Came here to find podcast discussion … am I in the wrong place?

For the most recent episode? There's a thread going here, but it's devolved a bit into how to not be sexist basics instead of the more advanced discussion that I'd (naively?) hoped this community was ready for. You might want to use this new thread to share your thoughts and we can start a deeper discussion.

Wait… there’s a Podcast?

Yes, it’s good but sometimes very long

Long and infrequent :)

The podcast always makes me shout at the radio

I read Feminism in Programming Language Design when it came out, and I'm dipping my toes back into the FoC podcast and community. I think it's a fantastic paper, and as you all acknowledge, it seems like Professor Hermans had to take on some personal reputational risk just to publish this introductory paper, and the fact that she had to do so speaks poorly of the academic CS culture we currently have. As a relatively new academic researcher, I definitely experience the tension between the HCI-centered work that I find fascinating and the "difficult" work that I worry other people see as more legitimate. I appreciated hearing all of your perspectives on the paper. Last week, I was at PLATEAU, which is a short academic conference about HCI + PL! I'm coming from the HCI side myself as an HCI PhD student and honestly expected it to be PL and Theory heavy, and I was pleasantly surprised to find it was very balanced or even leaning HCI. We saw talks about languages for visualization, making formal methods more usable, CS education for advanced PL topics, and more. There were even some talks with user studies. It was a great experience and I'll definitely be returning next year. Paraphrasing a friend of mine who is a PL academic, "after assembly, everything else was in some way HCI". All of the work on functional programming and type systems and formal methods and proofs is ultimately in service of humans being able to do programming better! I like to think of HCI as due diligence: if we aren't trying to rigorously validate our ideas (which includes qualitative studies!), we're not doing serious academic work.

I don’t actually buy that “it’s all HCI”. Even if it were only computers writing code, they would still build abstractions. I’m pretty sure that the relational model, purely functional and idempotent operations etc would still be useful. These might not be expressed as human programming languages per se, but even there, I suspect that the machines would invent something like programming languages as well.

That seems a little abstract. Are we assuming that these code-writing machines are intelligent? I think I'd still call that HCI.

It’s hardly Human Computer Interaction if this is just machines writing code without displaying it for humans in any way.

Someone has write it and debug it - even if it not for human consumption.

My favorite talk was probably the keynote by Shriram Krishnamurthy, who is a professor at Brown who has done some work on PL and formal methods, but also CS education. There wasn't a recording of it, but it seems like he gave a talk that he's given before, which is recorded here: It's basically an overview of some research that his lab has done around the understandability of formal methods. They found some surprising results, such as around ~9:30 in this video, where he explains this study interrogating whether maxims like "simpler examples help people understand better" are true in the context of helping them understand formal specifications. And he draws on material from education research about the "principle of contrasting cases", in which examples need to be sufficiently different to help the viewer inductively generalize their intuition of the specification from the examples.

He is an excellent speaker. Ok what was your second favourite?

Probably more directly relevant to the Future of Code community was the "Code Style Sheets" talk by Sam Cohen, which is based on this paper: https://people.cs.uchicago.edu/~rchugh/static/papers/css-jan-2025.pdf The rough idea is "what if you had alternate presentations of textual source code besides syntax highlighting? This diagram from the paper visualizes a few different ways to "style" code to show type errors and operator precedence. They also do a bunch of work to make it possible for programmers to write their own stylesheets so that they can get these sorts of visualizations in their code editors on-the-fly, making code viewers more malleable.

Nice. TYVM.