Whether subroutine-threaded code or direct- or indirect-threaded code is faster is determined
 

Whether subroutine-threaded code or direct- or indirect-threaded code is faster is determined by the particular architecture of the processor. It can shake out one way or the other. Have you actually calculated the timings both ways for the '12?

Even when subroutine-threaded code is slower than other threadings, there are optimizations that can speed it up which are available only to it, such as converting the last call of a colon definition to a jump, eliminating a return. One can go further, using register allocations especially advantageous for a particular word, and redundant opcode elimination. By the time all of these are done, one has a native-code compiler. These will almost always be faster, even on processors where subroutine-threading alone loses speed. But the compiler itself can get pretty big.

I guess that what I really want to say is that the subject isn't simple, but that you are right at least some of the time. Whether you are right about MAXforth 3.5 remains to be seen. (Personally, I doubt it. When I back myself into the corner of thinking that the whole world is crazy but me, I try to remember to think again. The guys at NMI are pretty smart.)

Jerry

----------
Engineering is the art of making what you want from things you can get.

what I really want to say is that the su
 

Tom,

what I really want to say is that the subject isn't simple, ...snip.... Whether you are right about MAXforth 3.5 remains to be seen.
------------------------------
Tom wrote : Since F' is the simplist to disassemble , why have you not done it ? Just look at the Inner Int' . Everything hinges on it . Is this too tough for you ? You'd rather just side with the majority . In a few seconds , any F' can be evaluated for it's big blunders . No other HLL allows this .

Edited by: administrator

You made these same points on comp.lang.forth (using a much more strident tone; thank you for your restraint here) and were refuted by experts. Authors of a variety of Forths, including subroutine threaded, patiently tried to explain why your ideas might fail, giving specific examples which you seemed to ignore. What do you actually want? Nobody stops you from writing the kind of Forth you want. In fact, you can expect help with difficult problems from a large number of correspondents. Do you want the entire Forth community to ratify your views? That isn't going to happen, and it you had been listening, you would know why. Why it won't happen has nothing to do with wanting to be on the side of a supposed majority, a commitment to historical accident, or any of the other derogatory reasons you have mentioned. Your shtick is the equivalent of claiming that all bridges should be suspension bridges because other kinds use more steel for a given span, or that all motor vehicles should have dual rear wheels because they can carry more cargo that way. If you made shoes, would you claim that one size fits all?

It's clear that you've worn my patience a little thin. I want to make clear also, that although I have high regard for NMI, and for the few people from NMI with whom I've corresponded, I am in no way connected with them. I do not speak for them, nor can they limit what I say except by deleting my messages.

Jerry
----------
Engineering is the art of making what you want from things you can get.

Edited by: jya