Dagstuhl Seminar 18061: Evidence about programmers for programming language design

Last week I was fortunate to attend a seminar at Schloss Dagstuhl. It was a unique, immersive, possibly once-in-a-lifetime experience of computer science flavoured academic inquiry. Seminar 18061 was titled: Evidence About Programmers for Programming Language Design. That’s quite a mouthful, but the basic idea is this – too often, design choices that go into new programming languages are not evidence-based. One of the main objectives of this seminar was to discuss what evidence is out there, what evidence should be taken into account that is not currently, and what are some possible routes to get more evidence-based decisions into programming language design. There were about 35 or so researchers from the HCI, Computer Science Education, and Programming Language communities along with several people from industry including Google, Microsoft and code.org. The organisers of the seminar were: Stefan Hanenberg (Universität Duisburg-Essen, Germany), Brad A. Myers (Carnegie Mellon University – Pittsburgh, US), Bonita Sharif (Youngstown State University, US), and Andreas Stefik (Univ. of Nevada – Las Vegas, US).

Andreas is the lead author of a programming language called Quorum and he and I had several (often very late night early morning) conversations about how our research can inform each other’s work. Quorum has several features that make it pretty unique as a programming language, chief amongst them being that the development team strive for Quorum to be as evidence-based as it can. Quorum itself will be the topic of a forthcoming post here, so watch this space.

Below is a list of some materials related to this seminar:

• Andy Ko has provided an excellent trip report that is well worth a read
• Felienne Hermans live-blogged many talks
• Amelia McNamara recorded her reflections
• The Dagstuhl seminar materials (slides, abstracts, etc.) are here
• The slides from my talk: Compiler Error Message Enhancement: What’s Been Done, What Needs Doing. As you can see below, I got a bit passionate at times! (photo provided by @Felienne). The talk was based on several of my papers, all of which can be found at www.brettbecker.com/publications. On that page, [23] is a recommended entry point into the literature on enhanced compiler error messages.

Below are some more photos that were taken during the seminar, and around Dagstuhl’s excellent facilities. Click for larger versions. I wish I took more, but I was super engaged, and normally very tired! The general consensus seemed to be that sleep was much less important than capitalizing on a rare opportunity to discuss things face to face with fellow researchers. Whether or not that was the most optimal approach, however, we did not test – there wasn’t enough time to introduce a randomized controlled trial! The first photo is from Johannes Hofmeister:

dig