The future of software testing

I have been reading JW on Test since I learned James Whittaker started to blog. In his post titled "the future of software testing (part 8)", he states that "test code should ship with the binary". That is an interesting concept and directly relates to the idea I've been working on which I call "On Demand Regression Testing". Shipping test code would be fine for new code but what about legacy code?

Recently, at the IWST I heard a very interesting quote in a presentation given by Anthony Panozzo on Test Driven Development. The quote is from Michael Feather's book "Working Effectively with Legacy Code". It basically states that legacy code is code without tests. I would say that legacy code is code without tests and/or documentation but that is essentially the same.

In my current position I deal mainly with legacy code. That is code written over a period of decades with little documentation and certainly no tests. Sure it was tested when originally installed but those tests were not kept as a corporate asset and now they are gone.

To achieve the goal of regression testing this type of code, we started with a functional decomposition of the particular function we are testing. Then we look at the inputs, process, and outputs within the function. We create a logic table for the inputs and also examine any interfaces that the function may depend. While we are doing this software engineering analysis, we try to understand the business process and types of users involved. We document our findings and publish them along with our tests that generally run on an automated basis.

The main roadblock behind "on Demand Regression Testing" is the data. If we can find a way to either create data based on certain characteristics and use them in our automated suite, then we would be much further along. Then we would be able to wrap our application with tests rather than embed them within our code, since we don't have the skills nor does the organization have the will to do that, at least for now.

September 2008 IWST - Testers who write code

I attended the Indianapolis Workshops on Software Testing in September along with a colleague of mine. We found the experience to be very worthwhile and informative. There were a good number of software engineering/testing professionals and it was great to see what they were doing in their shops. I gave a presentation on integrating HP Quick Test Pro and Empirix Hammer applications. Mike Kelly blogged about it here. The notes from the presentation can be found here.

First post

The purpose of this blog is to provide information about software engineering, project management, and testing and how to do these activities in a more efficient and effective manner. I will share my experiences, strategies, and philosophies and a few stories along the way. The title of the blog represents my feeling of how software projects should be viewed. Like a cruise in the Northern California Wine Country, software testing should not be difficult, it should be an enjoyable and sometimes difficult struggle. To do it effectively and consistently you need the right tools. I am not talking about automated software testing tools either. I am talking about tools such as the proper frame of mind, environment, experience, knowledge, and support. These are the tools that I hope to help you develop so that we can all build better software and have fun doing it.