IRC Log - 2025-03-12 - ##forth

Channel: ##forth
Total messages: 63
Time range: 09:22:43 - 23:51:55
Most active: xentrac (25), veltas (19), MrMobius (10)
09:22:43 ##forth <veltas> Also Forth has a certain flexibility about it, this redefining stuff is what allows us to define any DSL we want for a problem
09:23:05 ##forth <veltas> This is certainly the early attitude about what Forth should be, anyway, whether or not it is that....
09:23:57 ##forth <veltas> And safety because you don't want to accidentally call the word currently being defined from within a [ ] block
10:40:42 ##forth <veltas> DKordic: MSP430 isn't really like a PDP-11
10:41:30 ##forth <veltas> It has some PDP-11 features, and not others. It's like maybe 20 times faster than a PDP-11, it has less floating point options that PDP-11 had but that's probably irrelevant with how much faster it is, more registers, etc.
10:42:10 ##forth <veltas> It is interesting that an inexpensive MCU will hands-down wipe the floor with a $40k minicomputer from 1970, for some tasks
10:42:39 ##forth <veltas> But I'm more interested in actually emulating a PDP-11 rather than getting the compute capabilities, which I can get anywhere cheap
11:46:23 ##forth <veltas> GeDaMo: Right now I'm thinking we might have some fun if I host a PDP-11 emulator and let people "dial in" to it via some simple protocol, and we an share a morth multi-tasking system
11:46:49 ##forth <veltas> With a proper classic forth on it, something that can run most of starting forth unchanged, but maybe also support the ANS words
11:47:38 ##forth <veltas> That's my idea, so not too interested in a physical reproduction, but I would like to emulate an authentic PDP-11 in some limited fasion
11:47:49 ##forth <veltas> I won't actually use a phoneline though.... unless you're on dial-up
11:49:27 ##forth <GeDaMo> pidp uses simh
12:00:37 ##forth <veltas> I'm considering simh, I've not settled on an emulator yet
12:00:55 ##forth <veltas> Mostly just been reading the manuals, which is a good read
12:01:13 ##forth <veltas> I don't know much about PDP-11 but can definitely feel the impact DEC had on UNIX and C
15:11:57 ##forth <xentrac> veltas: have you written 68000 code? the PDP-11 is a lot like a 68000
15:19:54 ##forth <veltas> Nope, predates me
15:27:09 ##forth <xentrac> oh, it still had significant usage up till the early 21st century in things like the Treo Handspring
15:27:29 ##forth <GeDaMo> Aren't there some DSPs derived from the 68K?
15:27:45 ##forth <xentrac> (although in ColdFire they removed some of the addressing modes that were harder to implement efficiently)
15:27:48 ##forth <xentrac> it's a very pretty instruction set
15:27:56 ##forth <xentrac> GeDaMo: it wouldn't be surprising, but I can't think of one
15:29:00 ##forth <GeDaMo> It might be Coldfire I'm thinking of
15:31:17 ##forth <xentrac> I can't think of any ColdFire-based DSPs, but it wouldn't be surprising if they existed
15:32:28 ##forth <GeDaMo> Eh, my memory is unreliable :P
15:35:11 ##forth <veltas> xentrac: The company I work for did 68K CPU boards back in the day, but the only stale arch I've worked with directly was PPC
15:35:49 ##forth <veltas> The main bus we support is essentially based on the 68K system bus, but many iterations later and doing a load of stuff that wasn't possible back then anyway
15:36:04 ##forth <xentrac> yeah
15:36:51 ##forth <xentrac> one of the first programming projects I ever worked on for somebody else was a 68K tutorial for an introduction-to-assembly-language class
15:37:18 ##forth <xentrac> that probably would have been in 01994
15:38:35 ##forth <xentrac> at that point I think Macintosh was still 68K, and the PalmPilot hadn't launched, though Newton was already ARM
15:39:13 ##forth <xentrac> yeah, that was the year the PowerMac was introduced
15:40:56 ##forth <xentrac> seems like the last 68k-based Macintosh was the Performa, discontinued in 01997: https://en.wikipedia.org/wiki/Macintosh_Performa
15:41:56 ##forth <xentrac> oh, no, they kept calling them "Performa" even after they switched over to PPC
15:43:31 ##forth <xentrac> so maybe the Quadra LC 475 discontinued mid-01996: https://en.wikipedia.org/wiki/Template:Timeline_of_Macintosh_Centris,_LC,_Performa,_and_Quadra_models
15:44:01 ##forth <xentrac> I was still using a 68k-based Sun 3/60 in my bedroom until 02000, though. As an X terminjal
15:49:24 ##forth <veltas> Yup we do and did do VMEbus boards, but haven't done 68k since a long time ago
16:25:41 ##forth <veltas> It does remind me of the Unibus
17:18:25 ##forth <xentrac> I don't know much at all about the bus design
18:03:21 ##forth <MrMobius> TI still sells new calculators with a 68k in them
18:04:04 ##forth <MrMobius> if you have a hankering for programming a system with one
18:08:23 ##forth <xentrac> oh, that's a good point, MrMobius
18:08:36 ##forth <xentrac> they're pretty outrageously overpriced though
18:08:50 ##forth <MrMobius> agreed
18:09:09 ##forth <MrMobius> I got a used one from craigslist for $20
18:09:25 ##forth <MrMobius> which was the prior model that didnt have the shit chevron keyboard the new ones have
18:09:42 ##forth <xentrac> what's a shit-chevron? Is that like a Dirty Sanchez?
18:09:51 ##forth <MrMobius> it is if you want it to be :P
18:10:53 ##forth <MrMobius> I just mean I like having a straight row of keys instead of some marketing goon curving the row of keys trying to sell more calculators
18:10:53 ##forth <xentrac> there's something to be said for coming with a keyboard and a screen already connected
18:11:17 ##forth <xentrac> but I suspect qemu-m68k is a better option for most people
18:11:26 ##forth <MrMobius> ya! that's why I only collect calculators. unlike retro computers, the whole collection fits in a briefcase. dont have to worry about finding a CRT either or replacing leaky caps on ancient PSUs
18:11:55 ##forth <MrMobius> xentrac: probably so
18:12:43 ##forth <MrMobius> I was working on setting up 10 or assembler and gdb combos in a VM to give assembly noobs who mostly ask about how to set things up and usually just wont use a debugger no matter how hard you try
19:09:06 ##forth <xentrac> a bit harder with something like : narray create 0 do , loop ; for example
20:24:29 ##forth <xentrac> yakubin: btw in gforth , is defined as : , here cell allot ! ; which I think is a definition that will work in all standard Forths
20:24:40 ##forth <xentrac> so it's true that it's based on the word here
20:25:34 ##forth <xentrac> usually you don't need to make your defining words immediate because you're already in interpret state when you run them, so immediate has no effect
23:35:03 ##forth <pgimeno> veltas: "And safety because you don't want to accidentally call the word currently being defined from within a [ ] block" Wow, didn't occur to me. Just tried this in the jupiter and id did print hi! : x ." hi!" exit [ x
23:47:46 ##forth <pgimeno> without the exit it just reset the machine
23:50:52 ##forth <pgimeno> as for yakubin's request, I was wondering about the len part in ( -- addr len ), I can't think of anything but: create array here val1 , val2 , ... valN , here : array array [ swap - ] literal ; but it's not portable
23:51:55 ##forth <pgimeno> and it's a good example of the need for redefining past words instead of performing a recursive call, by the way :)