I believe the strength of the CAST conference is in its ability to provide a space where testers can come together to query and contemplate any subject related to testing. It’s a unique environment and the conference organizers have made a great effort to make everyone feel welcomed and involved. Lightening talks were open to any tester who feeling inspired by the talks and people around them could stand up and give a five minute talk on any subject they felt passionate about. About 15 testers took up the offer, which I think goes to demonstrate the impact the conference was having on people. I had my five minutes of fame, and got a chance to talk out IM tester coaching and the challenges involved in creating a learning environment in such a medium.
The topic of the conference was ‘skills in testing’ and the talks content was roughly split between the softer skills such as communication against the technical skills such as testability and automation. Both equally important to any tester. I tried to focus my learning on testability and went to some great talks on the topic. In particular, I liked Cem Kaner’s talk on exploratory test automation, where he challenged testers to ‘test for the unthinkable” (paraphrased,I forget the original quote) and one way to do that is to randomize your testing.
But Monica Wodzislawski had a different take on testability and gave a clever approach to testing large-scale scientific computations which have large data inputs but small data outputs where the challenge is to test without an oracle. She calls it the ‘short-circuit’ method where you include a checksum by minor additional computing.
What they both agreed on was that systems were only going to get more complex, and as testers our challenge will be to use our ingenuity to find ways to test large, fast, parallel computations.
I can’t complete this post without a mention on Greg McNelly’s talk on Testability and Technical Skill. Greg started out his career as a programmer but found the cognitive challenges in testing more compelling (I can understand that!). After testing for a while Greg ended up in development again, determined this time to bring the insights he’d learnt as a tester into the development program and therby enlighten developers on how to code for testability. Instead he found himself quickly resorting to his old developers habits. The insight there was that developer (and for that matter tester) behavour patterns are incredibly strong, and rather than try and change developer culture, its better to use your ingenuity and other tools to develop your own testability.
CAST 2010 is run soley by volunteers. It’s a grass roots community (a bit like the software testing club). There was a call for volunteers so please if you want to give something back, contact the association for software testers. CAST 2011 will be in Seattle and is being organized by James & Jon Bach. If you want a taste of CAST 2011 go to James & Jons video on YouTube.