Ten years after the Whyline
This year I received a 10-year most influential paper award for my dissertation work on the Whyline. This awards are a big honor, and represent some of the most substantial signals of the impact of a researchers ideas. I was humbled that such an innovative community recognized my work as a young researcher as something worthy of this distinction.
As part of the award, the conference asked me to give a 20 minute award talk. Now, these award talks are an interesting genre. They risk being self-serving, because they’re inherently about work that one has already published and already communicated to the community. Because it’s been impactful, it’s work most people probably already know. So talking about the work isn’t all that useful. But talking about it’s impact isn’t all that helpful either; that risks being too self-congratulatory. What can one say that is both about one’s work, but also useful to the community for the future?
I decided to focus on the big ideas that shaped both the Whyline and the other work I’ve done, but also the big ideas I think the Whyline revealed to me more then a decade since publishing it.
There are three ideas I shared that have long been true in science and scholarship, all of which shaped my discoveries:
- Accelerate innovation by reading. The body of knowledge published in academic archives is rich, powerful, broad, and deep. If you mine it for ideas, your discoveries will be dramatically more innovative and meaningful.
- Develop a personal intuition for SE practice. Reason isn’t enough for research. We need a deep personal instinct for the phenomena we study and impact, and that expertise goes far deeper than our ability to reason about the phenomena and far deeper than what we’ve written.
- Explain why your tools work. Don’t just describe and predict, but explain. Explain why tools work, because those underlying theories that tools embody are the knowledge that stands the test of time.
There were two ideas I believe the Whyline helped clarify, and that I argued are critical for the future of software engineering research and practice:
- Automation alone is insufficient. Humans are always in the loop, and must always be accounted for in the design of tools, systems, and processes.
- Augmentation > automation. Developers know far more than we think. We should be designing for augmentation, not automation, combining the insights that both developers and machines have to make much more powerful systems that automation can achieve alone.
It was incredibly fun to put together the talk, and fascinating to see how people responded on Twitter, in Q&A, and after the conference closing. I think I most reached the junior researchers in the audience; I hope I changed their views on what it means to be a scholar.
For my talk, here’s a version I recorded the day before my presentation, and ICSE’s recording of my actual talk.