It's evals all the way up

In 2022 I wrote a LISP interpreter in PDP-11 assembly, hand-translated it into machine code, and toggled it in directly to the front panel of my PhD advisor's PiDP-11. I kept as close as possible to the system described in John McCarthy's original paper (other than fixing the known bug in eval).

This is that project. The original plan involved using it to bootstrap an even better interpreter, but I lost some momentum when I realized the thing I created had dynamic instead of lexical scoping, and also when the PiDP-11 got turned off and lost its emulated memory contents.

I gave a little presentation on it anyway; the slides for that can be found here. In short, S-Functions manipulate S-Expressions, but descriptions of S-Functions are S-Expressions. eval is the S-Function that reads an S-Expression which describes an S-Function, and does the thing that that description describes.

If you want to know more and know how to find me in real life, please do so and I'd be happy to talk about any of this stuff in more detail.