2010 July 12 23:07

Rumours of muforth’s death are exaggerated

While it might appear that work on muforth has ceased entirely, such appearances are mere illusion. A look at the Github repo will show several bursts of activity. I spent most of September & October of 2009 working feverishly on the HC08 and HCS08 support, developing a nifty simple bootloader for the chips, and mastering eldritch art of programming the flash memory (which is different on the two families!).

Then, this past February, I returned to the ARM support and hacked feverishly on that for a month or so – creating a completely new target compiler, and re-writing and augmenting the ARM kernel to be direct-threaded (rather than subroutine-threaded), and self-hosting. That was the first time I had written a target compiler that compiles names (ie, a dictionary) on both the host and target. The definitions of defining words in such a system really become mind-bending...

I’m currently experimenting with some novel (for muforth) features, such as 32-way hashing of the dictionary, robust & portable double number support (in C – God help me), and a refactoring of the “platform” support code (so muforth could, in principle, run on non-*nix-alike platforms).

I look forward to someday returning to the HC08 work. I was having a great time last fall working on it, and still have high hopes for using the chips in practice.

Stay tuned!


Read the 2009 journal.