Java Tales from a Trading Desk

I've moved to WordPress

2005-12-22T13:49:00.000Z

I'm now blogging here.

FIT everywhere

2005-08-24T08:43:00.000+01:00

XPlanner - Could not execute query

2005-08-13T19:45:00.000+01:00

Have seen this a few times now, anyone got any ideas?

javax.servlet.jsp.JspTagException: net.sf.hibernate.JDBCException: Could not execute query at com.technoetic.xplanner.tags.db.UseBeansTag.doEndTag(UseBeansTag.java:133) at org.apache.jsp.index_jsp._jspService(index_jsp.java:83)

XPlanner Install Problems

2005-08-10T07:48:00.000+01:00

I got this stack trace after starting XPlanner on Tomcat 4.1.31 with MySQL:

javax.servlet.ServletException: java.lang.NullPointerException
at org.apache.jasper.runtime.PageContextImpl.handlePageException(PageContextImpl.java:495)
at org.apache.jsp.index_jsp._jspService(index_jsp.java:170)
at org.apache.jasper.runtime.HttpJspBase.service(HttpJspBase.java:92)

This wiki suggests adding the database settings to resource/xplanner-custom.properties. But that didn't seem to help either :( I found a few other ideas to try here and here.

Nice review of Synergy here.

Agile good practices.

When faster is wrong, and slow is right

2005-07-29T09:11:00.000+01:00

Gosling has an interesting post on the Java sin() and cos() methods - Sun's bug detail.

Simian, and some wacky results

2005-07-26T12:46:00.000+01:00

What do you do when you see this from Simian?

<summary duplicatefilecount="2286" duplicatelinecount="229743" duplicateblockcount="16659" totalfilecount="3684" totalrawlinecount="1472768" totalsignificantlinecount="815444" processingtime="33657">

I suspect the answer is ...run

Agile 2005 Conference

2005-07-25T08:32:00.000+01:00

A colleague is blogging from the Agile 2005 conference this week.

TDD with Velocity/xmlunit

2005-07-21T21:26:00.000+01:00

A component I am working as part of a pricing system needed to generate HTML reports. A fellow consultant - from a rival consultancy :) - suggested I use Velocity. Velocity turned out to be an ideal fit for the problem I was solving. However, before I could write the code I needed to write the test - TDD. xmlunit was an ideal fit for the test cases - specifically the assertXpathEvaluatesTo method.

Evo - an old (1976) agile development methodology with 5 day iterations.

Know your software development tools

2005-07-19T08:15:00.000+01:00

As Kevin Johnson states in Extreme Programming Applied (page 174) "pairs will be stronger if they both discover the most efficient way to use their tools". He also notes that shortcut keys "minimize the friction caused by back seat drivers". Having migrated to IntelliJ a few weeks back (from Eclipse) I have to agree with Kevin's shortcut point - I can't learn the Idea's keymap fast enough.

Is this the start of Microsoft moving towards Open Source?

2005-07-12T12:18:00.000+01:00

OSDL boss hints at Microsoft collaboration

GridCache

2005-07-12T07:58:00.000+01:00

DataSynapse GridServer has the concept of a GridCache. Only problem with GridCache is that it has no locking, no transactions, and is not a distributed data repository. Maybe DataSynapse should consider integrating Blitz into their product. GigaSpaces on the other hand has a distributed cache, but doesn't (I could be wrong) offer all the grid services that DataSynapse offers. Maybe DataSynapse and Gigaspaces should merge :)

Unit Testing Database Access code

2005-07-12T07:41:00.000+01:00

HSQLDB appears to me to be a clean way of testing database access code. Just use the in-memory feature (jdbc:hsqldb:mem:), and implement your setUp method to create the appropriate tables, and tearDown to drop the tables.

Interesting to see that the Deutsche is listed as one of the openadaptor adopters - DrKW originally build openadaptor. Suprised no US investment banks are listed.

Came across Nagios today. Nagios is a monitoring program for host, service and networks. Nagios is fully pluggable, with the Nagios Exchange offering a selection of plugins. NSJS looks particularly interesting.

JBehave

2005-07-09T07:35:00.000+01:00

A colleague mentioned JBehave yesterday. Not knowing much about the project, I googled and came across Magpiebrain's blog. The Velvick Underground reflects on the BA's view of JBehave, as does the Agile Business Coach.

Damian has a good overview of Grid Computing, particularly relevant to the project I'm now on.

Another day, another client

2005-07-08T08:21:00.000+01:00

Having just finished working on a large Foreign Exchange (FX) project at a US client of Finetix, I'm now at a European bank (for Finetix) working on Grid Computing. The grid project is particular interesting for two of reasons:

DataSynapse is the application we are working on
Agile principles

About six years ago I wrote a grid computing framework and application. Unfortunately, neither saw the light of day. However the project did give me a good grounding in grid applications, so the new project I'm working on is very much a flash back to past experiences.

With regards to Agile, I'm a support of the Manifesto. This particular project has a number of ThoughtWork consultants attached to it, giving the project a real Agile flavour, not the more typical investment banking half baked Agile approach. Today out development environment is IntelliJ, Cruise Control with Emma, CheckStyle etc

Off-heap data management using memory mapped files (Java NIO)

2005-05-03T14:12:00.000+01:00

Tangosol

Bytecode of method variables

2005-04-29T10:05:00.000+01:00

Having installed the Bytecode Outline Eclipse plugin, and read the online instruction reference it would appear that the following two code extracts produce similar bytecode:

for (final java.util.Iterator iter = categories.iterator(); iter.hasNext();) {
final String s = (String)iter.next();
}

String s=null;
for (final java.util.Iterator iter = categories.iterator(); iter.hasNext();) {
s = (String)iter.next();
}

ALOAD appears to be used in both cases, implying that there is no performance advantage gained by moving the String outside the for loop

Why the SPARC architecture is dead

2005-04-01T07:22:00.000+01:00

The project (Java application, consisting of a number of server processes) I work on currently uses a Sun Fire 15K server in production (32G RAM, 12 CPU). Today we decided to test the same application on a Sun Fire v40z server (AMD Opteron 4 CPU, 32G RAM).

Running the same performance test on both server configurations, we found that the Opteron server was 3x faster than the SPARC server. Another group have done a similar server comparison but using there multi-process multi-thread C++ server and found a 3-4x speed improvement.

It's quite interesting to compare the cost of both systems - 15K vs V40z.

Given the performance improvement and cost savings, it's difficult to see where Sun is going with the SPARC architecture.

Oracle JDBC Drivers

2005-03-10T09:27:00.000Z

It's implied from the Oracle web site that if you are using JDK 1.4, then you should use ojdbc.jar. Anyone know if there are issues using classes12.zip (Oracle site implies classes12.zip is for JDK 1.2 and JDK 1.3) with JDK 1.4? A quick Google search return this issue.

Why can't Sun (Java) do this as well?

2005-03-10T07:23:00.000Z

Blog Day from the Microsoft BCL Team

TCP_NODELAY - Nagel algorithm

2005-02-15T09:17:00.000Z

The Nagel algorithm has the potential to influence performance - this article talks a bit about the algorithm

Who started my defunct processes

2005-02-03T15:18:00.000Z

/usr/proc/bin/ptree `ps -ef | grep -v ^\ \ \ \ root | awk '// {print $2}'` | less

Finally, OpenSolaris.org goes live

2005-01-26T07:12:00.000Z

Today we can only download the source to DTrace, but there is a hint of things to come from the home page; "Soon, you'll be able to download the OpenSolaris distribution.". With the OpenSolaris release we are finally seeing Sun developers blogging source code - Adam Leventhal and Bryan Cantrill are just two examples. Calvin Austin's recent leaving sun blog entry hinted at problems within Sun with regards to blogging source code - which is a shame since the Java 5.0 source has been available for some time. Hopefully soon we might actually being to see more JVM Sun bloggers in the same calibre as Adam and Bryan, together with a blogroll of JVM developers equivalent to the Solaris Kernal Developers.

Solaris Internal Links

2005-01-14T11:39:00.000Z

Load Averages Explained
Performance Analysis and Monitoring Using Hardware Counters
prstat
Solaris Internals and slides
Hows to check your box is running 64-bit: isainfo -kv
Check Solaris patch levels: showrev -p
LWP
Managing Thread Execution and Wait Times

A Comparative Study of Persistence Mechanisms

2005-01-06T13:12:00.000Z

An interesting article from Sun Labs. If anything, the conclusion is worth a read.

"Acceptable EJB performance seems unattainable at present unless dramatic changes are made to the application object model to avoid fine-grain objects when mapped to EJB. While this approach is now reflected in standard design patterns for EJB, the extra effort that it implies is disturbing from the overall application design perspective. In contrast, JDO manages to achieve reasonable performance at much lower impact on application design, while remaining agnostic to the nature of the external data store. Therefore, at this time, JDO would seem to offer the best overall persistence mechanism for demanding, object-oriented, applications. Note, however, that at the time of writing, a new specification for entity bean persistence [Sun04] was being proposed for EJB 3.0 that would bring it much closer to JDO in spirit."

MythTV

2005-01-06T07:12:00.000Z

Now this looks cool. All I need to do is fine time to install it on my Linux box.

New Year, New Threading Model

2005-01-04T11:32:00.000Z

Actually its the Alternative threading model (T2) on Solaris 8. Over the Christmas period we finally tested our application with Solaris's alternative threading model. Overall we are seeing a 30% improvement in performance, with a 3x reduction in context switch activity, system CPU usage is also down.

To answer the question of why did it take us so long to move to T2 - well I currently working on a project for a large corporation, and they have only just moved our production hardware from Solaris 6 to Solaris 8.

Solaris 10

2004-12-03T22:26:00.000Z

So I finally installed Solaris 10 build 69 on my Blade 100 - and the next day Sun releases build 72 :( The main reason I installed Solaris 10 was to play with DTrace. I'm particularly interested in DTrace and jstack, although there currently doesn't seem to be a lot of information about jstack - Calvin's blog being one source.

I also need to install some more RAM in my Blade 100 - time to surf over to crucial.

The Blade came with Solaris 8 pre-installed, so this is the first time I've ever installed Solaris. Overall the install is simple, but I have to say, finding the ok prompt and typing "boot cdrom" was not really what I was expecting to have to do. The Java Desktop System is a massive improvement over the Common Desktop Environment (CDE), although I don't quite understand why Sun didn't add an option to the GNOME Launch menu for the Solaris Management Console.

Things that appeared broken after the install included the Java Media Player doesn't appear to work, and clicking Java in Desktop Preferences gives "Cannot launch entry" since the command was pointing to /a/user/java/jre/bin/ControlPanel when it should have pointed to /user/java/jre/bin/ControlPanel. Another problem I have is that I used the Solaris Management Console to create a new user. If I login as this user using CDE or the Java Desktop System, I never actually manage to login, and always end up back at the Welcome screen. If I login using the Command Line option, I can login fine. Anyone know what log file I should look in to try and resolve this problem?

Update: It appears the reason the new user I created couldn't login, was because of a corrupted .dtprofile. Taking a copy of roots file fixed the problem, and I can now login to the Java Desktop System using the newly created user.

DTrace and Java

2004-11-25T08:05:00.000Z

Calvin has a nice blog entry on DTrace and Java. Shame you have to run Solaris 10 and J2SE 5.0 :( jstack and DTrace are definitely an interesting combination. Adam also has a nice blog entry on this stuff.

String.intern()

2004-11-19T07:23:00.000Z

The current project I'm currently work on doesn't currently make use of String.intern() - which is probably not that surprising since part of the codebase dates back from 1998. Genady has a good overview of the inner workings of intern(). Assuming you agree to the license over at the Sun Community Source License Download, you can download the appropriate JDK source code and look at the VM implementation. I have to admit I found the source very interesting, especially how the C++ classes have changed from JDK 1.3.1 to 5.0. One thing that surprised me was that intern() is a native method, I'd expected intern() to be implemented in Java and use a HashMap or similar. It's a shame that Sun's JavaDocs don't identify which methods are native.

Sun's memory allocator article made me wonder if the Java VM uses a special malloc for multi-processor machines. The JDK 5.0 appears to perform memory allocations via a macro that wraps os:malloc. Does anyone have any information on this?

HotSpot Source Code

2004-11-17T08:13:00.000Z

It's nice to see that the Java Research License (JRL) source download includes the source code to the HotSpot Client and Server VM's. Assuming the JRL HotSpot is identical to J2SE 5.0, it's a cool download for the developer who wants to know what the VM is doing, and how its doing it. It's a shame Microsoft's Shared Source Common Language Infrastructure (SSCLI) doesn't follow Sun's openness - maybe the up and coming Rotor release will fix this

As blogged elsewhere on the web, its nice to see JDK 6.0 snapshots are available.

MDA and MDD

2004-10-28T08:58:00.000+01:00

Looks like we are moving down the Model Driven Architecture and Model Driven Development road. Yesterday everyone on my development team had a quick training session on Borland Together. We are using the Eclipse edition. The reverse engineering features of the product are very cool, although I'm not convinced yet about the Pattern Recognizer. The Quality Assurance metrics and audits look interesting - hopefully you will be able to run these from the command line, and generate them via the nightly Ant build.

Database Usage

2004-10-27T10:32:00.000+01:00

This is an old blog entry by Werner, but is still relevant today. I'd be curious to know if anyone has any information on how these sites have evolved since 2003.

Gigaspaces and GemFire

2004-10-27T09:08:00.000+01:00

Gigaspaces and GemFire both appear to be off-the-shelf products that can offer caching (distributed). Both products implement JCache (JSR 107). GemFire offers the option of using an operating systems shared memory segment to improvement performance between processes on the same machine. Has anyone used GemFire? Any idea which is the better product?

C++ Migration to Java

2004-10-08T07:57:00.000+01:00

After 3 months of work, I'm close to being done with the re-architecture of three legacy C++ servers. The original servers were written back in the late 90's, and used an OO database as a repository. The re-architecture has involved moving and code to Java, and migrating off the OO database. What follows is a summary of the performance timings taken from a JUnit calling both the C++ and Java CORBA servers - these timings should be taken with a pinch of salt, since the internals of the C++ and Java servers are different (but produce the same results, and expose the same IDL interfaces), as is the repositories:

(11998 method calls) for Test 1
C++:Average milliseconds 21(total milliseconds:253846)
Java:Average milliseconds 1(total milliseconds:18438)

(11998 method calls) for Test 2
C++:Average milliseconds 26(total milliseconds:314209)
Java:Average milliseconds 3(total milliseconds:38745)

(5999 method calls) for Test 3
C++:Average milliseconds 40(total milliseconds:242797)
Java:Average milliseconds 1(total milliseconds:10684)

(5999 method calls) for Test 4
C++:Average milliseconds 53(total milliseconds:322152)
Java:Average milliseconds 11(total milliseconds:69485)

(10 method calls) for Test 5
C++:Average milliseconds 717(total milliseconds:7174)
Java:Average milliseconds 168(total milliseconds:1680)

(10 method calls) for Test 6
C++:Average milliseconds 101(total milliseconds:1019)
Java:Average milliseconds 25(total milliseconds:259)

(10 method calls) for Test 7
C++:Average milliseconds 81(total milliseconds:813)
Java:Average milliseconds 28(total milliseconds:287)

(813 method calls) for Test 8
C++:Average milliseconds 3(total milliseconds:3133)
Java:Average milliseconds 2(total milliseconds:2053)

(813 method calls) for Test 9
C++:Average milliseconds 16(total milliseconds:13629)
Java:Average milliseconds 1(total milliseconds:1237)

Timings were taken on 8 CPU (1200Mhz) Sun Sparc Solaris box running JDK 1.4.2_05 with Iona Orbix 6.1

Q: Why are there two ps utilities on my Sun box?

2004-10-06T13:47:00.000+01:00

Answer
Historically, SunOS 4.x was based on BSD unix. Solaris 2.x (= SunOS 5.x) was based on SYSV, with a bunch of commands having different syntax and behavior. To ease the transition, the/usr/ucb directory was created to hold the incompatible BSD versions. People who really wanted BSD could put /usr/ucb before /usr in theirPATH.

A New Version of Java

2004-09-30T21:48:00.000+01:00

On the day that Sun release Java 5.0 the project I am working on moves to Java 1.4.2_5 and Iona Orbix 6.1.

Who are the Sun JVM bloggers

2004-09-29T08:12:00.000+01:00

You can get a list of the Solaris kernel developers here. The Microsoft CLR team hangs out here. Can somebody point me at a list of JVM bloggers?

On the GC Tuning Road

2004-09-29T07:23:00.000+01:00

I'm code complete for the three Java CORBA servers I've re-architected over the last three months. I now just need to re-run the last performance tests, and fine tune the JVM options. Since I'm forced to run on the Solaris JVM 1.3.1_09 I initially started using this AWK script (later migrating to Perl) to take the -verbose:gc output and graph the results in Excel. JVM options currently being used:

-server -XX:+PrintTenuringDistribution -verbose:gc -Xss256k -XX:TargetSurvivorRatio=90 -Xss256k -ms512m -mx512m -XX:NewSize=192m -XX:MaxNewSize=192m -XX:SurvivorRatio=4 -XX:+DisableExplicitGC -XX:PermSize=16m -XX:MaxPermSize=16m

Joseph Mocker's recent blog on "A Collection of JVM Options" is a worthwhile read.

Sleepycat

2004-09-27T07:39:00.000+01:00

Late last week I was asked to look at BerkleyDB for possible use within one of our servers (JBossCache also make used of BerkeyDB). Certain team members thought that BerkleyDB might be a better fit that using Oracle. I've never used Berkley DB, but having done a few google searchs , and reading this article, BerkeyDB could be an interesting replacement for Oracle in our servers that needs a high performance transactional storage engine.

Finally, access to ClearCase

2004-09-24T10:43:00.000+01:00

After weeks of waiting, my machine has been setup for ClearCase, and I've got access to the projects ClearCase VOB (ClearCase version 4.2). I've never used ClearCase before, so checking in the last 2 months worth of work was interesting and slightly different to CVS. Supposedly if you can afford ClearCase, it is the best source control system available.

Essentially ClearCase VOBs use a proprietary database format, the Raima database. ClearCase allows a user to have either a snapshot or a dynamic view. In my case, I'm using a dynamic view. The view is essentially your view of the elements in the repository, allowing you to decide what branch, versions etc you want to see and work with. One of the most useful commands I've used in the last few days is :

cleartool lsco -cview -recurse

Essentially a way to check my view to see what files are I still have checked out. The next ClearCase exercise I have on the list is merging my branch with the main branch - I just need to re-run all my JUnit tests against what I have added to my ClearCase view and performance tested the code on the development Solaris server instead of my Windows XP dev box.

ClearCase CheatSheet
Eclipse ClearCase Plugin

Explain Plan - Oracle

2004-09-07T07:56:00.000+01:00

I never remember how to check out the execution plans on Oracle, so this time I thought I'd blog the answer. Basically, you run the following:

explain plan for

You can then see the results with:

SELECT * FROM TABLE(dbms_xplan.display)

This article gives a lot more detail, as does this. Its a shame Oracle didn't integrate the explain plan output into SQL Plus in a similar way to what Microsoft has done with SQL Analyser.

Annotations - one of the many differences between .NET and Java 5.0

2004-08-21T22:31:00.000+01:00

If you are a software engineer using both .NET(C#) and Java 5.0, there are some subtle differences between both metadata implementations - attributes/annotations. The first one you normally run into is the fact that Java's annotations are available at source level only by default. Try running the following code:

public @interface SomeAnnotation { String value();}

@SomeAnnotation("hello")
public class SomeClass {
public static void main(String[] args) {
System.out.println(SomeClass.class.isAnnotationPresent(SomeAnnotation.class));
}}

If you come from .NET land, you might expect the output to be "true", but what you actually get is "false". To make the annotation available at runtime, change SomeAnnotation to:

@Retention(RetentionPolicy.RUNTIME)
@Target({ElementType.TYPE,ElementType.METHOD})
public @interface
SomeAnnotation { String value();}

IntelliJ 4.5.1 and Java 5.0

2004-08-21T22:00:00.000+01:00

So I upgraded to IntelliJ 4.5.1 last night. As stated by other bloggers, IntelliJ has very nice support for Java 5.0. I'm going to start playing with Java 5.0's annotations tonight since the last project I worked on (C#) used attributes in a big way - attributes before the .NET equivalent of Java 5.0 annotations.

final Object[] aa = {"hello", 22};
I have to admit the autoboxing feature in Java 5.0 is nice - especially if you write Java and C# code.

for (String s : a)
It also feels natural to use the new enhanced loop - maybe again because is some ways its similar to what I write in C#

The New World

2004-08-20T18:09:00.000+01:00

I finally moved to Eclipse 3.0 today at work. I've installed the plugins are per the previous blog entry. Anyone got any recommendations for other must-have plugins?

Stuck in the Old World

2004-08-20T07:28:00.000+01:00

The current project is interesting, and yet painful. Interesting in that I am back in Java land after a year or so, but painful in that all the current work is on Java 1.3.1 (and Eclipse 2.1), and hence I'm missing out on the expanding world of Eclipse plugin's that make development life easier - Jupiter, FindBugs, JDocs, Profiler etc. With Java 5.0 out soon, I hope we at least move to Java 1.4 on the current project. At least then I can start using Eclipse 3.0 again.
In the last week or so I started using Hibernate. Very cool, easy to use/install, with good documentation. What more could you ask for.

The Sun-JDocs issue is sad. I have to agree with Dion with regards to the Microsoft comment.

Last night I upgrading my Fedora 1 box to Fedora 2. The upgrade went fairly smoothly apart from when I rebooted, and got an error message during Gnome startup. Gnome did start after I ok'd the error message. I just need to dig into some log file to discover and fix the issue.

Finally, one of these days I'm actually going to get time to install and play with Blitz.

10g

2004-08-05T08:01:00.000+01:00

Today I had reason to install Oracle 10g. I always find installing Oracle painful - maybe its just the initial slowness of the actual installer, or maybe its the size of the download. One benefit of 10g is that it offers access to the Oracle Enterprise Manager via a browser (http://localhost:5501/em by default). SQL Plus still sucks badly, maybe one day they will actually fix this crappy product.

Once 10g was installed, I knocked up a simple Java program to ensure I could access the database, and ended up with the classic JDBC error "java.sql.SQLException: ORA-00600: internal error code, arguments: [ttcgcshnd-1], [0], [], [], [], [], [], [] ". The problem as usual turned out to be that I had classes12.zip on the CLASSPATH instead of classes12.jar.

One nice feature (or maybe not a feature) of 10g was that I didn't seem to have the TNS setup issues I had in Oracle 8 - maybe this was due to the thin ("jdbc:oracle:thin:") connection string I was using.

Distributed Caches

2004-07-28T16:02:00.000+01:00

John Davies has an interesting comment on the TheServerSide with regards JavaSpaces and GigaSpaces - I'm guessing its BNP Paribas. Would love to know more details about the "seriously sexy architectures" he's be working on.

J2ME and my Nokia 7250i

2004-07-15T09:09:00.000+01:00

It's been a few years since I did any mobile work (a WAP web site was the last), so I thought I'd have a go at developing a simple application for my Nokia 7250i. Nokia's forums have a lot of good stuff available for developers. For starters the device details page for the 7250i gives details on what Java technology I can access - it appears I need to develop an MIDlet (max 64Kb) that I can then download Over-the-Air (OTA)

An MIDlet appear to be very to an applet in concept. The MIDLet class can be found in the javax.microedition.midlet.* package. Once you have coded your MIDlet application, you can use the Nokia Developer's Suite to generate the .JAR and .JAD (Java Application Description) files. The Nokia Developer's Suite also has an emulator which avoid round-tripping to a real phone. One issue I found with the Nokia Developer's Suite is that it comes with integrationg for Java Forte and JBuilder, but no integration for Eclipse.

The easiest way to get the MIDlet onto your phone is to place the .JAR and .JAD onto a web server, setup MIME tags for .jad and .jar, add a HTML page that links to the .JAD files, and use the Nokia phone to browse to the page, and install the application.

midlet.org and other sites on the web have a load of free applications that you can download.

J2ME code for "Hello World":

package com.First;
import javax.microedition.midlet.*;
import javax.microedition.lcdui.*;

public class First extends MIDlet
{
private MyCanvas canvas;

class MyCanvas extends Canvas
{
protected void paint(Graphics arg0) {
arg0.drawString("Hello World", 0, 0, Graphics.TOP|Graphics.LEFT);

}
}

public First()
{
canvas = new MyCanvas();
}

protected void startApp( ) throws MIDletStateChangeException
{
Display.getDisplay(this).setCurrent(canvas);
}

protected void pauseApp( )
{
}

protected void destroyApp( boolean p1 ) throws MIDletStateChangeException
{
}
}

.JAD file for the above application:
MIDlet-Name: MyMIDlet
MIDlet-Version: 0.0.1
MIDlet-Vendor: MyCompany
MicroEdition-Profile: MIDP-1.0
MicroEdition-Configuration: CLDC-1.0
MIDlet-Jar-URL: MyMIDlet.jar
MIDlet-Jar-Size: 2301
MIDlet-1: First, , com.First

Eclipse Rich Client Platform

2004-07-14T21:31:00.000+01:00

Given that the client is back, the concept of using Eclipse as a rich client platform offers some interesting ideas. I know of a few development teams within tier-1 banks that are considering this Eclipse approach to consolidate exiting Java Swing rich client applications - at least SWT looks a lot better than Swing (a Swing application never looks quite right, irrespective of the hours/days spent on it).

Typesafe Enum etc - Pre Java 5.0

2004-07-14T21:04:00.000+01:00

Here are a few comments having jumped back onto the Java bandwagon after spending the last 15 months using Microsoft .NET:

Move from C# which has enum's to Java which doesn't is a real pain, even Bloch's Typesafe Enum pattern doesn't help in the context of switch statements. I'm amazed Sun never added enums prior to 5.0.

Events in C# are far easier than in Java. Maybe Sun can fix this in Java 5.1

Eclipse is a really nice IDE these days. The debugging experience is finally enjoyable. There is even a plugin for ClearCase

The Java class library is much richer than C# with regards to collections.

Bruce Eckel

2004-07-10T10:37:00.000+01:00

TheServerSide had a link to Bruce's latest blog entry. The video is worth watching, since he makes some very valid points - I have to agree that the classpath is a nightmare. He also talks about Java vs .NET, and J2SE 5.0.

Garbage Collection

2004-07-09T21:51:00.000+01:00

Nice article on tuning the garbage collection for a real application. It occurred to me today that tuning the Java GC appear to be more of a black art than tuning the Microsoft CLR GC.

DTrace appears to be getting a lot of press. There is a Flash demo of DTrace available here. Nice to see Bryan and Adam are blogging.

GregorianCalendar and Foreign Exchange

2004-07-09T18:25:00.001+01:00

Calendars are esential in trading systems. For Foreign Exchange (FX) system in particular, calendars are very important since if a trader is executing an FX currency pair GBP/USD trade, they'll need to know the date of Spot (http://www.fx-forex-trading.com/glossary.htm#S) and other relevant tenors; Spot Next, Spot Week, x Day, x Week, x Month, x Year. Examples of tenors in short hand form are: 1W 3W 3M 1Y.

So what does the above have to do with java.util.GregorianCalendar? Well actually a lot. Since to work out spot for a currency pair, you first need to figure out what the first business day is after the transaction day - the day the trade was executed. A business day is not a weekend, and not a holiday, so for GBP/USD, we need to look at the holidays in both calendars. Assuming today is 9th July 2004, then the next business day would be 12th July 2004, which means that spot will be 13th July 2004 (assuming settlement within 2 business days of the transaction date).

A description of FX value dates can be found here.

GregorianCalendar and Foreign Exchange

2004-07-09T18:25:00.000+01:00

Calendars are esential in trading systems. For Foreign Exchange (FX) system in particular, calendars are very important since if a trader is executing an FX currency pair GBP/USD trade, they'll need to know the date of Spot and other relevant tenors; Spot Next, Spot Week, x Day, x Week, x Month, x Year. Examples of tenors in short hand form are: 1W 3W 3M 1Y.

So what does the above have to do with java.util.GregorianCalendar? Well actually a lot. Since to work out spot for a currency pair, you first need to figure out what the first business day is after the transaction day - the day the trade was executed. A business day is not a weekend, and not a holiday, so for GBP/USD, we need to look at the holidays in both calendars. Assuming today is 9th July 2004, then the next business day would be 12th July 2004, which means that spot will be 13th July 2004 (assuming settlement within 2 business days of the transaction date).

EJB 3.0

2004-07-04T19:52:00.000+01:00

I have to agree with the list of common complains in this article, especially the 4 files per EJB - at least XDoclet makes life easier today. It's nice to see that Spring and other light weight containers are going to have an influence on EJB 3.0. Hopefully EJB 3.0 will be heavily influence Java 5.0 Metadata

Java Corba

2004-07-02T22:23:00.000+01:00

It’s been a while since I did any CORBA work - that last time was probably back in the late 90's with the Iona product. So what am I doing with CORBA? I'm re-architecting a few C++ Solaris applications that use the Versant database. The replacement servers will use Java 1.4 and an Oracle database. Its kind of strange to read the legacy code, with the TIE and narrow().

JavaOne - "The client is back"

2004-06-30T11:50:00.000+01:00

Its funny how in the software industry go round and round. Schwartz's comments about the client being back are interesting. Maybe Microsoft .NET taking a share of the client market from Java has influenced this comment - .NET WinForm's do appear to be going places inside investment banks. Not long ago rich clients where supposedly dead and the browser was king.

Looking Glass looks cool. It will be interesting to see how the project evolves, and how it stands up to Microsoft's Longhorn Avalon technology.

Java Performance

2004-06-30T10:48:00.000+01:00

Java Performance Tuning by Jack Shirazi is an excellent book - ok I have only read the first 100 pages or so. It's the first book I've read for a while that contains some interesting geeky performance/jvm internal stuff. There is also the well known web site that supports the book.

The Java GC Portal article provides a nice utility to mine the verbose:gc logs, as does Ken Gottry's AWK script.

The is a load of information in Improving Java Application Performance and Scalability by Reducing Garbage Collection Times and Sizing Memory Using JDK 1.4.1

With regards J2SE 1.5, I mean J2SE 5.0 (why did they leave the 2 in J2SE ?), it would be nice to get some idea of how they managed to get the improvements in startup time and memory footprint.

JFluid appears to be moving along nicely

Java Desktop

2004-06-29T15:47:00.000+01:00

Allied Irish bank moves to Linux and Java Desktop

JVM Internals

2004-06-29T09:45:00.000+01:00

I've been looking for blogs on JVM Internals for a while and came across Moazam Rajas blog last night. Anyone know of any other good source on JVM Internals - apart from the Sun site?

Hibernate and Eclipse

2004-06-26T21:26:00.000+01:00

Hibernate seems to be getting a lot of press at the moment. I personally haven't played with it yet, but when I do, I'm going to check out this Eclipse plug-in.

Atomicity

2004-06-25T08:32:00.000+01:00

Concurrent Programming in Java (Doug Lea) is really a gem of a book - its been a good few years since I last picked it up.

Atomicity (page 93) is a worth while read - access and updates to the memory cells corresponding to fields of any type except long or double are guaranteed to be atomic. volatile long and volatile double are also atomic.

volatile (page 97) could possible be cheaper than synchronization, or at least no more expensive.

synchronized has a double meaning: it locks permitting higher-level synchronization protocols, plus deals with the memory system (sometimes via memory barriers) to values in sync across threads.

java.blogs

2004-06-23T15:09:00.000+01:00

So, given that I am going down the Java road with this blog, I thought I would add this blog to java.blogs. Thanks to FeedBurner, I now have an RSS feed, and am part of the java.blog community.

Books

2004-06-23T13:45:00.000+01:00

I thought it was time I bought a few new Java books, since my current Java books are quite outdated. Here's what I bought:

Java Performance Tuning (2nd Edition)
Concurrent Programming in Java(TM): Design Principles and Pattern (2nd Edition)
Java Extreme Programming Cookbook
Server-Based Java Programming

I recently read Hard Core Java which wasn't a bad book - the 'final' chapter was pretty good.

Here we go again

2004-05-23T17:23:00.000+01:00

This is my second blog. My first is here. Its primarily for Microsoft technology - since its hosted my Microsoft. This blog is going to be Java centric, with maybe a little Python, C++ and Linux