September 26, 2011
I attended PyCon UK 2011 at the weekend, which was loosely run as an unconference. As I was about to start writing about the conference, I realised I haven’t written anything about PyCon AU last month in Sydney, which I also had the pleasure to attend. Two conferences in
two months! I’ll try to mention my Australian experiences in another blogpost, but will mention some items of relevance in this post…
Once again, John Pinner and his trusty team put on an excellent event. The TechnoCentre in Coventry provided a superb venue, and the on-site catering was very good. You couldn’t go anywhere without finding a water cooler, fine teas, coffee and other beverages. Delegates were fuelled up in the mornings with breakfast baps and pastries, offered a decent lunch (especially on the Sunday) and served a fine conference dinner on the Saturday evening.
It’s been a good year for keynotes for me: PyCon AU had Audrey Roy, Mary Gardiner and Raymond Hettinger. Audrey’s talk on diversity was particularly important, emphasising that diversity isn’t just about encouraging more women into computing. Worthy though that goal is.
PyCon UK had keynotes from Allison Randal, Laura Creighton and Lorna Jane Mitchell. Allison kicked things off with “The Fallacy of the Zero-Sum Game”, discussing why free software is the future of technology because the values of cooperation and collaboration provide a better environment for improving technology. Even though I was clutching my freedom-hating MacBook, free/open source software is something I believe in for the same reasons Allison does.
Laura Creighton is always a joy to watch in action. “Reflections on the work of Sociologist Charles Perrow, and what he can teach software developers” did seem to leave a few people in the audience scratching their heads, but I really enjoyed it. The general idea is understanding the lessons learnt by Charles Perrow on analysing “risky systems” and applying to software development. He knows a bit about this through his study of accidents such as Three Mile Island. I’ve discoverd that simple things are more complicated than I first thought, and that tightly-coupled, complex and unpredictable systems need to conform to two mutually incompatible states to be controlled and understood. Which is why they are both very risky and lead to catastrophic accidents when things go wrong: aircraft crashes, spacecraft breakups, nuclear reactor explosions. The sort of thing you don’t want to be involved in.
It’s one of the nice things about the Python community that PHP developer Lorna Jane Mitchell can be welcomed wholeheartedly. I don’t have her talk title to hand, but she conducted a wonderful, slide-free talk about contributing to open source projects and the benefits outside the project that can result: job offers, recognition, personal improvement. It was well-timed, as I was about to undertake my first conference talk…
I took the stage for the Testing Workshop alongside Michael Foord. I wasn’t quite sure what level to pitch the workshop at, or even what to cover. I’d abandoned the tutorial style I’d originally thought about, although there was a bit of legacy from that era in some of my slides. It was on researching the subject that I’d realised a fundamental issue with testing: no one really agrees on the terminology, and most books and training material for developers don’t really cover it in much breadth. I took a gamble and decided to start from basics, to cover the widest possible audience and to try and get people thinking about the types of testing and the terminology used. I also opted to cover the very basics of Python testing, ditching some material on TDD and dealing with legacy code.
As I was leading up to the presentation, I was beginning to think it might’ve been the wrong decision when I looked out at a very large audience and spotted some very knowledge people in the Python community looking back at me. Uh oh. I did give a clear warning, but no one left which was a good sign. Also, despite preparing my set-up in advance, there were some additional technical hurdles to overcome – notably, coordinating with the live streaming! At least the internets don’t heckle.
My first section covered what testing is and the types of testing. The second section was really more of an overview of testing with Python and so most of the audience knew the options already – but important to ensure everyone is aware. However, it lead to some interesting questions and responses from the audience, which was good. I then handed over to Michael who covered his Mock library and served the audience with some much-needed in-depth material. Jonathan Hartley followed-up with a talk on modifying the Django test runner, a modification I wish I’d had earlier this year while debugging a nasty issue with test fixtures and multiple databases.
Afterwards, I read some very nice tweets in support of the subject matter being presented, and I had some good bits of feedback post-talk. I’ve spoken a couple of times at the London Python Code Dojo, but have had the advantages of a smaller audience, smaller venue and a subject of my choice – not much different to when I stand up in front of a small group of people and teach them to snowboard. The venue was definitely a lot bigger, the audience a lot bigger and strangely the podium and lectern made me feel the most awkward. The first five minutes my throat was the driest it has probably ever been and I then worried I’d spill my bottle of water over the laptop and other electronics surrounding me!
In the end, I rather enjoyed it. To use a phrase from Russel Winder, I’ve picked up the bug. Even while I thought I was failing badly, I would look out across the audience and see someone smiling, nodding their head in agreement or looking like they’d just discovered something new. Which is all that matters. If you attended, I’d like to say thank you for your support and hope you came away with something of use.
Next year, I’m toying with either a talk on beginning Test Driven Development or something a bit more technical: common patterns for testing legacy code. Although I’ve been dealing with PHP legacy code recently, I’ll aim to keep it in Python. The principles are fairly similar across the dynamic languages.
Turning to other matters, teaching school kids to program was a recurring theme during the weekend. The BBC Micro featured prominently in discussions, as did the new Raspberry Pi – which is a wonderful piece of kit scheduled to cost just £25.
I discovered programming through the BBC Micro at school. I’m not sure the exact age I was, but I was probably about 5 or 6 and the Beeb had just been made available in schools. When we weren’t playing educational games, we could fire up the BASIC interpreter and do some very simple things. However, it wasn’t until my parents bought me a Commodore Plus/4 a couple of years later that things really took off for me. Doing it as a career wasn’t something that occurred to me until university, or just before. It was always a fun hobby.
Logo on the RM Nimbus came a bit later, and I remember returning to my old primary school a few years later and teaching a group of very excitable kids to draw on the screen using Logo. Kids of all abilities loved it, and my hope is that one of those kids ended up programming.
Back in my day (now I sound really old), computing wasn’t on the syllabus because computers were new, a bit of a novelty and the teachers had little or no experience of them. That was a reasonable excuse in the 80s. We live in the 21st century now, and I was horrified to discover recently that computing skills in the classroom consist of learning how to use Word or Powerpoint. The computer is this magic box that comes with games, Facebook and something to type out homework with. If anything, we’ve actually regressed.
At PyCon AU, I attended a talk on the NCSS Challenge which teaches school kids across Australia to program. It even caters to different levels of ability, from complete beginners to those who already have programming skills. You can watch the talk on the technical side from the PyCon AU YouTube channel.
That kind of scheme is just wonderful, and it makes me sad that the UK has nothing similar. Instead, we’re raising a new generation of kids who think a computer is indistinguishable from magic, the software just appears, and everything has a nice, neat Microsoft brand emblazoned across it. We’re not teaching kids to fully explore technology at all levels, to create applications themselves and become producers not consumers. It makes me wonder how many great programmers of the future are not discovering their calling earlier in life?
Anyway, thank you to all the crew at PyCon UK – the organisers, the venue staff, and my fellow delegates. See you all next year!