00:00:00 --- log: started forth/20.01.05
00:09:47 --- join: gravicappa joined #forth
00:32:24 --- join: ryke joined #forth
02:34:18 --- join: dddddd joined #forth
03:01:14 --- quit: dave9 (Quit: dave's not here)
03:13:34 --- join: rdrop-exit joined #forth
04:09:30 --- quit: rdrop-exit (Ping timeout: 265 seconds)
04:09:45 --- join: rdrop-exit joined #forth
04:39:56 --- quit: gravicappa (Ping timeout: 268 seconds)
05:03:10 --- quit: iyzsong-x (Ping timeout: 260 seconds)
05:16:41 --- quit: rdrop-exit (Ping timeout: 265 seconds)
05:20:20 --- join: rdrop-exit joined #forth
05:28:31 --- quit: rdrop-exit (Ping timeout: 260 seconds)
05:33:55 --- join: rdrop-exit joined #forth
06:00:00 --- quit: jsoft (Ping timeout: 258 seconds)
06:05:39 --- quit: ryke (Ping timeout: 268 seconds)
06:25:29 --- join: gravicappa joined #forth
07:09:30 --- quit: rdrop-exit (Quit: Lost terminal)
07:55:48 --- quit: WickedShell (Remote host closed the connection)
09:25:41 --- quit: inode (Ping timeout: 246 seconds)
10:30:37 --- join: ryke joined #forth
10:54:36 --- quit: gravicappa (Ping timeout: 260 seconds)
11:11:39 --- join: jsoft joined #forth
12:16:53 --- quit: jsoft (Ping timeout: 268 seconds)
13:00:34 <tpbsd> G'day Forthlings
13:15:11 --- quit: ryke (Ping timeout: 260 seconds)
13:39:18 --- join: dave9 joined #forth
13:43:29 --- join: jedb_ joined #forth
13:46:18 --- quit: jedb__ (Ping timeout: 268 seconds)
14:14:56 --- join: learning joined #forth
15:05:13 --- join: ryke joined #forth
15:35:57 --- quit: learning (Remote host closed the connection)
15:43:26 --- join: learning joined #forth
16:03:48 --- join: learning_ joined #forth
16:04:21 --- quit: learning (Ping timeout: 265 seconds)
16:05:20 --- join: jsoft joined #forth
16:24:12 --- join: dave0 joined #forth
16:24:30 --- quit: tabemann (Ping timeout: 260 seconds)
16:37:37 --- quit: learning_ (Remote host closed the connection)
16:46:24 --- quit: ryke (Ping timeout: 265 seconds)
16:51:49 --- join: tabemann joined #forth
17:47:16 --- quit: dave0 (Quit: dave's not here)
18:14:19 --- join: smokeink joined #forth
18:16:20 --- join: jedb__ joined #forth
18:19:29 --- quit: jedb_ (Ping timeout: 268 seconds)
18:26:00 <crab1> tpbsd: how are ya
18:26:24 <tpbsd> I'm good crab1, you ?
18:27:24 <crab1> good, might be getting a free oscilloscope :)
18:27:43 <tpbsd> I've just added another page to my doc site: https://mecrisp-stellaris-folkdoc.sourceforge.io/ihex-32bit.html   the marvels of Intel Hex extended addressing
18:27:51 <tpbsd> awesome!
18:28:36 <tpbsd> there are a lot of very impressive DSO's about these days, some are really cheap compared to what they used to cost
18:29:41 <tpbsd> my HP is ancient but I prefer it to cheaper faster Chinese scopes
18:31:57 <crab1> The one I'd be getting would be an older one, my father said he had it when he used to work at Motorolla 10 years ago and it was old then, so it's only older now
18:32:17 <crab1> I believe it's two channels, but I don't know the brand or anything
18:32:34 <tpbsd> he would know exactly what is a good scope
18:33:27 <crab1> I would have asked more, but I think he had to go eat dinner
18:33:30 <tpbsd> mine was made in 1993, so it's 27 years old
18:34:24 <tpbsd> I bought a new 300Mhz, 2GS/s Rigol about 2 years ago for $1200 AUD, but I hated it, and sold it after a year
18:34:57 <crab1> what was wrong with it?
18:35:11 <tpbsd> I couldnt understand the menues, I couldnt drive it easily, I found it so hard to use
18:35:47 <tpbsd> the old HP Digital Storage Oscilliscope is so easy to use I dont have to think about it
18:36:40 <tpbsd> Ive written a GUI program, I write tons of Forth programs, Perl, C, Shell, I know them all (to some degree) but that Rigol and I just didnt get along
18:37:15 <tpbsd> oh, plus the bloofy thing took forever to boot, it was painful
18:37:28 <tpbsd> my OLD hp boots almost instantly
18:38:36 <tpbsd> the Rigol had tons better resolution LCD, color, mega faster, but the menues were a POS
18:39:17 <tpbsd> HP used to have a whole department just for designing the UI, maybe Rigol will one day
18:40:41 <crab1> Well, one person is probably enough to design the UI of an oscilloscope, if you got the right person
18:40:48 <crab1> was it really that bad?
18:41:03 <tpbsd> it was horrible to me
18:41:16 <crab1> I would think that the people designing oscilloscopes are people that use oscilloscopes, and would know better than to make shitty menus
18:41:27 <tpbsd> I think that UI design is probably one of the hardest jobs ever
18:42:06 <tpbsd> the menu wasnt 'shitty' I just didnt think like the designer did
18:42:38 <tpbsd> I spent 12 months trying to like it but was using the HP most of the time because ... easy
18:43:28 <crab1> Hmm maybe UI is harder to design than I think
18:43:56 <tpbsd> I know my UI design skills are abysmal
18:44:30 <tpbsd> my first and only GUI app really looks horrible, but it's functional
18:44:55 <tpbsd> I used it over the last 15 years every day until the start of last year
18:45:23 <tpbsd> was just GTK widgets and Perl doing the work
18:47:22 <crab1> well I will definitely say that 95% of UIs I run into are stupid as fuck, I end up needing to work things a way where I can manage UI myself
18:47:44 <tpbsd> it's definitely a skill!
18:48:18 <crab1> I enjoy the ACME editor for consistent UI across programs
18:48:52 <crab1> but ACME only lives up to its name if you run it inside plan9, unfortunately
18:50:23 <crab1> And then ACME can't be your UI for an oscilloscope :(
18:51:03 <tpbsd> I heard plan9 was good, Ive never seen it, but I'm not that much into computing, Im a electronics tech
18:51:45 --- join: X-Scale` joined #forth
18:52:58 <crab1> I enjoy it, but it doesn't like my wifi card and I can't use ethernet at the moment, so OpenBSD it is
18:53:24 --- quit: X-Scale (Ping timeout: 268 seconds)
18:53:24 --- nick: X-Scale` -> X-Scale
18:53:31 <tpbsd> yeah old software and modern hardware rarely get along
18:53:55 <tpbsd> what window manager do you use ?
18:54:03 <crab1> 9front is actively used and developed, but the group is small and they all have ThinkPads
18:54:33 <crab1> I use cwm, it ships with OpenBSD, I find it does a good job of staying out of the way
18:54:37 <tpbsd> I hear the FreeBSD devs all use mac laptops
18:54:46 <crab1> and I love the little application menu cwm has
18:54:49 <tpbsd> I have used icewm the last 20 years
18:54:55 <tpbsd> maybe longer
18:55:52 <crab1> it has a bar by default, doesn't it?
18:56:19 <tpbsd> yeah, I'm a virtual desktop addict
18:56:32 <tpbsd> I use about 47 of them
18:56:42 <crab1> I do not like the bar
18:56:46 <tpbsd> but I hate icons so dont use them
18:57:13 <crab1> I want the screen empty besides whatever programs I'm running, I find myself running out of space as it is
18:57:14 <tpbsd> cwm looks pretty good to me, very functional
18:57:26 <tpbsd> yeah, it's always a issue
18:57:50 <tpbsd> i do consider a second monitor from time to time
18:58:06 <tpbsd> I may add one soon
18:58:18 <tpbsd> just for stuff I cant fit on the one screen
18:58:38 <crab1> I was going with just my laptop's monitor for a while, which is 17" so pretty good size laptop, but I have my monitor back now and I really did miss the second screen
18:58:58 <tpbsd> ahh
18:59:31 <crab1> Often it's "Do things on this screen, read things on second screen," but if I have acme on my system I will do one screen for acme and the other for everything else
18:59:39 <tpbsd> I tried 2 screens a few years ago and found it was too wide, but Ive changed since then I think
18:59:57 <tpbsd> yeah, thats a bit like what I had in mund
19:00:00 <tpbsd> mind
19:01:05 <crab1> I'd recommend giving it another try
19:01:19 <tpbsd> I will
19:01:48 <tpbsd> I'm upgrading the OS on this box in a few weeks and will set up a second screen then
19:01:57 <tpbsd> do it all together
19:02:05 <crab1> to what OS?
19:02:12 <tpbsd> same os, FreeBSD
19:02:23 <tpbsd> just a later version
19:02:57 <tpbsd> I can't leave FreeBSD now, Im a ZFS and jail addict
19:03:19 <tpbsd> if I did it would be to openindiana
19:03:22 <crab1> ah, I have encountered your kind before
19:03:30 <tpbsd> addicts ?
19:04:00 <crab1> specifically the ZFS and jail sorts
19:04:23 <tpbsd> yeah, hard core addicts on the really strong drugs!
19:05:10 <tpbsd> FreeBSD starts you on the soft drugs like jails, then slowly gets you hooked on the strong stuff like ZFS
19:06:49 --- quit: tabemann (Ping timeout: 240 seconds)
19:07:32 --- quit: jedb__ (Ping timeout: 265 seconds)
19:07:45 <crab1> I feel like an immense amount of effort goes into making the terminology used by mathematicians sound as intelligent as possible
19:07:51 --- join: jedb joined #forth
19:08:14 <tpbsd> probably not just mathematicians
19:08:28 <tpbsd> look at lawyers and Latin ?
19:08:41 <crab1> That's true too
19:09:09 <tpbsd> i can understand the reason for latin tho, everyone should probably learn it
19:09:15 <tpbsd> doctors also lean it
19:09:36 <tpbsd> plus the language kind of defines the technology ?
19:09:47 <tpbsd> how many romans were mathematicians ?
19:10:06 <tpbsd> with no concept of zero, probably not many
19:11:17 <crab1> concept of zero is actually a quite unintuitive jump, though it may seem simple
19:11:30 <crab1> lots of maths can be done without a concept of zero
19:11:41 <tpbsd> true
19:11:42 <crab1> perhaps even most maths
19:11:56 <tpbsd> the romans managed without it
19:13:07 --- quit: X-Scale (Ping timeout: 258 seconds)
19:13:27 <crab1> this is true
19:13:57 --- join: X-Scale` joined #forth
19:15:02 --- nick: X-Scale` -> X-Scale
19:15:18 <tpbsd> but then they didnt build spacecraft or 2nm transistors
19:15:31 <tpbsd> all tech gets us only so far
19:17:51 <crab1> tech is our salvation!
19:18:19 <tpbsd> absolutely
19:19:12 <tpbsd> sadly tho, and Im not complaining, I feel as tho I have lived at 0.0000000000000000000000000000000000000001% of the tech era I really want
19:20:07 <tpbsd> I want to live in a era when the unified field theory is complete and understood by children
19:20:17 --- join: rdrop-exit joined #forth
19:20:48 <tpbsd> when a 100kw fusion generator is the size of a USB stick, perfectly safe and lasts for 100 years
19:21:25 <crab1> but hey, if it's already been done, you get less of a role in the achievement
19:21:55 <tpbsd> i was never after fame, only understanding
19:22:42 <tpbsd> i dont care now, nor have I ever cared, if zero people in the world know of me
19:23:09 <crab1> not the fame, but the discovery
19:23:14 <crab1> discovery is great fun
19:23:17 <tpbsd> being alive with a working brain is all I care about
19:23:25 <rdrop-exit> Good morning Forthwrights
19:23:29 <crab1> maybe you prefer understanding to discovery, but I like discovery
19:24:01 <crab1> ello rdrop
19:24:11 <tpbsd> there will *always* be tons of understanding to be acquired by every individual. The fact you may understand calculus doesnt mean I do
19:24:15 <rdrop-exit> hi crab1
19:24:20 <tpbsd> gday rdrop-exit !
19:24:58 <rdrop-exit> Likewise Master Forth Technician (tm) and Youtube celebrity
19:25:18 <tpbsd> crab1, this is my opinion on 'reinventing the wheel: https://mecrisp-stellaris-folkdoc.sourceforge.io/rants.html?highlight=rant#reinventing-the-wheel
19:25:30 <tpbsd> thank you, Zen Forth Guru!
19:25:59 <tpbsd> rdrop-exit, yes, I'm famous now, I have 43 subscribers on my Forth channel!
19:26:11 <rdrop-exit> we are not worthy
19:26:23 <tpbsd> one is a alien as there are only 42 Forth users in the world
19:26:33 <rdrop-exit> I'm one of those 43
19:26:55 <rdrop-exit> Even though I don't speak Australian
19:26:56 <tpbsd> really ? remove yourself immediately, my channel is not worthy!
19:27:12 <rdrop-exit> :)
19:27:42 <tpbsd> I have a new page and it's about your *favorite* thing !!! Intel Hex ;-)
19:27:57 <rdrop-exit> :))
19:28:11 <rdrop-exit> link please
19:28:24 <tpbsd> I have mastered Intel Hex Extended Addressing type 04!
19:28:33 <rdrop-exit> ooh!!!
19:28:45 <tpbsd> https://mecrisp-stellaris-folkdoc.sourceforge.io/ihex-32bit.html#ihex32
19:29:22 <tpbsd> yes, I can clone a onforth image of up to 4GB now, should 4GB flash ever become common on small MCU's 
19:29:38 <tpbsd> onforth = Forth on mcu
19:29:50 <rdrop-exit> Reinventing something is a great way to gain a deep understanding, that's how Feynman worked
19:30:04 <tpbsd> so I guess I wont need to modify it for at least another couple of years ?
19:30:11 <tpbsd> oh I agree
19:30:40 <tpbsd> I see reinventing the wheel as a absolute right of passage for everyone who designs
19:31:22 <rdrop-exit> All that ihex stuff is unecessary ;)
19:32:24 <tpbsd> so you have said, many times
19:32:45 <tpbsd> however it's necessary here
19:32:52 <rdrop-exit> My tether bias is showing again
19:33:11 <tpbsd> well there are alternatives, but I love doing as much as I can *on chip*
19:33:17 <crab1> I like that you have an example that literally uses a wheel
19:33:37 <tpbsd> it is, you are certainly tethered to you anti ihex bias :)
19:34:00 <tpbsd> crab1, at my age one has plenty of real life experience to draw upon
19:34:34 <tpbsd> and I cant really convey the sense of utter loss that mums plastic wheel engendered in me
19:34:41 <tpbsd> wheels
19:35:15 <tpbsd> we kids had the worlds strongest, most bulletproof motorised go-cart with the metal wheels
19:36:49 <tpbsd> the feeling I had when driving it was exactly the same as if *right now* a man from Aglient knocked on your door and delivered a 10Ghz, 100GS/s, 10 channel digital scope to you for free
19:37:28 <tpbsd> and it still had the $1 million USD price tag on the box
19:39:47 <crab1> I wouldn't have a clue what to do with it lol
19:40:08 <rdrop-exit> resell it for $0.5M
19:40:30 <crab1> true
19:40:50 <tpbsd> that wouldn't change the emotion of having it there, of imagining the possibilities ?
19:41:00 <tpbsd> hahah
19:41:21 <tpbsd> and buy a $1000 Rigol with brain killing menues ?
19:41:48 <crab1> don't worry, you wouldn't like the menus on the million dollar one either
19:42:20 <tpbsd> Aligent is what HP used to be
19:42:44 <tpbsd> Im pretty sure the menues would be *exactly* what Id like
19:43:09 <tpbsd> even if they were the same ones I have now in my 1994 HP Digital Storage Oscilliscope
19:44:20 <rdrop-exit> One nice thing about the RISC-V compressed instructions is that they are an extension of the ISA rather than an alternative ISA (unlike Thumb and MIPS16)
19:44:42 <tpbsd> OH ?
19:44:48 <rdrop-exit> (getting ready to write assembler words for the compressed ISA)
19:44:51 <tpbsd> well thats really nice to hear
19:45:28 <tpbsd> I know matthias said that although his latest Mecrisp-Quintus for RISC-V uses compressed isa, that one can use either
19:45:35 <tpbsd> at the same time
19:45:59 <tpbsd> RISC-V is sounding better all the time
19:46:57 <rdrop-exit> It's funky but nice
19:47:07 <tpbsd> I havent yet set up hardware handshaking and 108Mhz clock on the GD32VF103 but I think it will be pretty nice
19:48:04 <tpbsd> it's not bad at 16mhz
19:48:09 <crab1> RISC-V is delicious
19:48:25 <tpbsd> crab1, it's still expensive
19:48:46 <tpbsd> the GD32VF103 is a real PITA to program the GPIOs
19:48:51 <crab1> But should get cheaoer, I think
19:48:56 <crab1> cheaper*
19:49:19 <tpbsd> I think that RISC-V could well alter the face of embedded in time
19:49:44 <tpbsd> in the same way that Linux pretty much chopped Windows off at the knees
19:49:44 <crab1> not only embedded
19:50:14 <tpbsd> Im only do embedded, so I'll leave prophecies in otehr areas to the experts like you :)
19:50:21 <crab1> "expert"
19:50:26 <tpbsd> well more than me
19:50:34 <tpbsd> I lead a sheltered embedded life
19:50:49 <tpbsd> I like $0.50 SMT cortex-m0's
19:51:01 <crab1> Now that I mentioned it earlier I want plan9 again :(
19:51:13 <tpbsd> plan9 lust ...
19:51:29 <crab1> have to ask my landlord if I can get ethernet down here
19:51:57 <tpbsd> wifi no good ?
19:52:33 <crab1> plan9 doesn't like my wifi card, and the wifi signal is often poor
19:53:34 <tpbsd> there are other options (to quote rdrop-exit )
19:53:59 <rdrop-exit> "RV32I instructions are indistinguishable in RV32IC. 
19:53:59 <rdrop-exit> Thumb-2 is actually a separate ISA with 16-bit instructions plus 
19:53:59 <rdrop-exit> most but not all of ARMv7."
19:54:01 <crab1> simple options?
19:54:02 <tpbsd> never use a wifi card anyway
19:54:32 <tpbsd> crab1, use 2 wifi units connected via ethernet at either end
19:54:58 <crab1> as in extend the wifi signal
19:55:21 <tpbsd> crab1, as in extend the network, pretend the wifi part is just a long lan cable
19:55:52 <tpbsd> crab1, using two wifi AP's in bridge mode and you have a 'intelligent' lan cable alternative
19:56:05 <crab1> I think it would be better to just get an ethernet cable run down here
19:56:19 <tpbsd> rdrop-exit, thats amazing, you have made me hungry to get back to the RISC-V 
19:56:41 <tpbsd> crab1, absolutely, only use wifi where cable isnt possible
19:57:56 <rdrop-exit> I really like that, the 16-bit instructions are just extra instructions you can choose to use, rather than a completely alternate ISA
19:58:12 <tpbsd> oh me too!
19:58:39 <rdrop-exit> You just need to make sure the chip supports RV32IC, not just RV32I
19:58:43 <tpbsd> and the latest mecrisp using compressed is only 18kB, the 32 bit version was 28.8kB
19:59:03 <crab1> you can even mix and match the two, right?
19:59:05 <tpbsd> the GD32VF103 does as I have both versions for it
19:59:16 <tpbsd> crab1, seems so
19:59:55 <tpbsd> crab1, I have no RISC-V assembly experience yet
20:00:33 <rdrop-exit> crab1, you always have RV32I, the C gives you compressed 16-bit instructions in addition to the normal 32-bit ones, and relaxes the alignment requirements
20:00:50 <crab1> now,
20:01:17 <crab1> how to run 40ft of ethernet through my landlord's kitchen and down the steps without him noticing
20:01:36 <tpbsd> crab1, hence my wifi suggestion
20:02:01 <crab1> but then I need extra hardware
20:03:04 <rdrop-exit> think out of the box, do you really need a landlord?
20:03:27 <rdrop-exit> ;)
20:03:30 <tpbsd> :)
20:04:53 <crab1> That's good thinking, I'll move out to the woods, and communicate with the world via HAM radio
20:04:57 * tpbsd has a landlord
20:05:47 <rdrop-exit> IP over carrier pigeons
20:06:14 <rdrop-exit> https://en.wikipedia.org/wiki/IP_over_Avian_Carriers
20:06:40 <rdrop-exit> RFC 1149
20:08:12 <rdrop-exit> I think the simplest, most Forthy solution to your problem is to bump off the landlord
20:08:43 --- join: tabemann joined #forth
20:08:52 <MrMobius> tpbsd, you should let him go soon then
20:08:58 <tpbsd> I think they have Internet in prison now ?
20:09:21 <tpbsd> MrMobius, I will when I retire next october
20:09:51 <tpbsd> MrMobius, my ex wife has a very nice 8.5 acres all paid for :)
20:10:17 <MrMobius> is she your landlord?
20:10:30 <tpbsd> no
20:10:44 <tpbsd> pretty unlikely
20:11:02 <tpbsd> as I was the one who paid for her land and home
20:11:17 <rdrop-exit> ouch
20:11:21 <MrMobius> at least you know what youd be getting if you rented from her
20:11:42 <tpbsd> which is why I dont :)
20:11:42 <crab1> this carrier pigeon idea seems good, but for IM it may not be practical
20:12:32 <rdrop-exit> oh well
20:21:51 --- quit: crab1 (Quit: WeeChat 2.6)
20:22:23 <rdrop-exit> Lots of bit scrambling in RISC-V's compressed instruction formats, e.g. offset[11|4|9:8|10|6|7|3:1|5]
20:22:57 <rdrop-exit> funky stuff
20:23:11 <tpbsd> I'm very confused by that
20:23:52 <tpbsd> please explain it in 10 words or less ?
20:24:17 <rdrop-exit> Let's say you have a 12 bit offset, offset[19:0]
20:24:30 <rdrop-exit> offset[11|4|9:8|10|6|7|3:1|5] shows how it would be encoded in the instruction
20:24:51 <rdrop-exit> i.e., the bits are redistributed all over the place
20:25:35 <tpbsd> I guess I have a long learning curve with RISC-V assy ahead of me then!
20:25:47 <rdrop-exit> "The immediate fields are scrambled in the instruction formats instead of in sequential order so that as many bits as possible are in the same position in every instruction, thereby simplify- ing implementations."
20:26:08 <tpbsd> sounds fair
20:26:37 <rdrop-exit> More trouble for the assembler writer, less trouble for the H/W implementer
20:27:38 <rdrop-exit> oops the above should be, let's say you have a 12 bit offset, offset[11:0] ...
20:28:47 <rdrop-exit> It won't affect application programmers, just assembler/disassembler writers
20:29:30 <tabemann> hey
20:30:01 <rdrop-exit> makes it harder to eyeball an instruction in hex and mentally disassemble it though
20:30:32 <rdrop-exit> end up cross-eyed
20:31:24 <rdrop-exit> hi tabemann!
20:31:44 <tabemann> ugh that'd be a fuckin pain to implement my compiler for, were I to write a compiler for RISC-V
20:32:12 <rdrop-exit> it's pretty easy to deal with
20:33:27 <rdrop-exit> just write some words to pack the format
20:33:29 <tpbsd> rdrop-exit, makes sense, the hardware is a million times harder to design
20:33:54 <tabemann> the thing is that this kind of thing smells like premature optimization to me
20:34:04 <tpbsd> rdrop-exit, as you say, no one writes in machine code except Forth designers :)
20:34:25 <tpbsd> tabemann, less complaining, more doing!
20:34:41 <rdrop-exit> tabemann, why? they're just making the hardware simple
20:34:50 <tabemann> Thumb-2 is annoying enough
20:34:58 <tabemann> and it's not nearly as bad as that
20:35:01 <tpbsd> it sure is
20:35:34 <tpbsd> rdrop-exit, what have you started ?
20:35:51 <tabemann> rdrop-exit: it's like archs that have a branch delay slot; sure, in the hardware of that age it sped things up, but it was not helpful in the future
20:36:03 <tpbsd> tabemann, matthias seems to have managed it somehow ?
20:36:07 <rdrop-exit> But as I mentioned before you got here, they were really smart about making the compressed instructions just an extension, unlike thumb
20:36:58 <tabemann> tpbsd: Thumb-2 isn't really that bad, it's just slightly annoying when it has discontinuous immediate fields
20:37:02 <rdrop-exit> tabemann, that's different, the branch delay slot in MIPS was something the application programmer had to deal with
20:37:17 --- join: ryke joined #forth
20:37:41 <rdrop-exit> Untested draft code ahead:
20:37:45 <rdrop-exit> source  RISC-V - RV32I - Instruction Formats - J-type
20:37:46 <rdrop-exit>    31 bit   constant %j.imm[20]
20:37:46 <rdrop-exit> 30 21 bits  constant %j.imm[10:1]
20:37:46 <rdrop-exit>    20 bit   constant %j.imm[11]
20:37:46 <rdrop-exit> 19 12 bits  constant %j.imm[19:12]
20:37:48 <rdrop-exit> : +j.imm[20:0] ( sx[20:0] 32b -- 32b' )
20:37:50 <rdrop-exit>   over 31 21 bits and triggers ximm
20:37:53 <rdrop-exit>   over     0 bit  and triggers xoddimm
20:37:55 <tpbsd> tabemann, I have programmed thumb-1 a small degree, I'm ok with it
20:37:55 <rdrop-exit>   over    20 bit  and %j.imm[20]    pack
20:37:58 <rdrop-exit>   over 19 12 bits and %j.imm[19:12] pack
20:38:00 <rdrop-exit>   over    11 bit  and %j.imm[11]    pack
20:38:03 <rdrop-exit>   swap 10  1 bits and %j.imm[10:1]  pack ;
20:38:11 <tpbsd> SPAMMER!
20:38:30 <rdrop-exit> where did my blank lines go? IRC eliminated them
20:38:43 <tabemann> tpsd: Thumb-2 is better than Thumb-1 though in that it greatly simplifies long literals
20:38:50 <tpbsd> rdrop-exit, why paste Japanese text here ?
20:39:23 <tabemann> because in Thumb-2 a long literal is just two 32-bit instructionsthat are very simple
20:39:34 <rdrop-exit> domo arigato Mr. Roboto
20:39:36 <tpbsd> tabemann, definitely not my favorite ISA but I can use it ok
20:39:57 <tpbsd> hi wakalimasta!
20:40:13 <tabemann> whereas in Thumb-1 if one wants a 32-bit literal you need a long series of MOVS, LSLS, and ORRS or ADDS instructions
20:40:34 --- join: crab1 joined #forth
20:40:42 <crab1> BAck
20:40:47 <rdrop-exit> forth
20:40:48 <tpbsd> tabemann, only if you cant write a macro or dont have a Forth that deals with it
20:41:03 <tpbsd> crab1, welcome back! 
20:41:18 <rdrop-exit> welcome forth, crab1!
20:41:24 <tabemann> hey crab1
20:41:44 <tpbsd> tabemann, I have 'registerliteral,' that takes care of that stuff
20:41:55 <tabemann> tpbsd: the difference is that one is guaranteed to use at most 8 bytes to represent a 32-bit number, whereas the other may use more bytes than that
20:43:10 <tabemann> I imagine that the former also executes faster
20:43:13 <tpbsd> tabemann, yeah but things usually average out I guess ?
20:43:18 <rdrop-exit> tabemann, I bet you'll find RISC-V cleaner than Thumb
20:43:30 <crab1> Who here has made a forth
20:43:37 * tabemann raises hand
20:43:41 <tpbsd> but it seems that RISC-V offers us the best of both worlds now
20:43:41 <rdrop-exit> guilty
20:43:58 <tpbsd> crab1, not me
20:44:17 <tpbsd> crab1, I'm just a Forth user
20:44:33 <crab1> Well I am trying to devise a forth, and something is troubling me
20:44:34 <rdrop-exit> tb, for now, never say never
20:46:23 <tpbsd> rdrop-exit, it's unlikely, and I have other things of higher priority
20:46:32 <tpbsd> but as you say ...
20:46:47 <tpbsd> crab1, what is troubling you ?
20:48:07 <rdrop-exit> take the couch and tell us your problem
20:49:25 <tpbsd> Dr rdrop-exit is listening ...
20:49:52 <rdrop-exit> wee have ways of making you better
20:49:54 <crab1> Well, I want my forth to support recursion, but I don't want to do TCO, and I don't want to have a word like -; as in colorForth, and I'm not sure if there's anything I can do
20:51:08 <rdrop-exit> explicitly branch back to the beginning to the word, i.e. explicit tail recurion
20:51:13 <rdrop-exit> * recursion
20:51:37 <crab1> then every word with recursion would have extra stuff
20:51:38 <rdrop-exit> although TCO is trivial in Forth, might as well have it
20:52:06 <rdrop-exit> what extra stuff?
20:52:28 <crab1> the explicit tail recursion
20:52:51 <rdrop-exit> e.g. in SRT, you're replacing a ret with a relative jump back to the beginning of the word
20:53:04 <crab1> yeah that's what you do
20:53:22 <rdrop-exit> : recur { -- }( -- ) myself jmpi, ; directive
20:53:56 <rdrop-exit> (that code is old, I need to replace it with a relative jump someday)
20:56:05 <rdrop-exit> TCO:
20:56:16 <rdrop-exit>  : curtail; ()( r: ra1 ra2 -- ra1| )
20:56:16 <rdrop-exit>    call| a@ here <> ?;
20:56:16 <rdrop-exit>    encoded jmpi  address 1+ here \- b!
20:56:16 <rdrop-exit>    bail; ; compiled
20:56:46 <rdrop-exit> (also old code, that needs updating to a relative jump)
20:57:07 <crab1> would a threading technique that's like subroutine threading but explicitly pushes and jumps instead of calling be bad? because then recursion is just "don't push before you jmp"
20:57:18 <crab1> I don't know if push jmp ret is slower than call ret
20:58:29 <crab1> or if call is superior to jmping everywhere
20:59:13 --- quit: smokeink (Ping timeout: 240 seconds)
20:59:31 <rdrop-exit> TCO is just replacing the final call with a jmp
21:00:06 <crab1> I know, I'm just asking if it would be bad to ALWAYS jmp instead of call
21:01:26 <rdrop-exit> that's basically what you do with RISCs, where a call is synthesized
21:03:17 <crab1> and then recursion is free, instead of needing a word recurse or -; or adding in TCO
21:03:46 <rdrop-exit> what do you mean recursion is free?
21:04:53 --- join: smokeink joined #forth
21:05:13 <rdrop-exit> tail recursion is just a jump
21:05:40 <crab1> I mean a word with its name in its defintion works without anything extra, instead of needing to change from calling as usual to jumping
21:06:03 <crab1> forth can do what it does as usual instead of needing to account for recursion will fill the stack
21:06:07 <tpbsd> nothing in life is free!
21:07:12 <crab1> tpbsd: BOGO sales are a lie!
21:08:26 <rdrop-exit> I'm not getting what you're saying, tail recursion is a special case that just lays down a jump instead of a call, normally in Forth it's done with RECURSE or RECUR or a similarly named word
21:09:12 <rdrop-exit> TCO is done by replacing the final call with a jump
21:09:19 <crab1> I understand all that
21:09:36 <crab1> I just don't like tail recursion being a special case, I don't feel like it should be
21:10:20 <crab1> but I'm crazy and everyone has told me the same about this specific issue of mine
21:10:29 <crab1> it continues to bother me though
21:10:31 <rdrop-exit> The reason we use a word such as RECURSE is we don't want to mask older versions of the word
21:10:54 <rdrop-exit> old version of test:
21:10:57 <crab1> I also really do not like RECURSE
21:11:08 <rdrop-exit> : TEST blah blah ;
21:11:23 <rdrop-exit> new version of test:
21:11:42 <rdrop-exit> : TEST bing TEST bang boom ;
21:12:09 <rdrop-exit> With you solution how wouuld you implement the second version of TEST?
21:12:47 <crab1> the second version is supposed to call itself or the first version?
21:13:00 <rdrop-exit> it calls the first version
21:13:16 <rdrop-exit> presumably they're in different vocabularies
21:14:02 <rdrop-exit> when you add the second vocabulary to your search order you get an "enhanced" test
21:15:29 <rdrop-exit> That's why we have RECURSE instead of just using the name directly, to avoid masking previous versions
21:15:36 --- quit: smokeink (Read error: Connection reset by peer)
21:16:51 <crab1> I don't like having multiple words with the same name though, besides even in yours if you created a TEST after that that also references TEST it would just mean the second TEST and the first would be lost at that point, right?
21:17:39 <rdrop-exit> not lost, I can remove the vocabulary from the search order, my old test is still there
21:19:40 <rdrop-exit> You manage your namespaces in Forth via the search order mechanism (i.e. vocabularies)
21:19:48 <crab1> then in this case, all you can't do is use the same name word from another vocabulary in your definitions that doesn't seem too bad to me honestly, I feel like : TEST bing TEST bang boom ; meaning "use the OLD test" is confusing anyway
21:20:36 <crab1> in ": TEST bing TEST bang boom ;" TEST looks like its calling itself to me, unless I already know there's some other TEST in existence I guess 
21:21:26 <rdrop-exit> it's just like bing bang and boom, look up the existing definition of bing and compile it, same with boom, same with TEST, etc...
21:22:29 <rdrop-exit> it's consistent, look up name in search order, compile it
21:23:09 <rdrop-exit> Normally the word that is currently under construction is not yet part of the search order
21:23:12 <crab1> and right after ": TEST" my intuition says that now a new definition should exist
21:23:34 <rdrop-exit> It doesn't exist until it is completed
21:24:28 <rdrop-exit> errors may happen while you're compiling it, you don't commit it until it's done
21:26:10 <crab1> it feels very uncomfortable to me idk why, this is the one part of forth that has struck me as 'this part is wrong'
21:26:19 <crab1> I understand your reasons, but it feels wrong
21:26:43 --- join: smokeink joined #forth
21:29:34 <rdrop-exit> weigh the advantages and disadvantages, experiment, maybe your way is more practical for your needs, I can't say
21:31:15 <rdrop-exit> See if you can back up your gut feeling, with an explicit list of pros and cons
21:33:32 <rdrop-exit> There are infinite varions on the Forth theme
21:33:59 <rdrop-exit> * variations
21:34:10 <rdrop-exit> Lunch is ready, catch you all later
21:34:18 --- quit: rdrop-exit (Quit: Lost terminal)
21:34:25 <crab1> buh bye
22:11:58 --- quit: dddddd (Ping timeout: 268 seconds)
22:31:03 --- join: rdrop-exit joined #forth
22:31:07 <rdrop-exit> c[]
22:47:13 --- join: gravicappa joined #forth
22:50:46 <crab1> c[]
23:01:33 <rdrop-exit> Hi crab1
23:34:47 --- join: dys joined #forth
23:51:43 --- join: proteus-guy joined #forth
23:53:35 --- quit: patrickg (Remote host closed the connection)
23:53:44 --- join: patrickg joined #forth
23:59:59 --- log: ended forth/20.01.05