Practitioners at Academic Conferences

Next year, my current home base, Gothenburg, can count itself happy to host the annual iteration of ICSE, the premiere academic conference on software engineering. I am pretty psyched about the event already — Ivica and the team are surely going to do an awesome job, and you should totally join.

Now, this blog entry is not so much about next year’s iteration, but about it ICSE in general. I have by now attended five iterations of this conference (Zurich, San Francisco, Florence, Austin, and Buenos Aires), and something that has stuck with me is the sheer amount of lamenting there usually is about how few practitioners show up. Florence in particular stayed with me as basically a chain of panels where academics talked to academics about how to get practitioners to attend ICSE. Clearly, this is a question near and dear to the heart of many of my distinguished colleagues.

I am going to offer a different view here, which I suspect is not going to be shared by many of my colleagues. ICSE is an academic conference, and it’s totally fine if the audience is mostly academics. Really. Yes, even in an applied field.

Please, critical reader, stay with me for a second here before you close this article, and let me explain myself.

First of, let’s speculate a bit why practitioners are a small minority in the audience of ICSE talks. Let me ask you this — how often do you go to academic conferences in other fields, say applied maths? Presumably the answer is very rarely to never — while the talks may often be interesting, you usually only understand a fraction of what’s going on, the talks don’t directly apply to what you do, you don’t know anybody, and generally the return on invest (both, in money and more importantly time) is just not there. I’ll venture that this is exactly how many practitioners feel about going to academic conferences, including ICSE.

You are now probably thinking that this is exactly the problem — we need to make the program of ICSE more interesting and directly applicable for practitioners. Well, conferences where speakers give practical talks directed at an industrial audience exist — we call those industry conferences! While industry conferences are awesome, ICSE does not need to become one of them. It’s simply not the point of an academic conference.

I get that we as software engineering researchers should try harder to communicate our work to practitioners, and I agree with that. I just think that desperately attempting to lure practitioners to our flagship conference and have them listen in on research talks that are, on a fundamental level, not targeted at them is not the way to achieve this. Here are what I consider to be some more practical approaches to get word about your work out there:

• Don’t hope that the practitioners magically come to us, go to them. Speak about your work at industry conferences. My student Jürgen Cito is presenting at Velocity just this week, and he is by far not the only one. For another example, Georgios Gousios spoke at OSCON 2016. Look, I understand that going to industry conferences is scary. You can’t half-arse your talk there like you can for some academic conferences, and you may have the nagging feeling that the audience knows more about your topic than you do. Part of this is to acknowledge that you are suffering from Impostor Syndrome — you are an academic, so you know different things than a practitioner, and the audience understands that (and if they don’t care about your scientific perspective they are free to skip your talk). There is also no need to directly try for a flagship conference like Velocity. Just presenting at a local meetup already gives you a feel for the water, and even at a small meetup you probably reach more practitioners than with your shiny ICSE talk.
• If you want to do something more systemic, you can consider organizing your own events with a practitioner or technology transfer focus. For instance, I was very happy to speak at the Open Cloud Days in Winterthur a while back. This awesome event series co-organized by my colleagues at the ZHAW Cloud Computing lab features a combination of industrial and academic speakers, and the audience is by design almost exclusively practitioners (from local SMEs, in that case). It’s great to have such industry-academia collaboration events. I just don’t think that ICSE needs to become one of them. Again, different events for different audiences.
• Finally, you can also push your work into practice by directly collaborating with industry. For instance, when I joined Chalmers, I was excited to learn that about 80% of our master’s theses are done in collaboration with companies. Such theses are an awesome framework to get in regular contact with people from industry, and a fairly safe environment to test-drive some of your wacky ideas on real products and data. For (smart) companies, this is also attractive because it gives them a low-risk vehicle not only for recruiting, but also for innovation — if the outcome of the thesis is that such-and-such is awesome for them, the company can start pushing for it more seriously; if things did not really work out, nobody lost a lot of time, money, or face.

I get that communicating scientific results to industry is important — but I also think that an academic conference isn’t the best vehicle for that. Complaining that ICSE talks are not interesting enough for practitioners is a bit like standing up after a Velocity talk and asking “It’s all well and good, but where is the scientific novelty in that case study you just presented?”. Different talks, and events, are for different audiences. Let’s leave it at that.