something that just occurred to me: had Chuck Moore been writing for an ARM, colorForth may have looked very different. it's a singularly poor fit for a RISC; i mean, it's not a great fit for the x86, to be honest (Intel have always assumed / supported only one stack; while it's less glaring than in the 8080, it's still a pain) but for a RISC it's even worse because of the use of jump & link instead of call instructions (and load/store architectures generally)

@millihertz
Did you have a rather extensive post about how C was particularly ill-suited for early Intel CPUs, and how C has been probably adopted for other reasons, not because it was "close to the hardware" and how later CPUs added instructions to smooth the edges.
I know that this is the opinion you're likely to hold, but I can't seem to be able to find that particular post. Am I making things up? 🤔
I think it was late last year or maybe early this one.

@m0xee it does sound like the kind of thing i would say... but the "search your own posts" function only goes back 30 days here, so i'll have to download my history and search it offline, and get back to you

@millihertz
Yeah, likewise here — this instance doesn't have full-text search at all, I have my own Pleroma instance, but as it's running on an old PowerPC Mac Mini, a pretty anemic machine, it fails to find anything.
No need to go into too much trouble, but if you come across this post, drop me a line. Thank you!