Last week we had Adam Knight present a talk on “Fractal Exploratory Testing” at the Software Testing Club Meetup here in Cambridge, hosted by the lovely folk at Linguamatics. FET is a concept Adam thought up in order to explain the benefits and workings of exploratory testing, drawing analogies to geographic expeditions and using fractals to visualise how your test focus might evolve on exploration. He has written about it in the past, from an initial draft to a revisited version, and there are also comprehensive write-ups from the evening:
- Fractal Exploratory Testing (Adam Knight)
- Fractal Exploratory Testing Revisited (Adam Knight)
- Just the Fracts Ma’am (James Thomas)
- Fractal Exploratory Testing: A Review (Neil Younger)
So I won’t be going into much detail here, but rather want to capture three points that were most inspiring to me:
- Adapt what you report about what your actually testing based on who’s asking and what that person wants and needs to know. If you’re asked what you’re doing, a business stakeholder might be more interested in the overall objective you’re working on (what is testing trying to achieve here in the overall picture?), a test colleague more in the details of your charter (what has testing already covered with this, what risky areas are yet to do?). Although the activity is the same, you might want to describe it differently based on context.
- The product development is often not completely predictable. So why would we expect testing to be predictable, scripted, static? Testing needs to be able to shape itself based on new experiences made. Make it possible to add discovered charters to your predicted charters as you go along.
- Although you might not be able to plan every test beforehand, that doesn’t mean you can not prepare. Exploratory testing is about having people ready with the right tools and resources – that can be all set up so that you can hit the floor running when you start your exploratory charters and will give you more flexibility while testing.
On other notes, I really liked that although Adam had covered the topic already in some detail in writing the talk was not a mere repetition. I had (re-)read both blog posts shortly before the Meetup but got loads of new thoughts and perspective out of the talk, accompanied by very helpful graphics and visualisations. And than, of course, the possibility for further discussion – we had planned to do about 30 minutes talk, about 15-20 Minutes Q&A and than pizza and drinks. Alas everybody was really involved, bringing up questions and exploring the topic that it felt bad to interrupt (which happened way later than planned) and drag people towards pizza – but happily, the discussion just kept on flowing. I just love that about the software testing community – a thought provoking talk where people pick it up straight away and engage. Thanks all for a splendid evening!