Talk Notes: testing at Microsoft

As I’ve mentioned recently (and probably way too many times by now), I’m in Munich to give two presentations at OOP 2011. My first presentation is tomorrow. Tomorrow evening, in fact, in the “Night School” track. Tomorrow’s schedule has tutorials in the morning, lunch, two keynote sessions, and then, after a short break, 4 concurrent sessions running from 6:45 to 8:15. I’m a little worried about attendance – I think this conference is pretty big, but after tutorials and two keynotes, I’m worried people won’t stick around for the evening sessions. It’s the first day of the conference though, so people may have more energy than I give them credit for.

My session is also the only Night School session in English. I’m not sure if that’s good or bad for attendance. A few months ago I was rereading some of the Richard Feynman books (Surely You’re Joking… and Adventures of a Curious Fellow). In one of those, Feynman tells a story of teaching in South America and taking the time to learn Portuguese well enough to give the presentation in Portuguese. I should have taken inspiration from this and at least learned an introduction in German, but for some reason, German has always been a difficult language for me. I once gave an introduction to a class, including a short personal bio – all spoken in Japanese, and I could do at least as well in French (both languages are a bit rusty, but I’m sure I could pull it off). In German, however, I’m lucky if I can remember thank you and please. I’m not sure why I have the block, but I don’t worry about it very much (pretty much never except for when I’m in Germany).

The advantage to giving the talk in English is that anyone attending from outside of Germany who wants to attend one of the sessions will be forced to attend my session (cue evil laugh…). I doubt, however, that there are many (or more likely, any) folks attending who do not speak German except for me and a few other speakers, so I’ll have to win them over with content.

Oh yeah – after 350 words about my talk you’d think I’d have mentioned what it’s about. The title of the talk is “Testing at Microsoft: Past, Present, and Future”. I was never 100% happy with this title, but someone must have liked it, because there’s a keynote on Tuesday called “Design Patterns – Past, Present, and Future”. My talk is me back up on the “let me demystify testing and software engineering at Microsoft for you” soapbox that I’ve been on for at least five years now. The book (HWTSAM) was part of that soapbox, but so are many other talks, blogs, and articles over the past several years – as well as the slowly progressing MSDN Tester Center.

I think there’s a lot of mystery about how software is made at Microsoft, and how we manage and run our huge(!) teams. Even with HWTSAM, knowledge of testing at Microsoft is still largely hearsay and myths, and for some reason (nobody has ever asked me to explain testing at MS), I keep dispelling myths and wondering where people get their ideas (no, testers are not paid one pay grade lower than developers, and our testers do actually test software (which is slightly different than “only writing tools and automation all day”)).

This talk has some background on how the test role evolved at Microsoft, and what kinds of people we look for when we hire testers. I’ll talk about org structures a bit (b-o-r-i-n-g, but gives some context of how we manage groups like Windows that have well over 1000 testers). I’ll talk about strategies and tools too. I’ll also talk about some of the newest stuff we’re doing – including my take on customer focused test design (which I didn’t get into at all in my webinar last week), and some discussion of productivity games (aks “serious games“), and how we’re using games at work to increase productivity and quality.

But here’s the really important part. The session (if you did the time math above) is 90 minutes, but I didn’t prepare anywhere near 90 minutes of material. I’m not a slacker, I did this on purpose. When I give talks like this, my goal is to present “just enough” information to get people to ask questions. If you’re attending this talk (or any version of it in the future), it’s your opportunity to ask any question you want about testing or software processes or anything else remotely relevant to Microsoft or me. If I fill the time with too many charts and dense slides, you’ll be way too bored to ask questions at the end, and neither of us will enjoy the talk as much.

I am going to work so hard to make this an interactive session, that I will be disappointed (and perhaps slightly ashamed) if I get to my last slide any time before 8:05 (8:10 is my target). A larger crowd will help, but great questions will help even more.

We’ll see how it goes.

Follow up: I was happy with the way Monday’s version of this talk went. People asked a lot of good questions and seemed engaged. I went two (maybe three) minutes over time (sorry again), but covered a huge amount of “stuff” about making software at Microsoft.