2022-07-06 00:18:48 That's just info, right? 2022-07-06 00:21:15 are you asking me or did I miss something 2022-07-06 07:47:04 is there a baremetal x86 forth that I could run off a flash drive? 2022-07-06 07:53:09 there's a ton 2022-07-06 07:57:19 is there one that's compatible with Starting Forth though 2022-07-06 08:20:28 you can easily make it compatible 2022-07-06 08:20:42 either change the book or change Starting Forth 2022-07-06 08:21:27 there's two editions; a badly regarded ANS one 2022-07-06 08:21:49 and an earlier FIG (if i remember correctly) 2022-07-06 08:22:22 the changes required were pretty minimal 2022-07-06 08:24:12 oh, i had heard somewhere that the changes were quite bad 2022-07-06 08:25:30 the changes required for the FIG-forth Starting Forth code to work under pforth or gforth on a 64-bit system 2022-07-06 09:40:14 Yes, I was asking. That git repo - it's just a file with the names of a bunch of open source projects? 2022-07-06 09:40:29 Well, projects, languages, resources, etc. 2022-07-06 09:41:41 dinklebink: there is a project out there called ForthOS that might let you do that. It's designed to run on standard PCs (well, PCs from some time in the past) bare-metal, and it also has full source and is capable of re-compiling itself. 2022-07-06 09:41:59 I've installed it on an old computer and played with it some, so I know it will run, but I did't explore it very deeply. 2022-07-06 09:42:14 https://github.com/Zer0cul/pijFORTHos 2022-07-06 09:42:29 I don't know for sure that's the one I saw in the past, but it's got the right name. 2022-07-06 09:44:15 I don't recall it being delivered in a USB stick deployable way, but the source is there, so you could probably make that happen. 2022-07-06 09:44:40 You might also look into Chuck Moore's ColorForth. It's available, I think. 2022-07-06 09:45:01 It's made to boot from a disk, but could probably be made to boot from a stick as well. 2022-07-06 11:10:33 thrig, is there an old forth that I could use, like for an 8-bit PC, with this version of Starting Forth http://home.iae.nl/users/mhx/sf.html 2022-07-06 11:10:46 is this the FIG or ANS version I don't even know 2022-07-06 11:14:29 why not find any usable forth and start working through the text 2022-07-06 11:20:07 I am but I kind of want to do something wonky 2022-07-06 11:20:34 Buy a cheap copy of the book, and use a netbook to do the book with, bc the internet distracts me a bit 2022-07-06 11:21:10 not even that, I just want to go to a cafe and do the book for a while 2022-07-06 11:21:48 also I have a DE-10 nano I wanna use 2022-07-06 11:24:05 imode: have you looked into equivalence graphs and equality saturation? 2022-07-06 11:24:20 might be interesting since you deal with rewrite systems 2022-07-06 11:51:04 siraben: we moved away from rewrite systems after the first iteration. 2022-07-06 12:17:13 I was just curious about headless forth system (for no dictionary overhead) and was wondering if anybody was using a Forth that had headless capability 2022-07-06 12:17:43 wouldn't you still need a dictionary? 2022-07-06 12:18:12 siraben: we moved to tiny queue machines linking with eachother and moving around eachother. 2022-07-06 12:39:26 for what: 2022-07-06 12:40:17 You mean a tethered system, with the dictionary on a connected host? 2022-07-06 12:46:59 KipIngram: yes 2022-07-06 12:47:22 Somebody just told me about this one, which is interesting: https://mecrisp-across-folkdoc.sourceforge.io/index.html 2022-07-06 12:47:36 but only for the MSP430xxxx targets 2022-07-06 13:21:19 $0.40 is a good price 2022-07-06 13:27:46 thrig: I suppose so, although RP2040 is available for $1/each 2022-07-06 13:28:00 non-bulk price 2022-07-06 13:28:43 though you have to have to buy externak flash for the RP2040 2022-07-06 18:02:28 That's still nice for a buck. 2022-07-06 18:03:26 Enough RAM to do a lot, too. 2022-07-06 20:16:07 Hmmm. That across compiler linked above says that one of its optimizations is that it automatically inlines definitions that are only used once. 2022-07-06 20:16:16 That would require multi-pass, wouldn't it? 2022-07-06 20:17:01 I presume that means it eliminates the definition and just puts in at the use point. But when the definition is made, and the first time it's used, a normal Forth would have no idea if it was going to be used again or not. 2022-07-06 20:17:22 With two passes it would be easy of course. 2022-07-06 20:20:11 It's a nice idea - you could factor based on logical function to your heart's content. 2022-07-06 20:20:28 Without any overhead unless you used a definition more than once. 2022-07-06 20:30:26 KipIngram, just joined. are you looking at Mecrisp Across by chance? 2022-07-06 20:41:17 Yes; lispmacs[work] linked it. 2022-07-06 20:41:25 I just took a brief gander. 2022-07-06 20:42:15 Also peeked at the rp2040 coming in at $1. 2022-07-06 20:43:30 it works by keeping all the source on the ARM host and only transferring machine code to the msp430 when youre ready 2022-07-06 20:45:17 Yes. 2022-07-06 20:45:21 Nice idea. 2022-07-06 20:51:25 it's really amazing if you have a chance to look at it again 2022-07-06 20:51:37 unbelievable optimization 2022-07-06 20:53:16 Yeah, that's what caught my attention, actually. 2022-07-06 20:53:25 I decided it more or less has to be multi-pass. 2022-07-06 20:53:25 like : foo 5 ; : bar 7 + ; bar optimizes to loading a constant into a register 2022-07-06 20:53:49 woops, i meant : bar foo 7 + ; 2022-07-06 20:54:26 Yeah, that seemed nice. The one I noticed was that it says it automatically inlines definitions that are only used once. 2022-07-06 20:54:39 I couldn't think of a way it could do that without making multiple passes. 2022-07-06 20:55:17 right. I think what happens on the ARM side is nothing like a traditional forth. it stores extra info to let it do all that crazy optimizing 2022-07-06 20:55:22 But, there's a powerful host creating code for a lightweight processor - makes sense to do everything you can to make the target code better. 2022-07-06 20:55:35 rather than keeping a traditional forth dictionary 2022-07-06 20:55:55 Well, it would have *something* of that ilk on the host, I'd assume. 2022-07-06 21:00:02 I need to give some thought to my setup - future plans include using it on gadgets I build. I'd mostly just assumed I'd put enough resources on the gadget to host it, though. 2022-07-06 21:00:13 But maybe I should think further than that. 2022-07-06 21:01:37 I have my headers in a separate memory region already, but the header part contains the CFA, and that's what gets pointed to by definitions. So both regions have to be present. At runtime the names and links are no longer needed, but the CFA and PFA pointers are. 2022-07-06 21:02:54 Honestly, though, for those really small and cheap micros going all the way to code feels like the only sensible way. 2022-07-06 21:13:02 would you consider using mecrisp? 2022-07-06 21:14:32 Probably not - part of this whole activity for me is rolling my own stuff. 2022-07-06 21:15:23 But I'm happy to garner ideas. 2022-07-06 21:20:03 hehe making the forth seems to be the main thing for everyone 2022-07-06 21:20:11 vs making something with a forth 2022-07-06 21:22:26 yeah. 2022-07-06 21:23:56 Indeed. It's just been a very entertaining thing for me over time. 2022-07-06 21:24:37 They say one needs hobbies. :-) 2022-07-06 21:53:23 hey guys 2022-07-06 22:33:01 Hi tabemann. 2022-07-06 22:40:21 sigh. I'm gonna do a demo on Saturday, but I don't know what I'm gonna demo except that it'll be zeptoforth and zeptocom.js 2022-07-06 22:40:31 on the STM32F411 "black pill" 2022-07-06 22:54:03 So, when I first started my career, programmable logic devices were just beginning to appear. Things like "16R8" and so on, and the "flagship" device was the 22V10. As the years went by of course these things got fancier and fancier - AMD had a line called "MACH" that was a nice step forward. 2022-07-06 22:54:17 Of course this all eventually evolved in to FPGAs, and those have just gotten bigger and bigger. 2022-07-06 22:54:35 Do small and medium programmable logic devices even exist anymore? 2022-07-06 22:54:47 Or has it become sort of FPGA only?