Scheme Workshop 2019


ICFP 2019 logo, based on the Ampelmännchen

Wed 30 Oct 2019

When working on software projects, it's easy to forget that one's collaborators are real people, not just disembodied voices somewhere in cyberspace. That's one of the reasons I like to attend the Scheme and Functional Programming Workshop each year. The talks are fascinating, and it's great to spend time with real people doing interesting things with Scheme. This year's workshop, like many in the past, was held in conjunction with the International Conference on Functional Programming.

For many of the workshop's twenty-plus years, I couldn't get away from work to attend it, so I only experienced it by reading the proceedings. Because many people miss it but would like to see the talks, I've been volunteering as videographer since 2016. I'm an amateur, but using a camera, a lapel microphone, a screen capture device, and a Scheme program (as yet unpublished) that wraps around FFmpeg, I'm able to produce videos with good sound quality that show the on-screen demos and slides clearly. Here are the YouTube playlists for the past four years:

I enjoyed all the talks this year, but I especially recommend two. First, Paulette Koronkevich's Reigniting Fuse, an Online Partial Evaluator for Scheme (YouTube). In this short talk, she does a great job explaining online partial evaluation, an old idea in static analysis that deserves more attention.

Second, I recommend Andy Keep's Clever Representations & Small Optimizations in Chez Scheme (YouTube). This was the afternoon keynote of the conference, so it was a bit longer, which gave Andy time to give a detailed explanation of how Chez Scheme's compiler and runtime system produce such fast code.

After the workshop, I explored Berlin for a few days, then attended Summer BOB 2019 (YouTube), a mini-conference that explores technologies beyond the mainstream and asks the question, What happens when we use what's best for a change? There were many excellent talks, including A Functional Reboot for Deep Learning (YouTube), a thoughtful talk by Conal Elliott about finding the essence of deep learning. I'm a novice at deep learning, so I'm not sure how to evaluate the strong criticisms he made of current practice, but since he did co-invent the functional reactive programming paradigm, it seems like he might have good ideas. He's definitely a clear thinker.

Attending the Scheme Workshop is one of the highlights of my year every year. I hope that the feeling of excitement and intellectual exploration from the workshop comes through in the videos.


[camera] The camera was donated by my friend Philip Greenspun, author of Greenspun's tenth rule. Thanks, Phil.