What sort of language tower are you building here.

it would be interesting to reduce the bootstrap path further. assuming the hardware target has enough memory, you could boot directly into Forth then run a Scheme interpreter.

Ah. Yes. That makes sense. Cool!

and then an Ada compiler 😈​

I like where this is going. 👍

Gforth and pforth both need a C compiler so that's not a huge improvement. but there must be some Forths written in assembly or like the GreenArrays chips that run Forth natively.

just kinda daydreaming... but I like the clarity of that lisp.fs! wish we can break the dependence on C

This is why I wrote Shoehorn. The idea is to be able to bootstrap a #Forth environment for a processor of your choice in a matter of days, not months or years.

Of course, I haven't yet written said Forth yet (that's still in progress), but Shoehorn itself allowed me to write a simple, graphical demo on an emulator in a matter of a day, in time for an SVFIG presentation.

I retargetted Shoehorn to 64-bit RISC-V in about an hour. Untested, but it should mostly work fine as-is.

It's not the easiest thing in the world to use, but once you get your head around it, you should find that it is extremely productive.