I wrote a couple Forth compilers for the Forth chips of the 80s. Yossi (the author) does a great job conveying how attractive Forth can be, and also how it fai

I wrote a couple Forth compilers for the Forth chips of the 80s. Yossi (the author) does a great job conveying how attractive Forth can be, and also how it fails to encode algorithms that are less than dead simple. I wonder how much of the stack manipulation problem could be solved with say four register words A, B, C, D (and A!, B!, C!, and D!) that access registers understood not to be trashed by primitives (CODE words) but might be trashed by definitions (COLON words). Otherwise, once you start trying to fix these problems, you end up with Scheme.