2024-11-25 04:46:35 https://en.wikipedia.org/wiki/Must 2024-11-25 04:55:56 nmz: I think if you want to call it Forth the return stack access is a must. That's just an important part of what Forth is. Could you build a system that didn't have that? Sure. Most straightforward programs don't involve you directly manipulating the return stack. It gets done for you by the system words. So it doesn't HAVE to be visible. I just wouldn't feel like calling such a system 2024-11-25 04:55:59 Forth, though. 2024-11-25 04:57:37 Basically >r and r> are the only words that modify it except for call and return operations and do loops. i and j touch it, but don't modify it. 2024-11-25 04:58:21 It would be easy to write >r and r> versions that used an array for the storage instead of the return stack. 2024-11-25 04:58:45 So you could still have them for temp data storage. You only really need the real ones if you are trying to play interesting control flow games. 2024-11-25 04:59:46 I guess there are some other words that use >r, r>, or r@ themselves, like compile. 2024-11-25 07:54:16 xentrac: Yes that's what you'd want from the standard 2024-11-25 08:08:44 And where there isn't consensus you want something worth adopting, capable of unifying 2024-11-25 08:10:53 Which I think ANSI C achieved, but not ANS Forth 2024-11-25 13:00:14 I think the whole idea of a Forth standard faced a difficult problem. Part of what Chuck had in mind was a language that fit whatever hardware you had like a glove - sort of a "thinnest possible overlay." But in order to get at the portability and so on they wanted the standards team had to abstract some things. I think those two themes are just kind of at cross purposes. 2024-11-25 13:02:23 I don't the the abstract language of the standard is necessarily FATAL, but it's already pushing you a step back from "your system at hand." 2024-11-25 18:14:24 testing, please ignore 2024-11-25 18:30:28 ignoring 2024-11-25 18:34:26 F A I L E D