Archive for October, 2010


October 29, 2010 Leave a comment

What is Scala?

In their book The Pragmatic Programmer, Andrew Hunt and David Thomas suggest “Learn a new language every year.” I looked around at the hot new languages out there today and decided that Scala was the most promising one to learn.

Scala, short for Scalable Language, is both a functional language and an object-oriented language. Like Groovy and JRuby, Scala runs on the Java Virtual Machine, which means it’s easy to integrate Scala code with existing Java code. It is a statically typed langauge, like Java, which means its type checking is performed during compile-time, as opposed to dynamically typed languages like Groovy and Ruby whose type checking is performed at runtime.

Scala code can be compiled and run as class files, like Java, or you can run it as a script directly from the .scala file. There’s even an interactive command line mode for quickly trying out some code.

Scala has a very concise syntax. It tries to reduce the signal-to-noise ratio of Java code, making the code more readable. It makes a lot of the dots and parentheses of Java code optional, which can be disconcerting for old school Java heads. Clearly the mark of an experienced Scala developer will be the conciseness of their code.

Scala supports a lot of nifty language features like tuples, closures and currying, which makes it a good language to learn what the heck those are and how they might be useful to you in your coding. In general, Scala feels like Java 2.0 in the same way that Java felt like C++ 2.0.

Quick History of Scala

Scala was originally released in late 2003 by its creator, Martin Odersky. The most recent version of Scala is 2.8.0, released in August 2010.

A free web framework called Lift supports Scala code in an environment very similar to Ruby on Rails.

A number of high profile websites have recently switched from Ruby on Rails to Scala & Lift, including Twitter and Foursquare.

James Strachan, the original creator of Groovy, famously said in a blog post in 2009 that if someone had shown him the book Programming In Scala, by Scala’s creator Martin Odersky, he probably wouldn’t have created Groovy.

Programming Scala Book Review

One of Scala’s big draws is its Actor model for concurrent programming. The book I read, Programming Scala by Venkat Subramaniam, showed its focus on this area in its subtitle, “Tackle Multicore Complexity on the Java Virtual Machine”. These days, the new microchips that power our computers feature more and more cores on a single chip. To take advantage of that multicore power, software must be written to take advantage of concurrent processing.

Anyone who’s worked with Java threads knows how painful they can be to get right. Scala’s emphasis on functional programming, and the Actor model, make writing concurrent code a lot easier than in Java. Functional programming avoids state and mutable data, making it ideal for concurrency. The Actor model helps you dispatch functions to perform work and gather the results when each task completes. Anyone needing to write highly scalable code should check out Scala for these valuable benefits.

The rest of Subramaniam’s book provides an excellent overview of Scala. Progamming Scala is part of The Pragmatic Programmer series of books, which strive to be highly readable and concise. At only 208 pages, Programming Scala gets through a lot of material but doesn’t go very deep into any of it. I found myself struggling to grasp the complexity of Scala’s syntax and wished that Subramaniam had gone into more detail in that area.

I’d recommend Programming Scala as a good introduction to the language, but one would probably need to supplement it with a few web based tutorials. I am looking forward to checking out Martin Odersky’s Programming In Scala book soon, which I hope to teach me a lot more about how to write good Scala code.

Categories: code Tags:

Maven & Eclipse

October 29, 2010 Leave a comment

Our project here uses Maven 2 to manage the dependencies and builds. I really like developing code in Eclipse, and I wanted to set up my Eclipse workspace easily from the existing Maven configuration.

The Maven Eclipse plugin makes that easy. Just run the following command:

mvn eclipse:eclipse

and the plugin generates the Eclipse project files, ready to be imported into your Eclipse workspace.

Well, that’s great and all, but when I tried it I saw this error:

[INFO] ------------------------------------------------------------------------
[INFO] ------------------------------------------------------------------------
[INFO] Request to merge when 'filtering' is not identical. Original=source src/test: output=target/test-classes, include=[**/*.java], exclude=[], test=true, filtering=false, merging with=resource src/test: output=target/test-classes, include=[*.properties|*.xml], exclude=[**/*.java], test=true, filtering=true
[INFO] ------------------------------------------------------------------------

A Google search led me to this answer:

Basically the problem is that something changed in the latest version of the Maven Eclipse plugin that prevents this from working with our Maven configuration.

The solution is to use the 2.6 version of the plugin instead.

mvn org.apache.maven.plugins:maven-eclipse-plugin:2.6:eclipse

Categories: code Tags: ,


October 28, 2010 Leave a comment

My most recent project, the Multimodal Operational Test Harness (OTH), is a thick Java Swing client used to gather biometrics and package them up into a properly formatted XML document. The Maven 2 build generates an uber jar including all the dependencies that can be run from the command line, but we also have a requirement to run as a WebStart application. It’s not generally required to digitally sign the software, but we have the requirement locally.

So here’s how I went about signing the OTH.

1) Generate the key pair and self-signed certificate.
I could have used the command line for this step, but instead I decided to use the Maven Keytool plugin.
Here’s a clip of the pom.xml file I created:

				<dname>cn=commonName, ou=organizationUnit, o=organizationName, c=countryCode</dname>

I know it’s not the best idea to put the passwords in the pom.xml file, but the security of this keystore isn’t very important and I wanted to allow other developers to access the keystore without needing to discover the password.

Take special note of the option. If you don’t specify this value, the default is only 90 days. 10 years is much better.

2) Instead of just relying on a self-signed certificate, I need to get my certificate signed by a Certificate Authority. To do that, I generate a certificate signing request (CSR) file.

keytool -certreq -alias oth -file oth.csr -keystore keystore.jks -v

3) I sent that through our security team to our Certificate Authority. They first sent me their root certificates. These have to be imported into our keystore before we can authtenticate the certificate reply.

keytool -importcert -file "rootcert.crt" -alias rootcert -trustcacerts -keystore keystore.jks -v

4) Then I receive my certificate reply from the certificate authority. Now I can import that into my keystore and my certificate is authenticated.

keytool -importcert -file "cert.txt" -trustcacerts -keystore keystore.jks -v

5) You can verify the contents of your keystore by running the following command:

keytool -list -keystore keystore.jks -v

6) Here’s the bit of Maven WebStart plugin configuration that uses the keystore to sign the war.

			<verify>true</verify> <!-- verify that the signing operation succeeded -->

Finally, here’s the URL of the documentation of the Java keytool tool.

Categories: code Tags: ,