I remember that a long time ago I was excited about the Guile Emacs project. It’s aim was to replace Elisp with Guile Scheme while maintaining comparability by making Scheme grok Elisp. To be honest, I can no longer recall why the idea excited me. It seems like empty engineering to me now.
Regardless, after a decade the project has been restarted. Robin Templeton gave a talk at Emacs Conf 2024 on the current status of the project and what they hope to accomplish. A lot of what Templeton has to say boils down to:
- Scheme being more efficient
- Scheme being a much more flexible and extendable Lisp
I think my original enthusiasm was probably the result of the promise of increased performance but now that we have native compilation, that doesn’t seem so persuasive.
A lot of Templeton’s argument involves the possibility of reducing the amount of C code. Unlike many others, I don’t see C code as bad in itself, but Templeton makes the point that those parts of Emacs written in C are not extensible in the same sense that those parts written in Elisp are.
To me, reducing the amount of non-extensible code is a worthy goal but I’m inclined to think that native compilation can do that just as well,
The takeaway is that my previous enthusiasm notwithstanding, I don’t really see the need for a Scheme Emacs. Still, it will be interesting to see how the projects plays out. Guile is a great platform and if Emacs were being written today, it would make sense to use it but that ship has sailed.