Exploratory Testing Erik Petersen, Emprove ecp at testingspot dot net
"Pirate Encyclopaedia": Exploratory testing is about exploring the problem space to find relationships.
Optical illusion: Problem with testing. You thing you're using your senses, but your brain gets in the way.
Nike Transformer: Anything we do, we bring models based on our experience, and we interpret our perceptions with reference to these models.
The Wind, always misunderstood: If you don't understand what's going on, use exploratory testing to try and find patterns.
20% of knowledge used is explicit, 80% is tacit, recognition of patterns already experienced . Exploratory testing is so powerful because it builds upon and makes use of this pattern-seeking ability.
Thomas Jefferson's letter: "Take Careful Observations"
Galileo: Plan your experiments before you undertake them. Decide what success and failure would look like. Then conduct the experiment. When you have the result, determine what previous held assumptions that changes, and the impacts of that.
David Harrison, racing balls. The intersection. Seats: Testing reduces the risk that the models we base our assumptions on are incorrect.
Samba Band: Based on what I've just learned, what's the next experiment to try?
The Clocks: Alan Jorgenson: Input, Output, Data, Computation. When you have constraints on any of these, you find bugs. Whittaker took this further, and proposed a set of 17 attacks that could locate the various types of bugs found through user interfaces.
The bilingual clock: Implemented in one model, which works. Add a second model, assumptions made in implementing the first model don't work in the second model.
Any role for automation in exploratory testing? Yes, e.g. in the setup. But the key to exploratory testing is in the human thinking. Things are heading in the direction whereby robots will be testing models, and when they find a bug, a human can perform exploratory testing.