This is the 2nd part of my series on gems from the QCon conference in London 2010. The first part can be found here.
Martin Fowler: What are we programming for?
This track was given a small room, which was not unreasonable considering unusual topics like work ethics, social responsibility, and charity work, but the organisers completely failed to understand the attraction of Martin Fowler as the first speaker. The room could have been twice its size and still not comfortably swallow the crowd that wanted to hear him. I was forced to listen to it sitting down, quite uncomfortably, on the floor. But I am glad I did.
Martin speaking from the heart
Martin’s talk was given without slides and what appeared as mostly from the heart. It was basically a sampling of situations where work ethics come into play in our profession. His main message was that many software developers seem to view themselves as “hired guns”, mercenaries that will do anything as long as you give them a spec. Questions like “For what purpose will the software be used?” and “Do I want to keep supporting this employer/client?” never seem to enter their minds. Well, maybe it’s time they did? Think about your own life for a second. Exactly for what are you giving away your talents?
Thoughtworks also presented some charity work they had done for tracking and uniting families that lose contact with each other, e.g. during the catastrophe in Haiti. Very inspiring! Finally, for the real fire, Martin talked about the lack of women in our profession and why that was a real problem (believe it or not, some people choose to view that as the way of the world).
I loved this talk and so did the crowd. Martin received very strong and warm applause. Coming from Adaptiv myself, a company with a strong CSR commitment, I was very happy to see these topics broached at a geek conference.
Simon Wardley: Situation Normal, Everything Must Change
In the Cloud Solutions track, Simon Wardley of Canonical (of Ubuntu fame) had a good talk on the ominous “cloud”. He explained the process of commoditization, being driven by both suppliers competition and customer competition (getting a cheaper, better deal than your competitors). Many products and services are subject to this process, from fridges to electricity to computing resources. Personal note: That doesn’t mean all things are subject to gradual commoditization. Creative, human-intensive services, for example, are not.
Simon also spoke about having different kinds of processes depending on the level of commoditization. It’s fine to use Six Sigma or some other defined process for a commodity service, but using it for innovative or custom-building services is quite mad. And vice versa, using agile processes for commodities is also a poor fit.
Simon said the biggest myth surrounding the cloud was this: Companies actually believe they have a choice.
Eric Evans: Folding Design into an Agile Process
Eric speaks rather slowly and quietly, but with an accuracy and preciseness of thinking that is enviable. This guy knows his stuff. Me and my colleagues decided to include Eric in our we-would-love-to-be-on-the-same-team-as-you people, right next to Michael Feathers and Martin Fowler.
Eric talked about how he deliberately had kept process stuff out of the DDD book. This, however, meant that some people actually had thought that he advocated a BDUF style of design. Nothing could be further from the truth. Eric explained how he thought about the process of finding driving user stories, exploring different designs, and evaluating the designs using simple story tests. Eric is currently working on a new book (yay!) on this topic. Apparently, there is a beta book online somewhere, but I can’t find it right now. Maybe one of the readers know?
User group: Erlang
We spent most of the evening in the company of the London Erlang community, where there’s a lot of activity. First, Francesco Cesarini gave an introduction. He has the unusual background of being half Italian and half Swedish so he caught us off guard at their booth, suddenly speaking Swedish. Francesco is one of the authors of the “Erlang Programming” book. After that, Ulf Wiger gave his account of the history of Erlang. Then, Justin Sheehy took the stage and gave an emphatic talk about RIAK; the hot, NoSQL database, written mostly in Erlang. To finish off the evening, Joe Armstrong, was to speak, but by then we were extremely hungry so we had to leave (sorry Joe, our loss).
The Blogging Terrier | Den bloggande terriern 
Pingback: Gems from QCon London 2010, Part 1 « Den bloggande terriern
Even though I was there this was a rewarding read.
Thanks! 🙂
Eric’s draft: http://domainlanguage.com/processdraft/
Ah! Cool, Peter. Thanks!
I heard Martin Fowler’s talk on women (and minorities) in software in Austin, and I’m glad to see he is (unsurprisingly) keeping it a part of his program. I agree with him that there really are not enough people talking about this issue _in the right way_. To often, the question becomes “How do we get more women and minorities interested in software?” when the real problem is “How do we stop driving them away?” The first is a bit of a condescending view that puts the root of the problem on women and minorities, but really the problem is that in many cases, we in the software community have not been appropriately welcoming to those who _are_ interested.
We have in many ways focused our field on competition and elitism and arrogance–all very traditional, hegemonic (and, masculine) ways of thinking. We claim that the IT world is a meritocracy, saying “if they don’t understand, then maybe they shouldn’t be doing the job”. Why can’t we instead say “how do we help everybody succeed?” or “what can we do to include everyone who wants to learn?” By focusing less on “how do I win” and more on “how can I help you win” we will as a natural course begin to create a more inclusive industry.
We need to change many things to achieve our goals. Obviously, hiring practices that focus on finding women and minorities will help (this does not mean quotas, but does require a realistic assessment of the demographics of the organization) but we also need to keep the pipeline of talented developers full too. To do that, we may need to change the way we teach Computer Science. At least here in the US, the dominant pedagogy is a “weeding out” of weak students early in the program, essentially refusing to teach the ones who are struggling. The inherent competitiveness and elitism in such a system is antithetical to inclusiveness in general, and specifically discourages women and minorities, who generally expect a more collaborative and cooperative environment. There is a lot of study into identity politics and social theory to support this claim, but I can’t provide any references because 1) I’m not a political scientist or a sociologist and 2) I haven’t actually read the seminal works but have merely seen them referenced or implied by other works. But my fiance, who _is_ a political scientist (and specialized in identity politics and politics of the oppressed) explained it all to me and I think I understood 1% of it:) As an industry, we need to lean on the universities that are training our future colleagues and employees to move towards a more inclusive mode of teaching. We can help, too, by making it a priority to mentor fledgling developers in general and women and minorities specifically. Once we take care of the pedagogy, the hiring practices should (hopefully) resolve themselves, since the population of available talent will better match real-world demographics.
So, in order to increase diversity in the field, we need to stop focusing on the need to “increase interest” and turn to inclusiveness. We have to do this not only in the way we hire software professionals, but also in the way we train them. So work on finding as many women and minorities to hire as you can, and lend a hand to the universities to make them more inclusive as well.
Of course it’s somewhat redundant that I’m arguing for inclusiveness on a blog that generally is in Swedish, but that the author has specifically decided to post in English to include those (like myself) that are linguistically challenged:)
Greg, thanks a lot for your excellent and insightful comment (more like a full blog post, actually). Like you say, it’s a multi-dimensional problem, which probably means that we need change in many dimensions. There are no easy answers, but if we don’t change, very little else matters.