SPLASH/ISSTA 2026

I have just returned from SPLASH 2026, which was colocated with ISSTA in Oakland, California. This was my first academic conference, and I thought it might be worth writing down some initial thoughts from the experience.

Talking to people is nice

I really enjoyed meeting lots of people working on lots of different things, although I seemed to have an uncanny ability to start conversations with people working on property based testing and/or fuzzing. I was happy with that, because I love property based testing! I've written a few tools in this space before (stateful-check in Clojure, Properb in Ruby, and in-progress Ferret for Guile Scheme), but I haven't had many opportunities to talk to people who are implementing these sorts of tools before, so that was really nice.

I also got to meet lots of people working in programming languages. There were people working on parametric polymorphism with subtyping, but more from the set-theoretic semantic subtyping angle, rather than the nominal types that I'm working with. I managed to mostly understand the conversations I was in, so I'm starting to believe I might not be an impostor after all.

PLMW wasn't quite what I was expecting

I was able to attend OOPSLA because of funding from the Programming Languages Mentoring Workshop. This was an extra day at the beginning of the conference, focused at late undergrad and early PhD students. It was a nice day with an icebreaker activity, several talks about technical and social topics, and two panels about research and academia.

On the whole, though, I felt like it wasn't exactly calibrated for me. Talking with my supervisor, he mentioned that one difference for me is that an Australian PhD is relatively short, so I've already settled into a project. Most people with under a year in the US system would still be looking for a project, so PLMW tries to give a survey of different topic areas, and helps people to think about how to decide on their project.

Overall, it was fine, and I really appreciated the explicit expectation that I would attend the rest of the conference.

A sense of dread

One thing I didn't expect from OOPSLA was the general sense of dread that I got from many of the longer-term members of the community. In one of the PLMW talks someone was presenting on the Hazel programming environment and said something along the lines of "I feel like we've got this system to the point where it's really nice to use, just as nobody writes code like this any more...". Similarly, in more casual conversations we often ended up discussion questions like "does anything we're doing in PL even matter when LLMs are writing all the code?" and "are the skills we're teaching actually worth learning any more?"

On Monday night there was a "40th anniversary of OOPSLA" which had this tone as well. The discussion involved a lot of looking back at the fun of the past, and a bit of hopelessness when looking to the future. There was some discussion of PL being important for verification of LLM-driven code and stuff like that, but it felt more like grasping at something, rather than a real enthusiasm for the idea.

I felt like across the week like there was a vibe of "industry used to care about PL, so we were cool and had lots of funding, but now those days are over".

All hope is not lost

I didn't get this sense of dread from everyone, though. There were lots of people doing cool work, and didn't seem particularly worried about LLMs taking over everything. I got to meet and hear from people working on all sorts of interesting things, like the insane ambition of Verse, or a really cool system for arena allocation.

I don't really know what to do with this whole experience yet. Due to the time zone shifts I'm still a bit sleep-deprived, but as I get some rest over the next few days I'm hoping I might get some time to process this experience.