JUnit is great. Jython and JRuby are even better. Unit testing libraries look the same everywhere, so why not use the one that lets you get your job done faster?

Unit testing has taken the world by storm. Almost every major language has a JUnit-like library. And here's the good news: These libraries are so similar that once you learn one, you can use any of them.

Given that many languages interoperate directly with Java, you aren't limited to JUnit. You can pick the library that provides the best language level support for writing concise, easily maintainable unit tests. Circa late 2005, there are good arguments for the unittest module in Jython, and Test::Unit in JRuby.

The talk is divided into three parts:

1. A crash introduction to unit testing Java code. We'll cover test-driven development, test cases, assertions, fixtures, and test suites. We'll be using Jython and JRuby but the concepts are exactly the same in any language
2. A look at the Python and Ruby features that make Jython and JRuby compelling choices. We'll see how tests are faster to develop and easier to maintain than JUnit tests. And perhaps surprisingly, we'll also see how tests scale better when applications get large.
3. A candid comparison of Jython, JRuby, and JUnit, including advantages and disadvantages of each. One size does not fit all.

Java has always provided a model for concurrency and threads. With Java 1.5, this model received a major facelift. Learn how to use the new concurrency utilities to build responsive, scalable, and correct concurrent applications.

Java's support for threads is broad and deep. From the early days of the platform, programmers have used threads, synchronized blocks, and monitors to build safe multi-threaded applications. Java 1.5's new concurrency utilities greatly reduce the need to use these primitives directly. Now, Java provides a set of lock classes and task scheduling tools that provide much more leverage in writing real applications. We'll explore java.util.concurrent, and also see how the Java Memory Model has been corrected in 1.5.

Most of the advantages of java.util.concurrent can also be had in previous version of Java. We'll discuss Doug Lea's concurrency utilities and the backport of java.util.concurrent to 1.4. Both of these are appropriate for production use.

Finally, we'll look at common mistakes in multi-threaded programming. The most common mistake is using threads when you don't need them. We'll look at alternatives to threads, and how to choose between them.

For centuries people have used crypto to build (and break) secure systems. Computers have only raised the pitch of conflict, providing enormous cryptographic power at commodity prices. Most programmers do not write their own crypto libraries, instead relying on the services of an operating system or virtual machine. But even with all this support, building secure systems is a daunting task.

This talk will cover three things all programmers need to know: 1. the basic tools of computer crypto 2. the programmatic APIs to these tools 3. common programming mistakes that can undermine otherwise secure systems You will learn the basics of hashing, message digests, public key encryption, symmetric key encryption, certificates, and key management. With each, you will see programmatic examples, with advice for correct usage.

The Java platform is built from the ground up with security in mind. This talk will introduce the security features of the J2SE, building quickly from the basic classes to realistic examples.

You will learn the core APIs:

SecurityManager, AccessController, Permissions and Policy

JAAS Subjects, Principals, and LoginModules

You will then see how to invoke these APIs in real application scenarios. You will learn how to:

Partition your applications to safely invoke downloaded code

Read and write Java policy files

Extend the architecture with custom permissions

Provide secure services through PrivilegedActions

Use JAAS to authenticate and authorize users