2021-11-16 16:16:58 <neuro_sys_  > Is REFILL similar to doing TIB #TIB ACCEPT?
2021-11-16 16:28:46 <neuro_sys_  > I'm doing subroutine threaded code, and it seems like there's no need for a code field, since colon definitions are a bunch of call instructions. (Or rather code field is the first cell following nfa). Am I missing something potentially? I noticed ;CODE doesn't seem to be needed either (i.e. no need to modify the code field of the last created word to jump into the next cell).
2021-11-16 16:43:50 <joe9        > Isn't colorforth subroutine threaded?
2021-11-16 16:44:25 <neuro_sys_  > I never checked it
2021-11-16 17:17:27 <lispmacs[work]> neuro_sys_: My understand, looking at MOVING FORTH, is that there are several different kinds of code fields usually, and a feature of FORTH is user is supposed to be able to define their own in addition to yours
2021-11-16 17:17:59 <lispmacs[work]> the typical 4 code fields are ENTER, DOCON, DOVAR, and DOVOC
2021-11-16 17:19:46 <lispmacs[work]> under Subroutine Threading: "STC is like DTC in that the CPU jumps directly to the Code Field of a Forth word. Only now there is no NEXT code, no IP register, and no W register. So, there is no choice but to use a JSR in the code field, since this is the only way to obtain the Parameter Field Address
2021-11-16 17:20:01 <lispmacs[work]> then a little later...
2021-11-16 17:21:08 <lispmacs[work]> "We can still speak of a Code Field and a PF in STC. In every "class" of Forth word except CODE and colon definitions, the CF is the space occuipied by a JSR or CaLL instruction, and the Parameter Field is what follows [...] the meaning of of Parameter Field becomes somewhat fuzzy in CODE and colon definitions
2021-11-16 17:21:58 <neuro_sys_  > Thanks for checking the article and finding it out!
2021-11-16 17:22:51 <lispmacs[work]> neuro_sys_: recommend buying the book
2021-11-16 17:23:00 <lispmacs[work]> by Brad Rodriguez
2021-11-16 17:24:00 <neuro_sys_  > On another note, I find the topological order/dependency between word definitions rather difficult a problem. Whenever I start defining a core word, I need another word, and then yet another, so on. The list of words ordered according to their dependency is not listed anywhere for a particular forth, right? Or perhaps that would be just the order of the words in the dictionary as they are inserted to some extent.
2021-11-16 17:43:39 <veltas      > neuro_sys_: Yeah, the code field is for indirect threaded only (and token... which is like indirect indirect threaded)
2021-11-16 17:44:06 <veltas      > It's mentioned in Starting FORTH etc because that is the classic implementation
2021-11-16 17:46:51 <veltas      > I think the usage of 'code field' in moving forth is a little fluid
2021-11-16 18:11:47 <neuro_sys_  > Right