CAST is not at all what I expected. Then again, that doesn’t surprise me. After all, the organizers are not what you would call ‘typical’.
The first thing that struck me was how relaxed everyone was. People are genuinely friendly and approachable. Considering the number of tester heavy weights around, it would be very easy to be intimidated but its not the case atall. I’m meeting so many people I’ve met virtually. Its fantastic to meet them in the ‘flesh’. Maybe I’m shortsighted but everyone looks a little different from their photo online. I guess my photo is about five years old, so most people don’t recognize me either.
Talks got on their way this morning with Tim Lister and very interesting talk on estimation and the difficulity (read impossibility) of accurately estimating at the start of a project. He had a pertinent and descriptive phrase for a typical project delay graph, calling it ‘the ‘steps to hell’ . In summary when estimating, give a time range, not a date, and be prepared to continuously adjust this estimate and make these adjustments public, throughout the project, regardless of wether the estimate changes are asked for.
A really neat (hey I am in the states!) and unique aspect of CAST is the ability to have discussion after the speaker. I was a bit nervous about this idea, I mean, with all those heavyweights in the room, it takes guts to get up and speak, but the friendly atmosphere dispelled that. The discussion afterwards are as and in some ways more interesting than the actual talk. I was slightly skeptical about this, but after one talk, I am an immediate convert.
What came out of the estimate discussion is that variety in domains fundamentally impact how estimating is performed. Thanks Tom for setting the tone for CAST this year.
I also went to Dot Grahams talk about test automation. She used a car analogy to automation, advocating that automation be split into automated testers who write the code, and testers who design the strategy. Her argument was that not all people who drive cars want to be mechanics and the same stands for testing, not all testers want to write code. The jury is out there for me on that one. I like testers to be curious enough to at least want to look under the bonnet.
To be continued…