Agitator 3.0 Testing Tool

If you have done any Extreme Programming, you probably know

that unit testing is one of its fundamental aspects. Moreover, in an ideal work

environment, the test cases should be automatic, they should test everything in

every class, and guaranteed to work all the time. Unit testing can be a

nightmare; you need to write lot of test cases. It will be nice if software

engineers test their code as they write it, it's called 'developer

testing'.

Agitator 3.0 Testing Tool  Price and Warranty: $ 40,000 for a ten-user team Meant For: Java developers, test engineers and development managers  Key Specs: Dashboard to ease testing, plug-ins for EJB, Struts and Log4j, support for Java SE 5 and Eclipse, easy integration with JUnit  Pros: Easy to build unit tests, code-rule validation, snapshots and coverage analysis Cons: Attempting to agitate using BEA JRocket causes slowdowns & VM crashes Contact: gitar Software, CA.  Tel: 650.694.7572. E-mail: support@agitar.com  RQS# E63 or SMS 131263 to 9811800601 Here we agitate the MedTracker application from the com.pcquest package

Thanks to Agitator, an intuitive test environment for Java

developers, developed by Agitar Software, that provides a new approach to

testing and agile development. Agitator enables organizations to test and manage

Java apps throughout the development lifecycle. Using Agitator is much more

beneficial than manually generating test cases.

With Agitator, you can analyze Java bytecode and observe

the behavior of each class and method in a variety of circumstances. In minutes

you can create durable test assets that would otherwise take hours or days. Bugs

discovered earlier in the lifecycle often cut cost and development time than

bugs found toward the end of development (system testing, load testing) or even

worse, post deployment.

Agitator comes with a set of command-line tools and a set

of corresponding Ant tasks that you can use to automate regression testing as

part of your regular build process. Using these tools, you can choose to agitate

classes based on different criteria's, for instance, classes that have changed

or only classes with changes to agitation configuration etc. Agitator further

helps in streamlining regression testing by monitoring just a specific class

rather than an entire project. Agitator saves generated objects in the object

pool so that it can reuse them in future method calls. As part of this process,

Agitator automatically adds new test cases as the code changes to ensure that

any tests stay up to date. Black-box testing by third parties is potentially

very powerful, but it can be ineffective if the code doesn't work properly when

delivered and QA resources are wasted finding trivial coding faults. 

We tested Agitator on Windows XP and integrated it with the

Eclipse. We appreciated its support for Java SE 5 for mapping generic types,

autoboxing, enumerators among several other features. We were impressed to find

the plug-ins for Struts and Log4j. Agitator is also capable of identifying an

item's state before and after a method call. This allowed us to form

assertions that tracked how a variable's state changed. We unit tested the

MedTracker application, and the



results were awesome. We used checkpoints. A checkpoint is a high-level summary

of a set of agitation results at a particular moment in time. Checkpoints

provide a baseline for comparison. You can use them to chart progress over the

lifecycle of a project. When you have saved one or more checkpoints, you can

compare your current status with any previous checkpoint. Large development

teams can leverage from Agitar's Management Dashboard which aids in testing

and can use these checkpoints to show which classes are new, which ones have

been modified, and which ones are unchanged since the checkpoint was saved.

Overall, a nice unit testing tool with rich functionalities.

Bottom Line: An award winning Java unit testing tool

for developers, which not only automates testing, but helps to



identify bugs during early stages of software development cycle and create

powerful regression suites.

Kunal Jaggi