00:00:00 --- log: started forth/18.10.22 00:00:03 Essentially you build (or simulate) a computer from the logic gate level 00:00:06 Yeah 00:00:08 Should be fun. 00:00:08 Cool. Even if you use Verilog, spend at least some time using it "structurally" instead of behaviorally. 00:00:16 siraben: and i had done the last step to CPU 00:00:24 Where you use it to actually instantiate and connect up gates, flip flops, registers, etc. 00:00:54 Should I use Verilog or write a circuit simulator myself? 00:01:07 When you use it like that you're using it to basically specify a schematic in Verilog. 00:01:16 siraben: and i think we could make an alternative option, like use its verilog like lang to build a stack vm, and build a forth on it, then use that forth to implemnt a tetris game 00:01:23 Well, a circuit simulator (a good one) is a fairly large undertaking. 00:01:25 you see, save lots of work 00:01:40 I think it's a great thing to do, but you'll spend a lot of time on that as opposed to becoming familiar with logic. 00:02:00 How did N2T make the final game? 00:02:16 Ah yeah, so better use a standard tool. 00:02:37 Well, until you decide you just want to do a simulator. I think I'd leave that for later. 00:02:39 Go full circle by having the computer emulate itself 00:03:00 So are verilog programs usually free software? 00:03:02 Icarus Verilog is good, and you can couple it to GTKWave for output timing diagrams. 00:03:04 As in freedom. 00:03:08 ^ that one is. 00:03:20 the riscv team use another tool 00:03:31 And if you start to do really big stuff you can use verilator. 00:03:42 It produces a C program that does a cycle accurate simulation. 00:03:54 Also free. 00:04:59 I'd probably use iverilog at first, when I was doing "pieces," then gradually shift to verilator as the whole thing came together. 00:05:39 But a lot of the learning you want to do comes before you start simulating stuff. 00:06:00 You'll want to learn how adders and other such things work, techniques for designing good state machines, and so on. 00:06:02 Right. 00:06:10 It'd remove the mysteries on how exactly registers and implemented, how addressing works and so on. 00:06:11 --- join: ncv (~neceve@2a02:c7d:c5c9:a900:6eaf:6ef7:3b81:d5f6) joined #forth 00:06:11 --- quit: ncv (Changing host) 00:06:11 --- join: ncv (~neceve@unaffiliated/neceve) joined #forth 00:06:18 Right. 00:06:23 Actually, some video games are good simulators for circuits 00:06:27 Like Minecraft and Powder Toy 00:06:36 So that might be an option before the more heavy stuff 00:06:52 Then as you get a feel for what sort of stuff is easy in hardware and what's hard, you get better at making the hardware/software decisions. 00:07:40 Playing Minecraft how I first learned a lot about circuits, and watching the videos of others 00:07:44 It is very satisfying to see it all come together with precise timing. 00:08:00 I love the idea of games being used to educate. 00:08:09 I don't think we exploit that as thoroughly as we could. 00:08:50 I agree. Minecraft is just one example of a potential goldmine of learning opportunities 00:09:13 I'm not actively aware of any other game that is generic enough to be able to be used for non-gaming purposes, but likely there are more. 00:09:44 Ah there's 00:09:44 https://www.kickstarter.com/projects/871405126/turing-tumble-gaming-on-a-mechanical-computer 00:09:55 KipIngram: which might interest you, being a hardware-oriented person. 00:10:45 One of my kids shared a video with me a while back, it was a little gadget with rectangular panels on a spindle. 00:10:55 0 was written on one side of each spindle, and 1 on the other. 00:11:26 How do you teach your kids hardware/software stuff? Do they know Forth by any chance? 00:11:28 and one end of each panel had a little tab that swiveled, and when it was in one position it extended over the next panel. 00:11:30 It would count. 00:11:38 Ah a binary counter. 00:11:41 When you flipped a panel from 0 to 1, the tab did nothing. 00:11:49 oh minecraft 00:11:54 red powder 00:11:56 When you flipped it from 1 to 0 the tab fell over to cover the next panel. 00:12:04 So yeah, it was a "carry flat." 00:12:04 yunfan: Yeah, redstone 00:12:06 flag 00:12:15 KIpIngram: Right, I recently saw one at school too 00:12:19 And it even *looked* like a flag when it was protruding. 00:12:23 I like little things like that. 00:12:34 Some students in the design technology studio made a model out of wood 00:12:49 That may have been the very video I saw. 00:12:51 Panels that you flipped and the transition to 1 to 0 caused the adjacent one to flip as well 00:12:57 Right. 00:13:07 Very evocative of what happens in a circuit. 00:13:13 Do your kids learn Forth? 00:13:15 Or programming? 00:13:21 No, no Forth converts yet. 00:13:51 It'd be interesting to see how kids would learn it. I think the unambigious nature of Forth should help. 00:13:59 I've got five daughters One's in biomedical engineering grad school, and one just starting biomed undergraduate. 00:14:13 Both of them use/study Python. 00:14:21 A fair amount of confusion results from them knowing math and the mathematical operators not corresponding to programming syntax 00:14:27 e.g. x = x + 1 00:14:50 Right. But see, my first semester in college I bought an HP-41CV calculator. 00:14:52 RPN. 00:15:00 Programmable, four element stack. 00:15:07 (cursiously that statement in Haskell leads to an infinite loop, because the = sign is very strong in that language) 00:15:08 That's the first programming experience I ever had. 00:15:21 So I burned "stack thinking" into my synapses before I ever learned anything else. 00:15:21 RPN is fast. 00:15:30 And intuitive. 00:15:58 That particular calculator had great keys mechanically too - I could hold it in my left hand and "touch type" on it with my right. 00:16:06 I was fast as shit - very dangerous with that thing in my hand. 00:16:11 Like a gunslinger. :-) 00:16:35 Key technology changed over the years, for the worse. 00:16:49 Huh why did the thought not pass my mind to make an RPN mode for my calculator 00:16:52 It became impossible to be as fast, just because the keys didn't work with you in the same way. 00:17:05 Um, that's what you're doing. 00:17:10 Instead of hitting ENTER, keys would trigger calculations 00:17:31 Well, ENTER is just there when you have two numbers without an operation in between. 00:17:40 In Forth it's a space. 00:17:44 Right but not nearly using the full potential perhaps 00:17:59 Especially with floating point 00:18:27 Yeah it is fast for quick calculations that need fixed point and don't exceed 16 bit 00:18:41 Although I have some words for 32 bit as well 00:18:46 Right. My Forth's going to support floting point. 00:19:40 A . in the number makes it floating point, and it just gets left on the FPU internal stack. 00:20:35 The main thing standing in the way of a "really good calculator mode" is the fact that + - * / are integer operators. 00:20:39 Without the OS under me things are pretty hatd 00:20:40 I'll need f+ f- f* f/ 00:20:41 Hard 00:21:06 I contend that none of it is "hard," really. Just a "voluminous" amount of work. 00:21:53 If my primary goal was numerical work, it would probably make more sense to make + - * / floating point, and have i+ i- i* i/ for integer. 00:22:04 But I want to be a bit more true to Forth than that. 00:23:06 But of course Forth will let me make a "calculator" vocabulary and define the operators to do whatever I like. 00:23:51 : + state @ if postpone f+ ;, then + ; immediate 00:24:02 KipIngram: https://github.com/siraben/ti84-forth/blob/master/forth.asm#L1450 00:24:10 Floating point stack operators 00:24:25 FDROP FDUP F+ D* FSQUARE FSQRT 00:24:26 etc. 00:24:31 Yes. 00:25:12 It's interesting that TI has a very stack based thing going on at the OS level, but doesn't expose it for the user. 00:25:17 Oh, sorry, that "then + ; above should be "then f+ ;" 00:25:32 siraben: you mean kjava? 00:25:33 Yes, tells you that HP really had it right all along in those days. 00:25:43 yunfan: kjava? 00:26:06 I've seen papers that present all the "basic math theory" using RPN. 00:26:15 yes, i remember ten years ago, when i were using dumb cellphone 00:26:23 We say for example that the commutative law means a + b == b + a. 00:26:30 Well, that's just a b + == b a + 00:26:30 they has a kjava enviroment which could support little .jar games 00:26:32 etc. 00:26:38 and since jvm is stack based 00:26:41 Losing the OS means that I have approximately 300% the work needed, so I should probably get it back in there 00:26:41 You can build it all up using RPN from the jump, and it works just fine. 00:26:46 Otherwise the time spent programming vs time spent using isn't worth it as much 00:26:49 So we could teach our kids RPN from first grade. 00:26:56 KipIngram: Right. Proving laws. 00:27:13 a b AND == b a AND 00:27:18 Right. 00:27:20 All of it. 00:27:30 Heck, you could even formulate De Morgan's and the like. 00:27:36 So infix is just a CHOICE we've made, and in my opinion a less than optimum one. 00:27:59 Well, I can't prove it exactly but there's a one-to-one correspondence with the tree-like structure in mathematics and stack-based processing. 00:28:14 Yeah, infix is only standard because we made it so. 00:28:15 Absolutely - it's been proven. 00:28:33 Strangely, we do have prefix here and there 00:28:44 f(g(x)) or ln(x) 00:28:49 Or integral(x,dx) 00:29:14 Math notation would be absolutely unacceptable as a programming language. It's just convention. 00:29:16 Right. That last could be dx x integral 00:29:28 dx x f integral 00:29:37 e.g. x^5 (d/dx), looks like a fraction! But it's a derivative. 00:31:17 Also the fact that exponentiation isn't usually indicated with an operator in math is interesting. 00:33:47 --- join: dave69 (~dave@90.20.215.218.dyn.iprimus.net.au) joined #forth 00:34:10 Well, sometimes it's indicated with ^ and in some languages as **. 00:34:29 Digital guys think of ^ as the exclusive or operator, though. 00:34:38 Yeah 00:34:57 So in Forth you might say " y^x" 00:35:21 I think it should be named EXPT 00:35:30 I'll probably go with ^ for exclusive or and ** for "power." 00:35:37 But it's a matter of taste, could be ^ or ** 00:35:38 What's the T? 00:35:58 I tend to like "terse symbology," but that's just a preference. 00:36:03 Right now I have and, or, not. 00:36:09 Not sure what the T is. 00:36:18 My exclusive or is written as XOR 00:36:24 But may switch to &, |, ~ at some point. 00:36:29 --- quit: dave0 (Ping timeout: 240 seconds) 00:36:31 Mine too right now. 00:36:57 I've gotten faster at typing on the non-ASCII layout of the calculator. 00:36:59 In my mind "exp" says "e^x." 00:37:04 Pretty interesting keyboard. 00:37:13 I mean non QWERTY 00:37:20 Well, and non-ASCII 00:37:23 --- join: rdrop-exit (~markwilli@112.201.162.180) joined #forth 00:37:42 I'm going to add SIN COS TAN and more soon, for floating point. 00:38:08 I'm curious as to how the OS can overload operators like +, so when you add fixed point it just uses fixed point. 00:38:11 Yes, I'll have those as well. 00:38:23 You have to keep up with types. 00:38:24 Or maybe everything's floating point and it just prints out the mantissa 00:38:33 I plan to add that as an alternate interpreter later. 00:38:45 Going to have a "type stack" that gets updated along side the integer and floating point stacks. 00:38:47 I only use fixed point 00:39:04 Matching the types at the top of the type stack becomes a word search impacting thing. 00:39:27 I do plan to do that so it gets worked out at compile time - it won't be something I do at run-time. 00:40:36 Or you could do what I said above - have a vocabulary that demotes + - * / to i+ i- i* i/ and makes the short symbols floating point. 00:40:43 That's not as fancy, but it's part of standard Forth. 00:41:53 How does gforth know about invalid memory access? 00:42:04 0 0 ! throws an exception 00:42:11 I think it uses the processor's memory management capabilities for that. 00:42:33 Hardware thing - the processor itself can throw an interrupt if you violate the memory access rules you've established. 00:43:00 GForth just has arranged to catch and handle that interrupt. 00:43:03 Ah. 00:43:07 You won't get that in a Z80. 00:43:59 Old chip. 00:44:54 What time is it over there? 00:47:48 2:45. I woke up thirsty a little after 1am, probably won't go back to bed. I usually wake up at 4am anyway. 00:47:55 Weekday mornings. 00:48:13 And sometimes that's a struggle - so given that I feel pretty awake I think I'll just run with it. 00:50:02 re: type stack, why not just have type tags on the stack entries? optimization to use the x87 stack without actually manipulating the type data most of the time? 00:50:51 I don't plan to integrate that capability into the base system. 00:50:53 That's one reason. 00:51:09 And the other is that I have an integer stack, like normal, but it's more efficient to keep floats on the FPU. 00:51:24 So two separate stacks - how are the entries logically interleaved? 00:51:31 The type stack will express the interleaving. 00:51:53 As far as the data goes, 1 1.0 and 1.0 1 will produce the same state. 00:52:52 same state except for the type stack you mean? 00:53:55 and you implement polymorphism by dispatching based on contents of the type stack? (apologies if that was covered in part of the scrollback i haven't read) 00:53:59 Right - adding the type stack will add the ability to distinguish them. 00:54:30 Yes, the word search process will require that the word arguments match the type stack "top." 00:54:42 So what I will need will be extra fields in the dictionary, to capture that information. 00:54:59 I probably will express it in the normal way as a stack effect comment. 00:55:12 : + ( f f -- f) ... ; 00:55:24 : + ( i i -- i) ... ; 00:55:37 Now 1 2 + will use the second one - 1.0 2.3 + will use the first one. 00:55:51 And as I noted, that will be tracked at *compile time*. 00:56:08 The compiler can update the type stack (a copy of it, at least) during compilation. 00:56:29 Once words are compiled, they'll just contain the right words, and it will no longer matter how that happened. 00:57:04 right, otherwise you'd waste a lot of cycles doing jump table lookups 00:57:10 (which is fine if performance doesn't matter much) 00:57:13 Once the stack effect of the primitives are specified, the interpreter / compiler will be able to *compute* the stack effect of other definitions, at least in many cases. 00:57:19 right 00:57:32 Yes, dynamic typing brings some added power, but I've decided to forego that. 00:57:44 I think I can make a big step forward with static / compile time handling. 00:57:50 attempted something similar a few years ago but it got too complex and i mothballed it 00:58:18 Yes, I suspect it could get out of hand pretty quickly if it's not done just right. 00:58:26 I tend to "think about" such things a long time before starting to code. 00:58:51 The "get it out the door as fast as possible" mentality in the industry sort of prohibits that sort of thinking most of the time, which is why I like doing this as a hobby. 00:58:58 I can let things "ferment" as long as I want. 00:59:08 Some ideas are "beer ideas," and ferment fast. 00:59:15 Some are "mead ideas" and take a long time. 00:59:31 A statically typed Forth would be enormously useful for beginners. 00:59:39 I've got 20 gallons of mead fermenting since last December. 00:59:57 Literally - metaphor not in play. 01:00:00 --- quit: dys (Ping timeout: 245 seconds) 01:00:16 Going to try it out during the Christmas holidays. 01:00:22 except instead of a type stack i had 40-bit "words" arranged into memory pages where each 4kB page had 768 8-bit type tags, a 256-byte scratch area, and 768x 32-bit tagless part of word 01:00:40 IIRC StrongForth was a statically typed Forth 01:01:45 had plans for doing compile-time type inference where practicable and a too-clever dispatch system for handling common dispatch cases 4 at a time 01:01:45 I've been reading about how the TI operating system does paging, and it's pretty interesting. The reason why I can't execute over $C000 is because that area the the page boundary and wouldn't really work anyway. 01:01:49 is the* 01:01:56 but it became a huge headache in practice 01:02:12 --- nick: dave69 -> dave0 01:02:14 I never cared for such things (statically typed Forths) 01:02:26 Looks like I'm short on memory, I'll need to find a way to get more under $C000 to still remain direct threaded. Because after experimenting indirect threaded it's around 50% slower :( 01:02:36 rdrop-exit: Me either, in a way - that's why I don't want that as "part of my system." 01:03:01 KipIngram: Should be an extension. 01:03:04 I think of it as an application program, designed to let me do stuff similar to what I'd do in, say, Octave (Matlab clone-ish). 01:03:24 So I'd want to be able to extend the operators, to vectors, matrices, etc. 01:03:32 KipIngram: Not sure if I asked you before but have you used Emacs Calc? 01:03:52 I see 01:03:52 No, I used emacs for a period of time as my editor, but didn't really get into all of the "extras." 01:04:06 RPN and has powerful scientific capabilities. 01:04:22 I think viewed that way, rdop-exit, it's not "counter to Forth mentality" at all. 01:04:29 Matrices, vectors, complex numbers, gamma function, bernoulli, rewriting system, bignums, all the things. 01:04:38 It's the nature of the problem I'll be trying to solve. 01:04:42 siraben: i'll check it out 01:04:48 Application-level I agree 01:05:08 Putting it in at the ground floor forces you to carry that baggage with you everywhere. 01:05:14 koisoke: The manual (type C-h i mCalc) is very good 01:05:24 Itâs at the application level that things get fancy and bullit-proofed too 01:05:30 And Forth DOES counsel us to carry baggage only where it's required. 01:05:31 koisoke: C-x * * opens the calculator for you 01:05:43 KipIngram: Precisely. 01:05:46 rdrop-exit: Right - totally agree. 01:06:00 My favorite calculator it the HP-16C 01:06:13 If your application is small enough and you can check that it works, it works, don't need extra baggage. 01:06:18 Never used that one. Is that the "programmer's calculator"? 01:06:19 i seem to have the mode .el installed but not the docs 01:06:31 You could vary the word size of the calculations 01:06:37 it's like the 15c but with base conversions and whatnot 01:06:37 Oh, interesting. 01:06:46 koisoke: What system are you on? Perhaps you could PM me and we could work it out. 01:06:55 https://en.wikipedia.org/wiki/HP-16C 01:07:32 KipIngram: The built in macOS calculator has a pretty good array of programmer oriented features. 01:07:33 Iâve had one since it came out 01:07:43 Although the lack of a stack-based nature makes it somewhat inconvenient. 01:08:01 I just shudder at trying to use non-stack calculator now. 01:08:01 siraben: ubuntu 18.04. i think i just need to install a docs package 01:08:13 I have HP-41C and HP-42S emulators on my phone. 01:08:16 koisoke: Try install texinfo 01:08:21 They're GORGEOUS - look just like the real thing. 01:08:33 siraben: i have texinfo just not that specific document 01:08:35 And emacs25-common-non-dfsg or emacs26-common-non-dfsg depending on your system 01:08:37 I also have a HP48GX, but rarely use it 01:08:44 Put them in full-screen and it's like holding the calculator, except for the different (i.e., nonexistent) key feel. 01:08:48 Huh. 01:09:00 They have haptic key feature, but I opt out of it. 01:09:15 I had one; don't even know where it is anymore. My wife uses that one. 01:09:24 She's 10 years younger, so it was "current" when she was in college. 01:09:48 I have an HP-34S too, which is a "re-flash" of some other HP calculator. 01:10:05 Based on HP-20b and HP-30b. 01:10:08 you mean wp-34s? 01:10:18 When I was in high school we werenât allowed to use caclulators for tests, only sliderules 01:10:26 Yes, sorry. 01:10:42 It's re-flashed and has key stickers and an overlay. 01:10:49 rdrop-exit: What? Really? 01:11:00 Yup 01:11:03 I'm lucky. My calculator really is darn powerful 01:11:03 :-) 01:11:07 Feels like cheating sometimes 01:11:10 How old are you, rdrop-exit? 01:11:19 Only 56 01:11:28 I'll be 56 in January. 01:11:30 KipIngram: what about you? 01:11:47 when i was in college we were allowed calculators but i used a slide rule because the ti-36 i had made me too angry to concentrate 01:12:06 Almost 18 here, so you guys have been alive for more than three times I have. 01:12:20 I bought a cheap plastic slide rule freshman year in college and forced myself to use it for a few weeks in my "Engineering 101" class. 01:12:32 Just so I could say I had some experience with it, and so I understood how it worked. 01:12:36 The scale of a decade is just unimaginable to me. 01:12:37 :) 01:12:56 I read about slide rules, seems pretty handy but I can't imagine how I would have used them. 01:12:58 Hm. 01:12:59 Yeah, the period since the 1960's has truly been the "electronics era." 01:13:04 AMAZING progress. 01:13:29 That's good. So we can finally use the advanced technology to solve problems like climate change 01:13:30 Believe it or not, when I was in 1st grade we had inkwells 01:13:32 Which has gotten so much worse over the same period. 01:13:40 When you consider that the transistor wasn't even invented until the 1950's. 01:13:54 A few months ago I read about a lab in Israel that had implemented a 12-atom transistor. 01:13:55 1947 i thought 01:14:01 Could be. 01:14:09 Maybe we picked it up in Roswell. ;-) 01:14:11 but not commercially useful until a bit afterthat 01:14:40 If they commercialize "few atom" transistors in, say, the next 30 years, we'll literally have gone from birth to theoretical limit in silicon electronics in a century. 01:15:11 integrated circuits becoming cheap was a big phase change i wish i could have seen in person 01:15:17 And despite this progress software latency has gotten worse. 01:15:24 Yes, it was pretty remarkable. 01:15:28 I suppose computers felt faster in the 80s even though they weren't, right? 01:15:31 Just always something new to learn. 01:15:47 Yes, turn it on and itâs ready 01:15:47 Well, yes, I think so. 01:15:59 They'd feel faster now if we hadn't bogged them down with software bloat. 01:16:25 I used to get pissed every time my Windows machine too five minutes to boot. 01:16:27 "A 1000x speedup should bring a 60 second boot time to 60 milliseconds." 01:16:29 No excuse for it. 01:17:03 eh, boot time is a special case since you are waiting for (sometimes mechanical) hardware to respond at the rate it responds 01:17:22 Well, I think the way they go about it has an effect too. 01:17:42 Why not just have a RAM image of the freshly booted system, and just copy it into RAM and say "go." 01:18:06 siraben: in the 80's, computers had BASIC in rom, so it never failed to boot and it took a fraction of a second to boot up 01:18:11 Loading a few GB from an SSD into RAM should be fast as greased lightning. 01:18:22 Of course, there's some hardware initialization to do too, but it shouldn't take long. 01:18:30 yeah you could almost do that if you designed the freshly booted system state to cope with arbitrary changes on disk 01:18:42 then it would be just a special case of hibernate 01:18:49 My Emacs "boots" up in 5 seconds flat, with a heck of a lot of configuration. 01:19:21 Thatâs just sad 01:19:22 KipIngram: Also considering the push towards things like ASLR for security measures 01:19:24 My Mac isn't as offensive - it will boot in a minute or less. Not as fast as it should, but fast enough not to annoy me. 01:19:48 I just never turn my Mac off 01:19:56 Well, that also bothers me - we've gotten ourselves into a bad situation security-wise. 01:20:14 It should be possible to have a "personal" computer that doesn't have to be as security conscious as a multi-user OS would need to be. 01:20:22 I tend to keep Emacs (and my computer) running for weeks so I don't worry about startup time. 01:20:25 Back to "baggage being carried where it's not needed." 01:20:35 Me too. Or months. 01:21:03 KipIngram: Funnily enough, it's still hard to have actual multi-user configurations. Try setting that up on Macs and see how isolated each user is. 01:21:26 Well, that wouldn't surprise me - I've never tried. 01:21:37 Iâm thinking of getting an old machine and trying out NetBSD as my development environment 01:21:50 And the fact that we "need that security" to protect us against attacks from the outside? Well, that's just a pathetically broken infrastructure. 01:21:56 KipIngram: Mobile phones in a sense are "personal computers", single user. 01:22:02 Yes. 01:22:04 But still with a lot of security measures in place. 01:22:23 Things get much more complicated them you have networked devices that can run arbitary code. 01:22:33 It'll be an endless cat-and-mouse game. 01:22:39 Right - "arbitrary code" being the key problem there. 01:22:52 I think we took a misstep when we broke down the barrier between "code" and "data." 01:23:00 "active content" on the web, and so on. 01:23:07 I understand why they thought they needed it at the time. 01:23:13 When networks were a lot slower. 01:23:17 But do we still need that? 01:23:19 I'm not so sure. 01:23:36 Apparently so. Browsing the web without JavaScript gets difficult quickly. 01:23:44 Barring reading articles and so on 01:23:48 Right. 01:23:58 try turning off javascript... the www becomes unusable 01:23:59 But that's a "self-imposed" need - not a "fundamental need." 01:24:08 We need it because we've decided to need it. 01:24:08 siraben: i agree 01:24:37 I use uMatrix so JavaScript is disabled by default 01:24:43 I suspect that with mature modern networking technology we could solve those problems a different way. 01:25:17 But there's just no way such a "fundamental culture reboot" is going to happen. 01:25:58 Not to mention the profileration of tracking on the internet, oh God. 01:26:08 proliferation* 01:26:21 Right - I was just about to mention that there are forces out there that don't WANT a better solution - they profit from the current model. 01:26:33 So the tech is there, it's just that our society has poorly adapted to its growth. 01:26:36 Commercial ones, government ones, etc. 01:26:47 That's what I think, yes. 01:26:50 I heard software gets really really bad at the government level. 01:26:55 Ironic because that's where it should be the best. 01:27:10 Hah hah - everything gets bad at the government level. 01:27:15 "Processes" of all kinds. 01:27:17 Can any of you recommend an IRC client for Macs, the one Iâm using is mixing up the incoming order of messages 01:27:21 Why should software be exempt? 01:27:32 rdrop-exit: Which one are you using? 01:27:35 I used to use ERC 01:27:41 Colloquy 01:27:42 (Emacs IRC client) 01:27:43 Sorry - I actually run a console client on a remote machine and use mosh to connect to it. 01:27:55 Ok thanks 01:28:02 But now I use Riot which connects to a Matrix-IRC gateway 01:28:25 rdrop-exit: Are you sure the order of messages is wrong? 01:28:35 Compare it to the log: http://bespin.org/~nef/logs/forth/18.10.22 01:28:43 ok, BRB 01:29:11 Actually my order of messages is a little off, too, but I don't mind. 01:29:36 Different order, it seems to clustering incoming messages by author, kinda confusing 01:29:40 I blame the latency of the gateway, but I gain 24/7 connectivity to these channels. 01:29:49 Isn't that just a feature of IRC? 01:29:55 What is? 01:30:04 I might not get the messages here in the same order that, say, siraben does. 01:30:17 Maybe it's just the internet. 01:30:22 I'd expect each of our clients to display them in arrival order. 01:30:32 Iâm new to IRC, hadnât used it in decades 01:30:34 I've NEVER seen my client insert a message anywhere except the bottom. 01:30:38 Most IRC clients just save your message when you hit return 01:30:45