Continued from Part 2
Last coffee break before the end and the trawl back on the m25 back home!
Talk 9 – Smart algorithms: Are we ready for this – Bill Matthews
Bill Matthews coming up, someone who I’ve known for a while and always has really in testing insights into all things testing.
“Smart algorithms – are we ready for this?” Bill talks about his experiences with machine learning and how prolific and accessible it is and how this affects testing.
He shows how machine learning that guesses ages or animal types changes even by providing a different resolution image, so while the image is exactly the same, just larger, the results are different.
What about testing… How can we test these systems. They have a higher level of complexity that most testers are used to. They simulate cognitive processes rather than procedural, it learns, adapts.
This makes standard input/output testing really difficult because the results may differ from day to day. This poses a really interesting question and I suspect it would require a more abstract approach to validate the behaviour. Perhaps there will be a series of scenarios that you could throw at it?
How can testers provide insights into the behaviours? Bill suggests creating a series of simulations, but the is an oracle problem, is it performing correctly? I wonder whether a rules based approach could help to determine the boundary cases, eg the Google car must not crash into another vehicle… But even then there are so many variables!
Start studying general systems thinking, cybernetics, control systems, simulation, statistics and probability, this will all help to prepare testers for this complexity.
Talk 10 – Nowhere to hide: Adjusting to being a team’s sole tester – Nicola Owen
Last talk for me is by Nicola Owen, “nowhere to hide: adjusting to being a teams sole tester”. Really interesting she did a graduate programme on software testing!! We need those here in the UK! However she knew the concepts of a lot of testing stuff, but didn’t know how much she didn’t know until she got into industry.
She explains that being part of a bigger test team, she felt safe from any consequences of her testing. That’s not to say that she wasn’t complacent, or didn’t care a bout her job or quality, but nevertheless, it was more comfortable.
Then moving to being a sole tester on a team, she gained more responsibility, more autonomy in how and what to test. However there was nowhere to hide, but it helped solidify what doing a great job actually meant. Personally, I hope that it wasn’t only Nicola taking the can for bugs getting out to production, and the whole team took responsibility.
She started using mind maps for testing to reduce time and provide a visual test plan and test progress. It generated comments such as “I hadn’t thought of that”, “you should probably add that scenario”, and testing became a share responsibility.
Nicola calls out BBST as one of the most influential courses she’s attended, and I can confirm that!
Nice talk from Nicola, great to see lots of new speakers this year, and another experience report which I prefer too.
Unfortunately I have to duck out before the 99sec talks, but a special shout out to Deborah Lee who I hope manages to get up and do her talk. She tweeted in response to my talk here back in 2014, and her kind and inspiring words will stay with me for a long time to come.
So, farewell Testbash 2016, it’s been awesome… AGAIN, and already looking forward to 2017! Thanks to Rosie and the gang for organising it!
Wow, my first attempt at live blogging and it’s been exhausting but has really helped me listen better.