Saturday, October 14, 2017

Heads Up

I tell you what: of all the things I might've expected to see on the first slide at Quality Jam London 2017, my own professional-work-photo-grinning, shiny-pated, blue-tinted face peering back down at me from behind a massive Thank You! wasn't it.

Expectations are grist to the working tester's mill, yet also often the bane of their lives. Tony Bruce, in  Manual Testing is Dead. Long Live Manual Testing, called for testers to set the expectations of the people that they interact with. The term "manual testing" undersells what testing is, or can be, with its connotations of manual labour, unthinking monotony, apparent separation from (woo! sexy!) automation and the historical association with scripted test cases.

For Bruce, testing is "the pursuit of information" but he doesn't necessarily rush into meetings spouting from that kind of lexicon (although he's singing my kind of song right there). Instead he promotes the use of PAC (purpose, audience, context) to guide conveying the desired message in a way that a recipient can take on board.

So he might refer to using questions, experiments, exploration, engagement, surveys, investigation, tools (which includes woo! sexy! automation), spending time thinking, iterating, and adjusting to new data. And he'll report on what he found, but not necessarily what he produced: "I discovered that the server will crash when there are 17 simultaneous requests" over "I filed a bug about a server crash". It's all testing, and it's on testers to explain and demonstrate how and why and what value it delivers.

In her talk, Succeeding as an Introvert, Elizabeth Zagroba led us through the traits she has that characterise her introversion and then gloried in the ways that she has found to harness them, to have them contribute to her success as a tester.

She's a passionate spectator, but that makes her focused, good on detail, and a great listener. Her preference for preparation helps give her opportunity to spend time asking "what if?", really think hard about things, and not rush into rash actions. If you're worried about her apparent reserve then consider that she says nothing when there's nothing to say, that she's comfortable in herself, and won't give in to peer pressure.

Strategies for dealing with the day-to-day were plentiful: find yourself an extrovert ally to perhaps open a spot in a boisterous conversation for you, be the scribe and write your comments direct into meeting notes (projected live or afterwards before circulation), allow yourself headspace by asking "can I get back to you?", carve yourself out some recharging time by booking solo meetings in another part of the building, ask for what you need (or risk not getting it). And, finally, some advice for extroverts: treat others how they want to be treated. (But, in fact, I think that's good advice for verts of any flavour.)

When asked how testing can be improved in an organisation, Keith Klain reaches for a simple model based on the kind of iron triangles you'll have seen in countless slide decks and blog posts. He wants to know how the notion of quality in the test team diverges from that of the business, he looks for transparency and honesty in the cost of testing activity, and he reviews time efficiency across the board.

His Debugging Your Testing Team covers the three main failings that he's seen this kind of analysis expose:
  • People who don't think about their work. This leads to wasted effort, poor coverage, bad feelings between team members, and worse. It can perhaps be turned around by finding ways to engage your staff, by coaching interactional expertise to enable them to converse on a level with others, by helping them to explore their craft.
  • People who don't trust their test team. For Klain, this is typically a process issue and addressing it requires reviewing attitudes, being open-minded about changing things, taking care to understand (to link back to Tony Bruce's talk) which nuanced version of X a stakeholder is thinking of when they claim X ... or Y, asking why (for example) particular metrics are being collected, and who is consuming them.
  • People who don't like testing. Technology is frequently at the heart of this problem, he finds. In particular a fetishisation of tools. Satisfying as it might be, automating all the test cases is not necessarily testing, which Klain views as a social activity. Step back and consider what the right thing to do is, for who, at this time.
Having given his common diagnosis, he proceeds to give his common prescription:
  • Focus on business risk
  • Make objectives based on business-aligned principles
  • Fire your test managers
Fire your test managers? I think I'd better keep my head down.

See also: my sketchnotes from Quality Jam London 2017.

No comments:

Post a Comment