not that long.

As conclusions, I’ve had to learn a lot of new tricks and deal with many libraries/procedures with scarce documentation, and I do some things that are not actually recommended to do in real life projects, like download the dependencies directly from public repositories, but I don’t regret at all going through all of this “pain”, as I learnt quite a few things to apply to real projects.

I’ll just some fresh air for couple of days and document the installation a bit, in case someone wants to have a go at it.

I might try to write a series of posts, or a full article, out of all the things I learnt, like using HTMLUnit/Jetty embedded to test a web application, using HTMLUnit/JSON-lib to test AJAX/JSON requests, using Ivy to download the dependencies for all the different “steps” (8 different Ivy configurations are used)… but we’ll see if I can find the time to do so…

In any case, I hope you found this post interesting and that it gave you some ideas to try & test. BTW, I've heard and read about Maven, yes, but up to now my build scripts were not that sophisticated and I have mixed feelings about Maven, so I went with what I knew better: Ant.

I'd like also to take the time to thank sincerely all the people that work(ed) in all those open source projects, one could not even dream of doing such things if it weren't for your generoisity. I might moan and bitch sometimes about lack of documentation, bugs etc. ;) but that does not mean I do not appreciate it.

Cheers!

PS: Some of the code used in the web application is auto-generated from the db schema, which is also read automatically using DDLUtils… I wonder if I should also remove the generated artefacts from the repository and start from a real scratch… mmmm ;)
PPS: WebLEAF, the framework that this application tests could also be applied the same principles and if I joined both "from scratch" builds together... mmmm ;)
JON - Jboss Operation Network

Posted: March 30, 2008, 5:52pm CEST by Rafael Benevides

Jboss Operation Network ou simplesmente (Jboss ON) ou simplesmente JON é uma excelente ferramenta de monitoramento, controle e inventário do parque de servidores Jboss.

Tive conhecimento desta ferramenta através do curso JB336 (ministrado pelo Alexandre Gomes da SEA Tecnologia) do qual venho partipando aos finais de semana. Controlar um servidor de aplicação geralmente é uma atividade tranquila para o ASA, mas quando "a coisa" começa a crescer com vários servidores (hardware) contando com pelo menos 4 instâncias do Jboss AS, é inevitável a necessidade de uma "console de gerenciamento" única.

É neste ponto que entra o JON. Este é uma aplicação WEB que permite em sua console administrar todo o parque de servidores Jboss. Nesta console é possível monitorar e inventariar algumas métricas tanto do servidor físico (CPU, Memória, tráfego da rede, etc), bem como as instâncias de Jboss (JVM, Filas JMS, Datasources, Regiões de memória, Thread, etc). Algumas outras métricas também são possíveis de serem coletadas de servidores Tomcat, Apache e outros.

Além de mostrar um gráfico com o monitoramento de cada uma destas métricas, é possível também definir alertas e ações a serem tomadas. Por exemplo: Caso o pool de conexões seja consumido em até 90%, envie um e-mail de alerta e aumente a quantidade máxima do pool em mais 10 conexões. Todos o histórico de alertas, fica disponível na console do JON, permitindo que análisar no exato momento da ocorrência quantas conexões haviam com o servidor, quantas Threas e qual a quantidade de memória e CPU estavam sendo consumidas no momento em que ocorreu o alerta.

Além do inventário e monitoramento, existe também a funcionalidade que achei melhor: A possibilidade de se controlar o servidor de aplicações a partir do JON. Ou seja, é possível fazer ou agendar deploys de aplicações; gerenciar pools e filas; parar, iniciar ou reiniciar o servidor sem a necessidade de sair da interface do JON.

Além da "console WEB", o JON também fornece um shell onde é possível realizar as mesmas atividades sem a necessidade de um browser (geralmente muito útil para quem estiver acesso a máquina, apenas via SSH)

Ao contrário da maioria dos produtos Jboss que são disponíveis gratuitamente via Jboss.org, este produto só é disponível apenas através da subscrição de suporte.

Uma demo do JON pode ser encontrada em: [https:]
StartUp Engineering-TheHardestDecision

Posted: March 30, 2008, 5:24pm CEST by Fred Grott

The hardest decision in start up engineering is not a technology implementation. Its the decision to change a technology implementation to deal with market realities to get a prototype out faster and users using that prototype. Its the type of decision where everyone may say no and you do not get paid.

Its the type of decision that makes some developers walk out. Its a decision that you stay up for days analyzing. It is also a lonely decision, it cannot be shared and no one else can make for you. But is the decision that makes a CEO or CTO. Jason's right there is a time when you have to be bad ass, all the freaking time!
Mumbai Agile 2008 Conference Finishes Great!

Posted: March 30, 2008, 1:57pm CEST by Dhaval Dalal

Wow...indeed for us in Mumbai, this was a 2^nd great conference in Mumbai after 2005, there were around 125+ participants from various companies across the country, and not just from Mumbai and Pune. All the sessions were quite interesting and delegates have benefited from them. This is quite clear from the retrospectives that we have had and I am documenting them, so that we are reminded of things to improve upon for the next time around.

Retro Day 1 Retro Day 2

Conference left its mark with the delegates and the speakers alike...At the outset, I would like to thank, our host Mukesh Patel School of Technology Management & Engineering, all our sponsors (BNP Paribas India solutions, Mumbai and TietoEnator, Pune) and Supporters (Agile Alliance, CSI, Directi, ThoughtWorks, PMI - Mumbai). Kudos to all the student volunteers. Here is our group picture after the retrospectives...
Left to Right: Angela Martin, Student Volunteers, Naresh Jain, Elizabeth Keogh, Pramod Sadalage, Ahmed Sidky (sitting), Self and Student volunteers...
NULLs Last: a lasting problem

Posted: March 30, 2008, 1:26pm CEST by Tiago Silveira

Last year I had very little time to work on Tiny Marbles. When I came back from vacation in January, I saw the perfect opportunity to work on a very important issue that wasn't solved completely: sorting. So I sat there for a few hours hacking away. We had broken compatibility with HSQLDB, so I had some failing unit and system tests that would tell me when my new solution worked. They were created back when the first implementation was written, then adapted to the implementation I was going to change, which was the second. It was pretty much a matter of getting the full green bar for all 500+ tests, of which less than 10 were red. The story is long and is posted in the Tiny Marbles blog.

So I got that green bar after eventually. Then, of course, I went to test it back on MySQL (I have the version 4.1 on my laptop). Bummer! I had a filter for a set of 5 objects, in MySQL it gives me the order "1, 4, 2, 3, 5", but HSQLDB returns "4, 1, 2, 3, 5". Comparing objects 1 and 4, I stared, jaw-dropped, at yet another poorly defined issue with relational databases: different RDBMSs put NULL values differently, either before or after the non-null results. Check this out:

PostgreSQL – Nulls are considered HIGHER than non-nulls.
DB2 – Higher
MSSQL – Lower
MySQL – Lower
Oracle – Higher

Please add "HSQLDB – Higher" to that list. And there's no standard to force them either way in a cross-database approach. Hence the title: sorting NULLs is the perfect example of an abstraction leaking through the ORM into the business logic; everybody has to think about their SELECTs, know in which database they will run their system, and if the NULLs come in the wrong place, add database-specific flags to invert the order. For Hibernate users, even that won't work, because Hibernate mysteriously removes the "NULLS LAST" segment.

Since Tiny Marbles now stores the sorting key as a string, we could make sure that nulls come first using the empty string to encode null values, but then the user wouldn't have the choice to make them come last. The Hibernate folks are taking the don't-complain-submit-a-patch-or-pay-commercial-support-you-wimp approach on this, so you bet this won't be solved any time soon.

Although this is a problem that affects a minority of the Tiny Marbles audience, when it does, it is pretty disastrous. So we'll need to come up with a solution, but I don't have one yet. Has it ever bitten you?

Bug in YouTubeG library (searching a video by its video ID)

Posted: March 30, 2008, 10:34am CEST by Farooq Ali

I just dug up a neat little gem, YouTubeG, which wraps the YouTube REST API. You can lookup videos with interesting queries, like these:

Basic queries

client.videos_by(:query => "penguin")
client.videos_by(:tags => ['tiger', 'leopard'])
client.videos_by(:categories => [:news, :sports])
client.videos_by(:categories => [:news, :sports], :tags => ['soccer', 'football'])
client.videos_by(:user => 'liz')

Standard feeds

client.videos_by(:most_viewed)
client.videos_by(:top_rated, :time => :today)

More complicated queries with boolean logic

client.videos_by(:categories => { :either => [:news, :sports], :exclude => [:comedy] }, :tags => { :include => ['football'], :exclude => ['soccer'] })

Funny enough, the most basic query, finding a video by its ID doesn't work for me. I managed to fix the bug by changing YouTubeG::Client#video_by to look like this:

def video_by(video_id)
  request = YouTubeG::Request::VideoSearch.new(:video_id => video_id)
  parser = YouTubeG::Parser::VideoFeedParser.new(request.url)
  parser.parse
end

Perhaps you should monkey patch it until someone checks in a fix.

An interesting discussion about Data Portability

Posted: March 30, 2008, 9:03am CEST by Phillip Rhodes

If you've been following the whole Data Portability scene with any interest, there's an interesting discussion going on here that is worth reading / participating in. Obviously there's still a lot of disagreement about certain aspects of this whole thing, but hopefully things will coalesce sooner than later and the discussion will become even more pragmatic. On a related note, I'm tentatively scheduled to present on the subject of Data Portability at next month's Refresh The Triangle event, so any Triangle area folks who are interested, please do come out and take part. In the mean-time, I have to do some catching up with what the current areas of focus, themes, patterns, etc. are among the players in this space. I'm a little behind on reading the DP list, so it's time to get back into that conversation a little more earnestly. And even more so if OpenQabal is going to implement "Data Portability" (which, of course, depends on what that turns out to be!). Also, the technical documentation section of the dataportability.org site has some interesting reading for anybody who's trying to understand what this whole discussion is about. In particular, see the use cases section.
GridGain: Creating Open-Source Pressure

Posted: March 30, 2008, 7:44am CEST by Nikita Ivanov
In the last 7 months since GridGain Systems came out of the hiding with its project two prominent companies had to introduce some free initiatives to counter-weight the pressure that we have created (at least partially).
I see it as a clear validation of the business model behind GridGain project. In all the honesty, it really didn’t need any validation as it was already largely validated by companies like JBoss, SpringSource and MuleSource - whose business models are essentially identical of GridGain’s one. But it is remarkable that two commercial-only grid-related companies were so fast to introduce some free initiatives: GigaSpaces introduces free startup program (free use for companies with revenue up to $5M but no open source) and Appistry introduced Community Edition (can be used free up to 10 cpus but it's not open source again). I think these types of marketing exercises are too late for 2008 as most of the developers and managers are well informed and educated by now about open source and the rest. But I leave it up to the respective companies.

For all of us here at GridGain the open-source is first or foremost a business model that we believe fits enterprise software middleware perfectly. I have explained this point many times and it really comes down to these simple principles:
- Middleware (including grid computing) doesn’t solve any single business problem – it is a tool that customer will use to hopefully build an application that will solve some of its business needs. Thus charging upfront license fee is counterintuitive for something that will only potentially benefit the customer in 6-12 months – thus GridGain is open-source software and free without any strings attached. Charging for support when this application does go in production, on the other hand, is totally logic and expected by the client.
- Open source (LGPL or Apache) is essential and a must for middleware software. On multiple occasions I learned a critical piece of information from JBoss source code or Mule source code instead of lengthy and unproductive support escalations. It is a huge time saver and a knowledge base available to me as a user of that middleware - among many other intrinsic advantages. Remember the axiom: support is for production issues NOT for development issues.
- Commercial support (indemnification, patches, and production issues), professional services and training in general benefit customers right away and are offered on commercial basis.
GridGain Systems runs on this traditional professional open-source model and we had it from the start. We didn’t need to open-source something that doesn’t sell and we didn’t need to cave to VC pressure early on in product development phase. I think these factors benefited the project immensely as we were able to concentrate on the software and the first customers we had. And we are not along or the first in this category – looking at MuleSource, for example, I see what GridGain can become in couple of years.
Enjoy grid computing!
Back to Progress Software in Hyderabd/ driving the SOA & Data integration products work

Posted: March 30, 2008, 6:20am CEST by Ramesh Loganathan

Early this month, I took over a new position, as the MD & VP-Products, at Progress Software (India).

Progress in Hyderabad is three years old now (as of April 1). We started off with the Sonic Workbench (IDE for building ESB applications)- actually the work started in September 2004 itself (in Pramati). Went onto become the first team at Progress Software's captive development center in Hyderabad. Formally started on April 1. I was driving the Sonic Workbench development work at that time, until the first release in April 2006. Two years now. In the meantime Progress changed from a pure resource augmentation model, to few products independently owned in 2007, to now in 2008 a competencies based model with few products in each competency. IN this model needed all the work to be driven and managed locally, with dotted line relationship to the product line management. I had the extreme privilege of being considered for the role. After many months of deliberation finally accepted the position. Some more months to transition. And finally on March 7th joined Progress Software India.

Progress India is now aligned around a Centers of Excellence based models. Initial COEs being around Dev Tools, Management tools, Adapters and Automated Tools QA. Additionally work is happening on .NET tooling and also pubs & support. All teams have independent ownership of product(s) or a component. Thru this year these groups will be aligned across common product goals including the integration and better user experience across products. This is important now given that Progress is offering most of its products as part of the SOA Suite (and some also in the Data Integration stack).

One other mandate is to add some technology vigor into the regional space, Progress being in the leading edge of technology space (even now we have products in Event-driven SOA space, Complex event processing, SDO and other new data abstractions in the data integration space and even a legacy access is more currently by building event abstractions into the 'DataDirect Shadow' mainframe access product). We will be more active in conferences and seminars circuit, and also see a more of Progress Architects in India writing about our technologies.

Many of my other indulgences are aligned with Progress regional goals. Including education institution involvement, student internship (for technology exploration), technology evangelism, encouraging innovation, and also getting new technologies into the Indian market place (even a report this week from SpringBoard suggested that the Indian IT *software* market is expected to cross $10 billion!)

In all, am looking forward to a great time at Progress. (Hopefully can replicate my great 8 years with Pramati- that was both greatly enjoyable to me, and hopefully useful to the company as well).

Sonic ESB & DataDirect (drivers, and more) are two key products in the Progress suite.
.
What is SOA ?

Posted: March 30, 2008, 4:15am CEST by Nanik Tolaram

What is SOA ?
Groodle updates & screencasts

Posted: March 30, 2008, 3:36am CEST by Andres Almiray
When I wrote Groodle #3 I mentioned that JavaFX had a better syntax to enable and handle dragging of a shape, GraphicsBuilder's version is definitely more verbose but allows you to drag shapes, images and groups. Well not anymore, now all displayable nodes have a new property autoDrag, that when set to true will allow dragging of said displayable. The actual code is exactly the same but it allows you to set additional behavior and react to a displayable being dragged. Groodle #3 can be updated as follows
// goatee triangle( x: 60, y: 210, w: 40, angle: 180, f: 'brown', id: 'goatee', autoDrag: yes ){ filters { dropShadow( name: 'shadow', enabled: no ) } } ... def startDrag = { e -> def shape = e.target shape.borderWidth = 3 shape.borderColor = 'red' shape.filters['shadow'].enabled = true } def endDrag = { e -> def shape = e.target shape.borderWidth = 1 shape.borderColor = 'black' shape.filters['shadow'].enabled = false }
I also updated the code to follow Joshy's version, now that dropShadow and reflection (from Swingx) are also available as filters, I can finally render a shadow whenever a piece of clothing is dragged. Instead of creating a new filter each time (as in the jfx version) the shadow filter's enabled state changes according to the situation.

Screencasts of all Groodles are available too (thanks to vnc2swf, I love being back to Linux)
Keep on Groovying!
RichFaces 3.2 - File Upload

Posted: March 30, 2008, 12:37am CET by Mert Caliskan

Since 3.2 version of RF framework will be out on Monday (31.03), I thought it'd be good idea to give the file upload component a shot. I used the CR7 release from [repository.jboss.com] repo. Here is the screenshot,

Actually it looks very nice and has some neat features. It supports multiple file uploads, you can customize the look with your own CSS, it has the support for i18n and it has a nice progress bar for showing out the process.

Here are some of the properties of the component:

maxFilesQuantity: To specify the file # to be uploaded and visual disables the add button when that # has been reached.

acceptedTypes: You can specify it for example like "gif, jpg, jpeg, png". Of course user can select other types (say doc, pdf blah) but they aren't gonna be added to the file upload list which is fair enough.

autoclear: For clearing the upload list after the progress.

fileUploadListener: Listener method that will be called after each upload for handling the bytes[].

Filter configuration for handling the multipart post is delegated to the org.ajax4jsf.Filter. old CR releases were using the Seam's MultipartFilter. You can specify parameters named maxRequestSize, for handling the max. number of bytes that will be uploaded, createTempFiles, for creating temporary files on disk or not, to handle the large files.

I also looked around for a reRender property but it wasn't there. It'd be great to have that ability with the component though. For now, I got the same result with some JS work.

You also don't have to specify enctype for the h:form while using the component. It enables that via JS.

And I bumped into a show stopper when I used MyFaces Extensions Filter with RF. File Upload simply doesn't work with that filter-mapping declared, so I had to remove it.
OpenSocial Foundation

Posted: March 29, 2008, 10:30am CET by Phillip Rhodes

Old news by now, but I wanted to comment real quick on the announced OpenSocial Foundation. In light of my previous comments about the status of the OpenSocial specification(s), I am happy to see this: Today we are pleased to announce that Google is joining together with Yahoo! and MySpace in the creation of a non-profit foundation for the open and transparent governance of the OpenSocial specifications and intellectual property. This foundation, modeled after the community-led and industry-supported OpenID Foundation, will seek to ensure that the technology behind OpenSocial remains implementable by all, freely and without restriction, in perpetuity. The establishment of the foundation is a necessary and exciting next step toward an open model for application syndication via an increasingly socially-enabled Internet, and we invite and encourage your involvement as we complete this process over the next several months. I'm not sure about the creation of a new organization to do this work,as opposed to working with an existing standards body, but I suppose this is - at minimum - a positive step in the right direction. At best, it means we can put fears about the status of the OpenSocial spec(s) to rest. More info at: Google
TechCrunch
CNet
Mashable
ReadWriteWeb
CenterNetworks
GigaOM
ProgrammableWeb
FaceReviews
EWeek
InformationWeek
InfoWorld
ComputerWorld
CIO Insight
New OpenQabal Release Coming Soon

Posted: March 29, 2008, 8:23am CET by Phillip Rhodes
The last month has been an interesting time for OpenQabal. Thanks to my previous employer graciously laying me off from my job about 5 weeks ago, I've had the opportunity to spend most of the last month essentially working on OpenQabal full-time. A lot of progress has been made, and I anticipate making a formal 0.0.3 release soon, probably within the next week. Some of the changes that will appear (this does not, btw, map exactly to the previous road-map. As they say, "planning is essential, but a plan is useless.")
- New, more modular build system using Ivy for dependency management and incorporating useful tools such as FindBugs, JDepend, PMD, TestNG, Cobertura, etc.
- New "User Dashboard" component that provides the main point of entry and ties the various components together visually
- Concordantly with the introduction of the new User Dashboard, all of the old Sitemesh stuff has been ripped out
- New configure and install scripts to automate most of the tedious parts of building and installing OpenQabal
- Addition of many unit tests (we still don't have 100% test coverage, but progress has been made)
- Various minor bug fixes and tweaks
- Probably some other stuff I'm forgetting
Unfortunately a couple of things that I'd hoped to get in a 0.0.3 release are going to have to be deferred, with the most notable being OpenID login support. However, a subsequent 0.0.4 release should be around in relatively short order, and will include OpenID and - if things go well - a number of other big improvements. Unfortunately I have to go back to work for "the man" now so that I can pay my rent, so I won't be able to spend quite as much time on OpenQabal. But things will continue to move forward. There may actually be some even bigger news in the near future, so stay tuned.
Discotheek 1.0 released

Posted: March 28, 2008, 10:52pm CET by Anthony Goubard
Today I've released Discotheek 1.0.

Features:
- Transform your computer in a color spot light
- Transform your computer in a stroboscope
- Transform your computer in a light
- Vary the speed of the light flashes
- Support multiple screens
- Fade colors between colors
- Accept command line arguments
- Works on Windows, Mac OS X, Linux
- You can run it online as Applet here
Screenshot
WingS 3.2 released

Posted: March 28, 2008, 10:37pm CET by Andres Almiray

The wingS development team just announced that version 3.2 of the WingS Framework has been released.

Wings is a web application framework based on Java Servlets, resembling the Swing API with its MVC paradigm and event oriented design principles. It utilizes the models, events, and event listeners of Swing. Like in Swing, components are arranged in a hierarchy of containers, whose root container is hooked to a frame.

There are many improvements related to performance, dialog management and JavaScript, but the feature that I find to be most compelling (at least for me) to try out in this release is PropertyChange support, which means I can add bind() to WingSBuilder. Hope the wingS team add the jars to maven's central repo soon.