2026-04-06 00:26:34 <KipIngram   > Guest4010: Create/does> implementation depends to some extent on the architecture of your system (is it indirect threaded, direct threaded, code threaded, etc.)  In my experience you can implement it fairly cleanly in an indirected threaded system, but in the others it can seem more klunky.
2026-04-06 00:28:22 <KipIngram   > We tend to think of old forth words as having two important fields - a CFA (code field address) and a PFA (parameter field address).  The CFA determines what kind of word it is - a variable, a constant, a definition, etc.  Then the PFA supplies whatever other bit of information that is needed - in a variable the PFA points to the cell that the data gets stored in, in a colon definition it points to
2026-04-06 00:28:25 <KipIngram   > the list of addresses that specify that definition, etc.
2026-04-06 00:29:40 <KipIngram   > But a create/does> word (that is, a word that is ultimately DEFINED by a create/does> defining word) needs three pieces of information.  It's sort of like a fusion of a colon definition and a variable - the colon definition part is what the does> part lays out, and the variable part is what gets specified by the create part.
2026-04-06 00:30:12 <KipIngram   > So suddenly you're faced with shoehorning three pieces of information into an architecture that's fundamentally designed to have two pieces of info associated with each word.  So some hoop jumping is required to do that.
2026-04-06 07:32:25 <pgimeno     > MrMobius: ah ok, I didn't see it implied that you were talking about adding external RAM
2026-04-06 07:35:12 <pgimeno     > I'm curious about CREATE / DOES> because the Jupiter has a different syntax. It has two kinds of words: one is DEFINER {definer word} {code to execute while defining} DOES> {code to execute at runtime, gets parameter field in the data stack} ;
2026-04-06 07:37:22 <pgimeno     > The other is {n} COMPILER {compiling word} {code to execute at compilation time} RUNS> {code to execute at run time, gets IP as parameter} ;
2026-04-06 07:39:07 <pgimeno     > But it appears that CREATE ... DOES> does not do the same. Will check.
2026-04-06 07:45:37 <pgimeno     > OK, got it, it seems that DEFINER {word} {creation-time action} DOES> {runtime action} ; is written in standard Forth as : {word} CREATE {creation-time action} DOES> {runtime action} ;
2026-04-06 20:54:35 <Guest7919   > Hi, is there any forth's source code from moore, other than the one from 1968?
2026-04-06 21:18:27 <tpbsd       > Guest7919, have you seen the 'starting Forth series of tutorials by Brad Rodriguez ? The Mecrisp-Stellaris forth release has tons of Forth source code for Cortex-M assembly
2026-04-06 21:19:17 <tpbsd       > Guest7919, and it's well commented
2026-04-06 21:23:27 <Guest7919   > thanks, yes I have read Brad Rodriguez, however I'm trying to find some source code from moore which is proving impossible tpbsd
2026-04-06 21:24:40 <tpbsd       > Guest7919, I think it was because all of More's work was proprietary. Back then FLOSS hadn't yet been invented
2026-04-06 21:25:28 <tpbsd       > all software had to be paid for, ATT Unix was $10,000 a seat back then
2026-04-06 21:25:32 <Guest7919   > I always thougth he got to release full source code of colorforth's tpbsd? maybe I got it wrong
2026-04-06 21:27:02 <Guest7919   > that's very sad, not need to see okad, latest version of colorforth or anything propietary, but man, just a reference implementation from it's original author I don't think is too much to ask
2026-04-06 21:28:52 <Guest7919   > or a paper or something explaining it from his point of view in a way one can follow along and implement it
2026-04-06 21:29:27 <Guest7919   > not even source coude would be needed that way. It's a real shame forth will die when he does
2026-04-06 21:56:16 <tpbsd       > Guest7919, I'm pretty sure Forth will never die
2026-04-06 21:57:28 <tpbsd       > Guest7919, Moore only invented forth, thousands of others made it famous, including the incredible Elisabeth Rather with Forth.inc
2026-04-06 21:58:23 <tpbsd       > Guest7919, Forth has moved far beyond what Moore imagined, it has a life of it's own
2026-04-06 21:58:29 <Guest7919   > yes, and what about those showing interest in the original thing tpbsd?
2026-04-06 21:58:44 <Guest7919   > that's not the same forth I'm talking about
2026-04-06 21:59:58 <tpbsd       > Guest7919, theyre in the same boat as the Citroen DS21 Pallas, the Tektronix 7000 series scopes and every other legend of the past
2026-04-06 22:00:47 <tpbsd       > Guest7919, everything turn to dust given enough time, only our memories last 'forever' :)
2026-04-06 22:01:23 <Guest7919   > I can only say I'd like to know more about the original idea, not about a standard's comitee understood it, or the many that have implemented it
2026-04-06 22:01:59 <Guest7919   > for sure more than one implementation brings something new, or pretty handy or usable, but again that's not what I'd like to see
2026-04-06 22:02:15 <tpbsd       > Guest7919, Moore never developed a 'Free Software' attitude, hes from the 'IP is precious' era when a computer was so expemsive that only a large company could own one
2026-04-06 22:02:41 <Guest7919   > that's a real shame, no need to release a full featured forth
2026-04-06 22:03:05 <Guest7919   > just something that shows his idea, or either to write about it as I said
2026-04-06 22:04:34 <tpbsd       > he was famous for Forth already, he had to elk out a living how he could for decades. Even Forth didn't make Moore rich. Have you read Moores autobiography series ?
2026-04-06 22:05:01 <Guest7919   > no, I haven't
2026-04-06 22:05:21 <tpbsd       > he was quite poor for much of his life, it's a fascinating read over several 'blogs'
2026-04-06 22:05:53 <Guest7919   > I'll look for them, nonetheless, I mean
2026-04-06 22:05:55 <tpbsd       > he simply had to survive, no one was paying him to release free software
2026-04-06 22:06:38 <tpbsd       > I think you wont find one piece of Free Software ever written by Moore
2026-04-06 22:07:05 <tpbsd       > he simply is from another era, and Im about his age, so I remember it well
2026-04-06 22:07:05 <Guest7919   > if you got famous at some point because of something like forth, even if you are poor and need to work for a living, wouldn't you spend a while now and then to leave heritage to the rest?
2026-04-06 22:07:46 <tpbsd       > I think thats entirely up to the individual
2026-04-06 22:08:47 <Guest7919   > that's so true, but again, it's not even about code or a reference implementation, some sort of essay explaining in a way others can understand, at least the core idea
2026-04-06 22:09:03 <tpbsd       > Moore is very old now and will pass on soon I think. He's probably more concerned about his health nowdays
2026-04-06 22:09:25 <Guest7919   > yes, I think the same
2026-04-06 22:09:50 <tpbsd       > in his 'blogs' he wrote about his love of long distance trail walking, so he did keep pretty fit most of his life
2026-04-06 22:10:25 <tpbsd       > he lived in remote towns that had lots of long senic trails
2026-04-06 22:10:32 <tpbsd       > scenic
2026-04-06 22:11:28 <tpbsd       > even if you offered him a million dollars to write up his life and code as open source, I think he's too old now and it cant happen anyway
2026-04-06 22:11:49 <tpbsd       > that time has passed, hes from a different era
2026-04-06 22:11:58 <Guest7919   > and why is it no vendor has shown interest, there are many fabs the us, microchip, texas instrument, dallas semiconductos, faichild, etc.?
2026-04-06 22:12:16 <Guest7919   > yes, it's a shame
2026-04-06 22:12:27 <tpbsd       > shown interest in Forth ?
2026-04-06 22:12:50 <Guest7919   > and in forth at vlsi level, chips
2026-04-06 22:13:15 <tpbsd       > Moore started 'green Arrays' with his GA144 chips, but it seems to have faded mostly away
2026-04-06 22:13:57 <Guest7919   > it seems to me a giant, like, say microchip, could have produced a line of these
2026-04-06 22:14:01 <tpbsd       > you can get Mecrisp-Ice , it's Mecrisp-Stellaris on a FPGA
2026-04-06 22:14:51 <tpbsd       > Forth has been efectively dead for a long time, there is no commercial interest in it
2026-04-06 22:15:33 <Guest7919   > well, if you don't make it available you are not really helping either
2026-04-06 22:16:25 <tpbsd       > it's only due to the fact that Forth is the easiest OS in the world to build/design that it still lives, with thousands of unfinished Forth implementations clogging up the repos of SourceForge and Git*
2026-04-06 22:16:55 <Guest7919   > maybe offering a free IP license to those willing to produce it would have helped, then you could still have the real chip to be sold
2026-04-06 22:17:39 <tpbsd       > sure, but it was Moore's IP, complain to him
2026-04-06 22:17:50 <Guest7919   > I'm sure if it would have been made available, plenty of universities would had use it, and that's the beginning of it
2026-04-06 22:18:07 <Guest7919   > haha, I'm not complaining, just giving out ideas
2026-04-06 22:18:10 <tpbsd       > there are hundreds of people that have worked their whole lives making Forth available
2026-04-06 22:19:01 <tpbsd       > and until RMS came along and invented FLOSS, no one had the tools to make a Forth
2026-04-06 22:19:19 <Guest7919   > oh, I didn't know that
2026-04-06 22:19:21 <tpbsd       > it all started with RMS, he's the hero here
2026-04-06 22:19:55 <tpbsd       > Moore is just another enginner making a living in comparison
2026-04-06 22:20:23 <Guest7919   > that feels sadly unfair
2026-04-06 22:20:26 <tpbsd       > what did you use to write programs before Gcc ?
2026-04-06 22:20:56 <tpbsd       > andswer = whatever you could *afford* because nothing was free
2026-04-06 22:21:08 <tpbsd       > remember 'shareware' ?
2026-04-06 22:21:10 <Guest7919   > I have always been programming C since 82, don't even remeber the names of all the compilers
2026-04-06 22:21:29 <Guest7919   > the computes I had access to came with full manuals
2026-04-06 22:21:46 <tpbsd       > academia ?
2026-04-06 22:21:53 <Guest7919   > no
2026-04-06 22:21:59 <tpbsd       > manuals cost a lot of money
2026-04-06 22:22:24 <Guest7919   > we just run a business as a family, and so I had access to computers long before the pc
2026-04-06 22:22:29 <tpbsd       > remember AT&T Unix was $10,000 USD a SEAT
2026-04-06 22:22:50 <Guest7919   > and these came with manuals for both, the hardware and the compilers and operating system
2026-04-06 22:23:01 <tpbsd       > All the unix's are free now, times are different
2026-04-06 22:23:25 <tpbsd       > but only because the manuals are in PDF do you get them etc
2026-04-06 22:23:59 <Guest7919   > yes, back in my days there weren't even much people you could ask
2026-04-06 22:24:14 <Guest7919   > it was either the manual or nothing
2026-04-06 22:24:27 <tpbsd       > I remember wanting basly to learn C back in 1996, and a DOS, Windows user there was nothing agailable that wasn't buggy shareware
2026-04-06 22:25:22 <tpbsd       > I stumbled across DJGPP which was a windows port of Gcc by DJ Delorie, and that changed my life
2026-04-06 22:25:33 <Guest7919   > c for me was easy to learn, even if I didn't know english which was the language used in the manuals
2026-04-06 22:25:41 <tpbsd       > I soon migrated to Linux full time
2026-04-06 22:25:53 <Guest7919   > yeah, DJ Delorie is a crack
2026-04-06 22:25:58 <tpbsd       > C is easy to learn
2026-04-06 22:26:19 <tpbsd       > Forth is also easy to learn, but hard to master ;-)
2026-04-06 22:27:13 <Guest7919   > what I like the less is the mix of postfix and infix, I'm rethinking it out of the box to see if a full postfix implementation would bring me something
2026-04-06 22:27:13 <tpbsd       > remember 'Borland C' ? that was about $90 iirc
2026-04-06 22:27:23 <Guest7919   > yes, used it under DOS
2026-04-06 22:28:26 <Guest7919   > that was the first time I got to see anything resembling an integrated ide
2026-04-06 22:28:27 <tpbsd       > I had just purchased the shrinkwrapped Borland C box set when I discovered Gcc, and I never used the Borland stuff. I threw it out a decade later
2026-04-06 22:28:28 <Riviera     > I have to admit that I miss this software; somehow I never felt as comfortable in any IDE / editor / whatever than I did in the borland products,
2026-04-06 22:28:39 <Riviera     > perhaps even including the similar Microsoft ones.
2026-04-06 22:28:58 <Guest7919   > you can still use it, there's freedos
2026-04-06 22:29:13 <Riviera     > Yeah well, that's not wrong, but it's limited.
2026-04-06 22:29:27 <Guest7919   > limited?
2026-04-06 22:29:39 <tpbsd       > Riviera, I'm the opposite, I could never get into the Borland stuff as the CLI of Linux and Gcc appealed to me
2026-04-06 22:29:50 <Riviera     > It's software from decades ago, I would limit myself for what I could use it for.
2026-04-06 22:29:54 <Guest7919   > I'd say it's even better than the original
2026-04-06 22:30:01 <Riviera     > Guest7919: Ah
2026-04-06 22:30:04 <Riviera     > Guest7919: misunderstanding
2026-04-06 22:30:14 <Riviera     > Guest7919: I thought you meant I could use the Borland products in FreeDOS.
2026-04-06 22:30:22 <Guest7919   > I meant to get ideas for your own ide, since you said you liked it
2026-04-06 22:30:28 <Riviera     > Guest7919: I did not mean FreeDOS was limited, but using the old software in FreeDOS.
2026-04-06 22:30:28 <tpbsd       > Guest7919, the original DOS was a unfinished heap of crap (please pardon my language)
2026-04-06 22:30:39 <Guest7919   > yes, you can, ah ok, hehe
2026-04-06 22:30:42 <Riviera     > :)
2026-04-06 22:30:55 <Guest7919   > wehn I used to use borland linux wasn't even a project yet
2026-04-06 22:31:09 <tpbsd       > who would ever use DOS having experienced Linux ?
2026-04-06 22:31:52 <Guest7919   > I still sometimes run programs I wrote with my dad in cobol, so I need it for rmcobol
2026-04-06 22:31:58 <Riviera     > tpbsd: I'm relatiely fluent in the CLI. Like, I know bash pretty much inside-out, also most of the POSIX standard, the common tools, etc. Using that stuff for over 30 years now, but still .. back then it "felt" better :D
2026-04-06 22:32:07 <Riviera     > tpbsd: But granted, it's surely my bad. :)
2026-04-06 22:32:52 <tpbsd       > Riviera, we all came from somewhere
2026-04-06 22:33:06 <Riviera     > also haha, thinking back: I remember when I first ever used Linux. I was so happy and impressed that all the fancy stuff of 4DOS or NDOS or whatever it was called was there in Linux, and even better ;D
2026-04-06 22:33:38 <tpbsd       > I remember seeing the first Windows, I thought it was a garish childrens toy software
2026-04-06 22:34:08 <Riviera     > I kind of never used Windows, used Linux before Win95 came out, and then never really had a Windows machine.
2026-04-06 22:34:20 <tpbsd       > in those days I was using Banyan 'Vines' on SCO Linux
2026-04-06 22:34:39 <Riviera     > There are phases in which I feel this to be a deficit of mine.
2026-04-06 22:35:06 <Riviera     > anyway :)
2026-04-06 22:35:25 <Riviera     > Guest7919 is right, I'll just run some of the Borland products again, perhaps that will cure me. :)
2026-04-06 22:35:28 <tpbsd       > SCO ran on the MC68000 iirc, but thats all arcane knowledge from a distant past now
2026-04-06 22:35:57 <Guest7919   > haha
2026-04-06 22:36:01 <tpbsd       > Riviera, you know the trouble with nostalgia ?
2026-04-06 22:36:16 <tpbsd       > it's never what it was cracked up to be !
2026-04-06 22:36:33 <Riviera     > tpbsd: haha, only this weekend I had a book in my hands with a chapter on how the 68k CPU is great for implementing Unix :)
2026-04-06 22:36:45 <Riviera     > (it also had another chapter saying the same for the 80386)
2026-04-06 22:37:23 <Guest7919   > many gnu tools were initially written for the Motorola 6800
2026-04-06 22:37:30 <tpbsd       > and Moore LOVED the MC6800 because it's opcoded mapped nicely into his Forth
2026-04-06 22:38:13 <tpbsd       > I learned to love assembly language on the 6800, back in 1984
2026-04-06 22:38:52 <Riviera     > On which machine? :)
2026-04-06 22:39:01 <tpbsd       > It's atill special to me, tho I now prefer the MSP430 assembly, and dont even mind Cortex-M assembly now
2026-04-06 22:39:13 <tpbsd       > SWTP 6800 !!!
2026-04-06 22:39:27 <Guest7919   > I prefer mips any day
2026-04-06 22:39:54 <Guest7919   > besides all these stm32's feel so synthetic
2026-04-06 22:40:03 <Riviera     > tpbsd: nice :)
2026-04-06 22:40:24 <tpbsd       > I purchased a second hand SWTP 6800 for $50 at a auction back then, including floppies, printer, terminal and a fantastic assembler etc
2026-04-06 22:41:25 <tpbsd       > it changed my life, tho the actuall SWTP 6800 box was a electronics nightmare of utilux connectors that caused constant problems
2026-04-06 22:41:50 <tpbsd       > and I think those connectors killed SWTP
2026-04-06 22:42:08 <tpbsd       > but what could they do ? connectors are expensive!
2026-04-06 22:43:15 <tpbsd       > Guest7919, I'm a retired electronics technician, I only use MSP430 and Cortex-m since 2014 now
2026-04-06 22:43:45 <Guest7919   > I can't stand arm tpbsd, in any form
2026-04-06 22:44:17 <tpbsd       > Guest7919, this is my Forth website : https://mecrisp-stellaris-folkdoc.sourceforge.io
2026-04-06 22:44:44 <tpbsd       > Guest7919, Im hardware agnostic being a electronics technician
2026-04-06 22:45:02 <Guest7919   > nice, I was perusing your code a few weeks back
2026-04-06 22:45:44 <tpbsd       > my site has been around a long time now, it was just a blog of when I learnt Forth initially
2026-04-06 22:46:02 <Guest7919   > I have always been so in love with mips, and nowadays anything you look at is some sort of more or less retarded son of it, with modern peripherals
2026-04-06 22:46:35 <tpbsd       > but I soon observed a total lack of modern Forth online doc, and hungered for more. I'd read all the old stuff a 100 times over by then
2026-04-06 22:47:22 <tpbsd       > so I figured even my crappy doc couldnt hurt in a world starved of Forth information
2026-04-06 22:47:34 <Guest7919   > I got to know about forth in the 90's, but never got my hands into it, is that that I have some spare time now that I want to devote to my own "standard" haha
2026-04-06 22:48:04 <Guest7919   > no, the opposite is truth, is another resource to have into account
2026-04-06 22:48:52 <tpbsd       > Guest7919, if a mcu cant run forth by itself, with no external components and with only a st-link/usb dongle for a PC terminal, Im just not interested
2026-04-06 22:49:02 <Guest7919   > in my case, I don't like the "standard", and from forth as it is, I don't like the mix of in and postfix
2026-04-06 22:49:24 <Guest7919   > or just a stub and have an umbilical system
2026-04-06 22:50:05 <tpbsd       > Guest7919, youre obviously a programmer, you and I are about 180 degrees oposite in our opinions as our specialities are so different
2026-04-06 22:50:18 <tpbsd       > Im hardware, youre siftware
2026-04-06 22:50:25 <tpbsd       > software
2026-04-06 22:50:29 <Guest7919   > I'd like to write a forth that would parse intel's xed to write and assembler, then an elf library, and let it dump itself fully
2026-04-06 22:50:39 <tpbsd       > most here are software of course
2026-04-06 22:50:45 <Guest7919   > yes, probably
2026-04-06 22:51:31 <Guest7919   > I have done embedded, though, using pretty much anything, not lately though
2026-04-06 22:52:38 <tpbsd       > Guest7919, pretty much all I do is embedded, even now in retirement
2026-04-06 22:53:31 <Guest7919   > I like it a lot, dealing with spi, sensors, etc., have done some for the industry
2026-04-06 22:54:01 <tpbsd       > Guest7919, I forgot to mention, I found Moore's autobiographic blog incredibly depressing, I felt really sorry for the poor guy, always strugling to survive and follow his Forth dream
2026-04-06 22:54:49 <tpbsd       > hehe, I have a SPI NFC sendor on my desk atm, works fine taking to my pc with FreeBSD
2026-04-06 22:55:09 <Guest7919   > haha, man, maybe it just wasn't meant to have invented forth or something, well, there's people with bad luck in life
2026-04-06 22:55:12 <tpbsd       > it's going to be talking to a Cortex-M in a box soon
2026-04-06 22:56:05 <tpbsd       > Guest7919, but I came away from reading his blog with a lot of respect for Moore, hes a legend to me now after reading it
2026-04-06 22:56:33 <Guest7919   > to me he is
2026-04-06 22:56:35 <tpbsd       > Guest7919, before I read it, Moore was just the 'inventor of Forth'
2026-04-06 22:56:57 <Guest7919   > he did something great in my opinion, simple and powerful
2026-04-06 22:58:02 <tpbsd       > and I have lots of historical legends, marconi, de flemming, brittain and bardeen, Hiram Maxim
2026-04-06 22:58:42 <Guest7919   > I see forth as the software implementation of the underlying hardware, like an emulator such that taking it to produce machine code might be much easier than with anyother languge
2026-04-06 22:58:57 <tpbsd       > to me Moore was more like the guy who struggled to commercialise the paperclip
2026-04-06 22:59:02 <Guest7919   > like a 1-1 translation
2026-04-06 22:59:08 <Guest7919   > haha
2026-04-06 23:00:00 <Guest7919   > I understand him as somebody who came with such a great idea, bear in mind forth predates C
2026-04-06 23:00:33 <tpbsd       > I have a purists view of assembly language as the 'true way' and Forth sits on top of that, and pays the price by being a lot slower for all the extra intelligence
2026-04-06 23:00:45 <Guest7919   > and the fact he got to learn, produce a cad system, and his own chips is remarkable
2026-04-06 23:01:08 <tpbsd       > I think theyre about the same age, or so close that it doesnt really matter
2026-04-06 23:01:14 <Guest7919   > yes, slower, when you dump it you dump machine code
2026-04-06 23:01:46 <Guest7919   > it does matter, in a world where c didn't exist, forth was about the best thing to be had
2026-04-06 23:01:52 <tpbsd       > I spent a couple of years writing only machine code, I hope I never have to do that ever again
2026-04-06 23:02:12 <Guest7919   > I highly enjoy writing assembler, at&t syntax, though
2026-04-06 23:02:23 <tpbsd       > finally having an assembler was like I died and went to heaven
2026-04-06 23:03:01 <tpbsd       > I designed and sold hardware I had to program in machine code with a hex keypad and 7 seg display
2026-04-06 23:03:20 <tpbsd       > it was hard, that was my 'Moore' days ...
2026-04-06 23:03:37 <Guest7919   > haha, that's a bit like, didn't you have a computer at hand'
2026-04-06 23:03:39 <Guest7919   > ?
2026-04-06 23:03:52 <tpbsd       > but even that was easier than soldering diodes into arrays to make roms
2026-04-06 23:04:13 <Guest7919   > man, some serial protocol or something
2026-04-06 23:04:36 <tpbsd       > a computer by itself without software is just a boat anchor
2026-04-06 23:05:02 <tpbsd       > I only has a Intel SDK8085 back in those days
2026-04-06 23:05:06 <Guest7919   > indeed, or without operating system given they get more complex by the day
2026-04-06 23:06:20 <tpbsd       > how far we have come!
2026-04-06 23:06:38 <tpbsd       > this is truely the golden age of technology!
2026-04-06 23:11:46 <Guest7919   > I wonder if there exists some reference about how "core" primitives develop using others?
2026-04-06 23:12:18 <Guest7919   > I have read in the TIL book some, in some vfxforth manuals others, but not much else
2026-04-06 23:12:53 <Guest7919   > I'm gonna have to gather a list and use brain power to make sense of it all
2026-04-06 23:14:24 <Guest7919   > the truly golden age of technology is gonna come, imho, when we can have processors using light
2026-04-06 23:15:42 <Guest7919   > and even if it would still being used to switch it on and off, hence, still binary
2026-04-06 23:17:22 <tpbsd       > I think that golden age is here right now :)
2026-04-06 23:19:11 <tpbsd       > in 2000 I paid $14 for a PIC16F84 mcu, last year I paid $0.20 for a MSP430 with the same specs and peripherals but 16 bit, compare to the PIC, thats a seventy times reduction in cost!
2026-04-06 23:20:13 <tpbsd       > in 1977 a MC6800 was around $170 USD !
2026-04-06 23:20:19 <Guest7919   > well, that's true, and these chineses riscv for 30 cents?
2026-04-06 23:21:00 <tpbsd       > only a programmer can love RISC-V, a electronics guy naturally prefers cortex-m
2026-04-06 23:21:12 <tpbsd       > (for all the peripherals)
2026-04-06 23:21:12 <Guest7919   > process has evolved, so costs have been cut down
2026-04-06 23:21:23 <KipIngram   > Why is that, tpbsd?
2026-04-06 23:21:33 <tpbsd       > KipIngram, ?
2026-04-06 23:21:34 <Guest7919   > I don't like riscv
2026-04-06 23:21:35 <KipIngram   > I've not worked in detail with either one.
2026-04-06 23:21:43 <KipIngram   > cortex vs. riscv
2026-04-06 23:21:51 <Guest7919   > it's a retorted free mips
2026-04-06 23:22:16 <KipIngram   > That doesn't mean anything to me; I don't really "know" either one.
2026-04-06 23:22:30 <KipIngram   > What features or lack of, etc.?
2026-04-06 23:22:52 <Guest7919   > both are risc, both sons of the same mother
2026-04-06 23:22:59 <tpbsd       > KipIngram, because cortex-m (I should have said STM32xx !!) is associated with a wide range of proven peripherals in chips using that ISA
2026-04-06 23:23:05 <KipIngram   > I'll be getting to know riscv in the coming future, but no experience yet.
2026-04-06 23:23:28 <KipIngram   > Ok - so better / friendlier peripherals.
2026-04-06 23:24:00 <KipIngram   > The ESP32 I'm looking at seems to HAVE a nice set of peripherals, but I can't comment yet on how easy there are to work with.
2026-04-06 23:24:03 <Guest7919   > while cortex-m gives your a two-stage pipeline (iirc, that's all armv6) has to offer, well cortex-m0, even a 28-pin mips from microchip gives you the original 5-stage pipelined processor
2026-04-06 23:24:31 <tpbsd       > KipIngram, but RISC-V has as yet only a subset of the stm32xx peripherals, and the ones in the RISC-V are Chinese ones
2026-04-06 23:25:05 <KipIngram   > Hmmm.  Ok.  Well, I guess I'll be finding out.  So it's actual hardware differences and not a difference in available library software?
2026-04-06 23:25:39 <tpbsd       > KipIngram, the think to keep in mind is that a sigle STM32H7xx mcu can have up to ** NINETY** complete peripherals onboard
2026-04-06 23:25:44 <KipIngram   > I'm planning to support them from scratch, so the library situation won't matter a lot to me.
2026-04-06 23:26:20 <tpbsd       > KipIngram, to me, thats correct, it's all about the peripherals
2026-04-06 23:26:55 <KipIngram   > My understanding is that there are some utilities for WiFi and Bluetooth in the ESP32-C6 immutable ROM, but it's not a fully complete stack.  Lower level stuff that has to be used as building blocks.
2026-04-06 23:26:57 <tpbsd       > KipIngram, as a Forth user, I dont care about the ISA as it's below the Forth and I only use the Forth
2026-04-06 23:27:51 <KipIngram   > I'm writing the Forth, so I have to care about that.  Though I'm planning to slap a vm of my own design in, so most of the real peripheral "development" will be done at that level, not at the hardware ISA level.
2026-04-06 23:28:15 <KipIngram   > And the vm is meant to be very "Forth friendly," so in a way I'll sort of be working at the Forth level also.
2026-04-06 23:29:19 <KipIngram   > But it'll be using things like @ and ! and cmove and so on - I'll have to create my peripheral specific lexicons.
2026-04-06 23:29:46 <tpbsd       > KipIngram, of course, as usual, youre precisely correct, Im the one lagging behind!
2026-04-06 23:30:43 <tpbsd       > KipIngram, I dont see the included WiFi and Bluetooth in mcus, as 'peripherals', tho of course technically they are
2026-04-06 23:31:30 <KipIngram   > Eventually I'll have to get on top of the instruction encoding, so that I can meta-compile the vm layer, but starting out here I'm cheating by using a Python library that does riscv instruction assembly for me.
2026-04-06 23:31:55 <tpbsd       > KipIngram, the fact that they require large and complex drivers and API's as opposed to the much simpler 'standard stm32 peripherals' causes me to think this way
2026-04-06 23:32:44 <KipIngram   > I share your feeling about "large and complex" stuff - it seems like al most everything is way more complex than it really needs to be.
2026-04-06 23:33:26 <tpbsd       > KipIngram, Im showing my age, as Im still amazed by the fact that PIO's are now included as standard in MCU's, where theyre called 'GPIO's' ;-)
2026-04-06 23:33:36 <KipIngram   > I'm going to try to avoid that - I don't have any intention of trying to write, for example, a full-function WiFi stack.  If I can just push data across a link I control both ends of, that will be good enough.
2026-04-06 23:34:21 <KipIngram   > I guess I'll want the notebook implementation to be able to do more, but it will be running under Linux so I can cheat there too if I want to and make OS calls.
2026-04-06 23:34:47 <tpbsd       > KipIngram, and as I only use Cortex-M0 atm, I have only 64KB Flash for everything including the Forth, so there is no way a WiFi and Bluetooth peripheral can ever exist on such chips
2026-04-06 23:35:21 <KipIngram   > I think my little chip has 512kB flash.
2026-04-06 23:35:27 <KipIngram   > Wait - no.
2026-04-06 23:35:59 <tpbsd       > KipIngram, no you ned external flash I think ?
2026-04-06 23:36:00 <KipIngram   > 512kB of RAM - I think it's got a few MB of flash.  It's a multi-chip module and they have an SPI flash integrated into it.
2026-04-06 23:36:03 <tpbsd       > SPI ?
2026-04-06 23:36:09 <KipIngram   > ^
2026-04-06 23:36:23 <tpbsd       > yeah, they have to do it that way
2026-04-06 23:36:44 <tpbsd       > and thats because they use different silicon processes for the MCU and Flash
2026-04-06 23:36:49 <KipIngram   > It's pretty neat, though - there's a cache sitting between the serial flash and the CPU, and you set a bunch of config for your flash and then the hardware makes it look like a RAM interface.
2026-04-06 23:36:52 <Guest7919   > there are cortex-m0 with wifi, and bluetooth tpbsd
2026-04-06 23:37:29 <KipIngram   > No cache between the onboard RAM and the CPU, though - RAM access is direct and un-cached.
2026-04-06 23:37:36 <tpbsd       > Flash nodes are gigantic, around 90nm for the cheap stuff, only STM have a 40nm Flash in their 'G' seriex STM32
2026-04-06 23:38:19 <tpbsd       > so they cant put flash and mcu plus peripherals on the same die
2026-04-06 23:38:35 <KipIngram   > I couldn't tell you anything at all about how the space inside the module is split up, but the whole thing is about a centimeter square.  The little board it's mounted on is about an inch square.
2026-04-06 23:38:57 <tpbsd       > Guest7919, that makes no sense to me, how does a M0 have the brainpower to spare for wifi and bt ?
2026-04-06 23:39:26 <Guest7919   > because are macrocells, by the same rule you wouldn't have usb tpbsd
2026-04-06 23:39:27 <KipIngram   > 14 through-hole header connections.
2026-04-06 23:39:46 <tpbsd       > KipIngram, it's a SOC, where there are a number of dies on the substrate
2026-04-06 23:39:52 <KipIngram   > Sure.
2026-04-06 23:40:17 <KipIngram   > What I meant above is that I don't know what fraction of the space inside is flash, what's cpu, etc.
2026-04-06 23:40:43 <KipIngram   > All I really know right now is what they managed to cram inside.  :-)
2026-04-06 23:40:59 <tpbsd       > but as I only design small devices using STM32F0 I dont use any of that advanced computing stuff
2026-04-06 23:41:09 <KipIngram   > And I felt like getting it all in that small package for just $6 was a pretty good bargain.
2026-04-06 23:41:29 <Guest7919   > have you tried nuvoton's tpbsd? no
2026-04-06 23:41:35 <tpbsd       > KipIngram, it is amazing tech for sure, but far too complex for me
2026-04-06 23:42:18 <tpbsd       > Guest7919, no, Ive only been using MSP430 and Cortex-M0 since 2014, it's taken me that long to master them
2026-04-06 23:43:11 <Guest7919   > that long? armvm6 is not that complex? or you mean the peripherals? I have never used ti's ones
2026-04-06 23:43:14 <KipIngram   > A while back I was looking at this one:
2026-04-06 23:43:16 <KipIngram   > https://www.analog.com/media/en/technical-documentation/data-sheets/MAX32655FTHR.pdf
2026-04-06 23:43:19 <tpbsd       > Guest7919, as a electronics guy I have limited myself to two ISA's, so that I may master their strengths
2026-04-06 23:43:51 <KipIngram   > That one has both a Cortex M4 main processor and a RISCV processor for low level hardware stuff (precise BT timing, etc.)
2026-04-06 23:44:19 <KipIngram   > I was pursuing the same idea - it has a nice flock of onboard peripherals, but it's quite a bit larger, and around $30.
2026-04-06 23:44:37 <KipIngram   > When I found the ESP32-C6, I pretty quickly decided I liked it better - smaller and cheaper.
2026-04-06 23:44:52 <tpbsd       > Guest7919, only programmers jump from one ISA to another as they admire the tech in use, but they generally have no idea about the peripherals or how to deploy the chip in a working product in the real world
2026-04-06 23:45:28 <KipIngram   > Yeah, there's just a lot less interest in that level of things than there used to be.
2026-04-06 23:45:51 <KipIngram   > A few decades back you could walk into any good bookstore and find scads of books addressing low-level hardware stuff.
2026-04-06 23:46:11 <KipIngram   > But that's pretty much all dried up, and these days everythings about web, big data, etc.
2026-04-06 23:46:14 <Guest7919   > admire or get bored of, I thoroughly studied armvm6 back in the days tpbsd
2026-04-06 23:46:17 <tpbsd       > KipIngram, I have wireless modules, ie 'zigbee' that I will use with a cortex-m0 if I need wireless, but the Zigbee is self contained, I dont need to worry about drivers or licences etc
2026-04-06 23:47:14 <tpbsd       > Guest7919, there is a difference, I havent studdied the ISA's but I have studdied and made working projects using every peripheral
2026-04-06 23:47:15 <KipIngram   > Having everything all bundled up and ready to go is nice.
2026-04-06 23:47:27 <KipIngram   > That's basically what I'm trying to create here.
2026-04-06 23:47:47 <KipIngram   > A little standard gadget with a Forth OS supporting all its peripherals, that I can just pop in wherever I want it.
2026-04-06 23:48:29 <KipIngram   > The "smarts" for any oddball project I ever fancy putting together.
2026-04-06 23:48:32 <Guest7919   > well, I also did that, as soon as I noticed the gpio macrocell in stm32's is the worst ever, I never came back
2026-04-06 23:49:06 <Guest7919   > I used to program arm7, arm9, and str755 for vfds sold in my country, though
2026-04-06 23:49:34 <tpbsd       > KipIngram, Ive always avoided the 'all in one' type of device. Get the 'super-dooper' 10 in one sawbench! it's a planer, a saw, a router, a grinder etc. It does 10 things but you have to set up for each one, and even then each thing isnt very well done, compared to specialist single devices
2026-04-06 23:50:22 <tpbsd       > Guest7919, are you referring to the first STM32, the STM32F1xx series ?
2026-04-06 23:50:33 <Guest7919   > and after all the hype, you end up noticing even nuc120's or nuc123's beat stm32's any day
2026-04-06 23:50:50 <tpbsd       > Guest7919, because I hate that model and wont use it simply because of the horror that is the GPIO
2026-04-06 23:51:07 <Guest7919   > stm32f1xx tpbsd the ones in the bluepill
2026-04-06 23:51:28 <Guest7919   > that's what I'm say, the exact same macrocell is copied and pasted to other models
2026-04-06 23:52:03 <tpbsd       > Guest7919, I dont balme you, and Im entirely the same! however ALL the STM32 after that series, i.e. the STM32F0 GPIO are a thing of beauty
2026-04-06 23:52:35 <tpbsd       > the STM32F1xx was the first and LAST to use that GPIO design
2026-04-06 23:52:37 <Guest7919   > maybe they got embarrassed at some point, it was a complete rip off
2026-04-06 23:53:03 <Guest7919   > then they complain about the chinese doing arm's for pennies
2026-04-06 23:54:30 <Guest7919   > I got to see that exact macrocell in other models by the way, tpbsd, as I say, I never came back to st neither have been involved in a hardware project since that era, at work I mean
2026-04-06 23:55:00 <Guest7919   > a company I highly respected, and ended up hating it for robbers
2026-04-06 23:59:57 <Guest7919   > the only mid-high end arm's I got to enjoy with was imx.7 series from nxp