Tuesday, 27 November 2012

2 of 3: Eurostar Testconference 2012 – Lessons learned and to learn...

Part 1 of this posting is here

Huib Schoots and changing the context at the Rabobank
The world of testing is changing: As we have been using methods like TMap and ISTQB in the past, which could fit perfectly in your project or organization, the big differences in organizations and projects lead us to defer from these methods or even throw them away completely. This to rethink testing and get a better fit with testing in certain ways. The Rabobank is doing so and is a good example of introducing these (new) ideas in bigger organizations, where you would expect it not to be possible or at least too difficult to get this accomplished.

This because organizations such as banks have big IT machines that exist for a long time and have certain ways of working and people that do not want to change after working for 40 years in the same job.

Not only testing, but the world of IT is changing and the way we think about software development. The agile revolution is one of these changes. And this was also introduced in the bank and was as a hook to improve testing too.

Low tech dashboards as a way to replace comprehensive reporting
Some actions were taken, as doing testing dojos, focus on learning and training for testers, let people speak on conferences, experiment with Exploratory testing, less documentation by using mind mapping techniques and low tech dashboards, introducing test workgroups to figure things out and get the testability on a higher focus within the organisation.

These are all things I can put on my own test improvement checklist. I will do that and search out some more about those subjects and will think about how to introduce those within organizations when I'm on a project of such kind. 

Graham Freeburn and mapping your way to better testing
And the change is also happening in the way we look at test process improvement. Graham Freeburn had used Test Proces Improvement (TPI) in the past. But he learned that the TPI method is too much focused on processes and has less focus on actual skills.

So he created another way to improve testing processes. Now he uses sessions such as kick offs, presentations and creates an action plan from these sessions. Because of this there is constant communication between him and the people and between the teams in  organisations. He uses the Goals problem tree and checks constantly with the business goals. For working out and collaborating he mainly uses mind maps.

  • Best quote: If you deliver value every time, the customer will worry less about costs
  • Important skill he mentioned: Influencing other people.

The sessions he mentioned that he uses are good to put on my test improvement list. So instead of asking questions and do interviews, communicate, think, explore within these sessions to get the quality of testing higher. The outcome would fit an organization better than the standard formal process of TPI.

Cognitive biases
Martin Jansson (from the test lab) already mentioned this to me while talking in the test lab: “Knowing the requirements could cause you to not see certain aspects of a product and could eventually lead to missing some bugs.”

I checked ‘cognitive biases’ on Wikipedia: “A cognitive bias is a pattern of deviation in judgment that occurs in particular situations, which may sometimes lead to perceptual distortion, inaccurate judgment, illogical interpretation, or what is broadly called irrationality.”

The inaccurate judgment and perceptual distortion would be issues we as testers are dealing with every day.
"It is the human tendency to make systematic errors in certain
circumstances based on cognitive factors rather than evidence."
A human is easily influenced. The speaker gave an example of a scientific research project. People were asked to write down (or think) about the last two digits of their social security number. After they were asked about this, they participated in an auction. The result was that people with higher numbers in their head because of a higher number at the end of the social security number ended up bidding more than the others.

Some ways to overcome biases and he mentioned are the use of checklists, always be cautious when making decisions (black hat thinking), reference class forecasting (check with other industries and compare).

Later that day I spoke with Srikanth and about the difficulty I have with biases, namely not recognizing them, he said, that it is good to be aware at least these exist. (so a bit of curing from biases here in his talk)

I will have to update my checklists I already have and will use them also constantly. Until now my own checklists are all over the place and not structured to my needs. Rethink, use new ideas and structure this for future use.

Good to check: Michael Bolton has a presentation about the confirmation bias here http://www.developsense.com/presentations/2010-04-ConfirmationBias.pdf

To be continued in part 3

No comments:

Post a Comment