00:00:00 --- log: started forth/18.10.16 00:13:38 --- join: nerfur (~nerfur@broadband-95-84-184-13.ip.moscow.rt.ru) joined #forth 00:20:13 --- quit: dys (Ping timeout: 252 seconds) 00:22:36 --- quit: kumool (Quit: Leaving) 00:31:27 --- quit: wa5qjh (Quit: http://quassel-irc.org - Chat comfortably. Anywhere.) 00:33:35 --- join: wa5qjh (~quassel@175.158.225.216) joined #forth 00:33:35 --- quit: wa5qjh (Changing host) 00:33:35 --- join: wa5qjh (~quassel@freebsd/user/wa5qjh) joined #forth 00:45:57 --- join: ncv (~neceve@2a02:c7d:c5c9:a900:6eaf:6ef7:3b81:d5f6) joined #forth 00:45:57 --- quit: ncv (Changing host) 00:45:57 --- join: ncv (~neceve@unaffiliated/neceve) joined #forth 00:51:31 --- quit: wa5qjh (Quit: http://quassel-irc.org - Chat comfortably. Anywhere.) 00:59:31 --- quit: smokeink (Remote host closed the connection) 00:59:54 --- join: smokeink (~smokeink@218.255.173.76) joined #forth 01:00:38 --- quit: smokeink (Read error: Connection reset by peer) 01:00:59 --- join: smokeink (~smokeink@118.131.144.142) joined #forth 01:02:25 --- quit: smokeink (Remote host closed the connection) 01:02:45 --- join: smokeink (~smokeink@185.189.254.154) joined #forth 01:02:46 --- quit: smokeink (Remote host closed the connection) 01:03:07 --- join: smokeink (~smokeink@61-216-40-75.HINET-IP.hinet.net) joined #forth 01:03:44 --- quit: smokeink (Remote host closed the connection) 01:07:45 --- join: wa5qjh (~quassel@175.158.225.216) joined #forth 01:07:45 --- quit: wa5qjh (Changing host) 01:07:46 --- join: wa5qjh (~quassel@freebsd/user/wa5qjh) joined #forth 01:18:59 --- quit: wa5qjh (Remote host closed the connection) 01:52:04 --- quit: jedb (Ping timeout: 252 seconds) 02:04:26 --- quit: ashirase (Ping timeout: 246 seconds) 02:08:39 --- join: wa5qjh (~quassel@175.158.225.216) joined #forth 02:08:39 --- quit: wa5qjh (Changing host) 02:08:39 --- join: wa5qjh (~quassel@freebsd/user/wa5qjh) joined #forth 02:14:21 --- join: ncv_ (~neceve@86.127.177.21) joined #forth 02:14:21 --- quit: ncv_ (Changing host) 02:14:21 --- join: ncv_ (~neceve@unaffiliated/neceve) joined #forth 02:15:35 --- quit: wa5qjh (Remote host closed the connection) 02:16:43 --- quit: ncv (Ping timeout: 252 seconds) 02:17:26 --- join: wa5qjh (~quassel@175.158.225.216) joined #forth 02:17:26 --- quit: wa5qjh (Changing host) 02:17:26 --- join: wa5qjh (~quassel@freebsd/user/wa5qjh) joined #forth 02:23:13 --- quit: wa5qjh (Remote host closed the connection) 02:24:10 --- join: ashirase (~ashirase@modemcable098.166-22-96.mc.videotron.ca) joined #forth 02:25:02 --- join: wa5qjh (~quassel@175.158.225.216) joined #forth 02:25:02 --- quit: wa5qjh (Changing host) 02:25:03 --- join: wa5qjh (~quassel@freebsd/user/wa5qjh) joined #forth 03:14:24 --- quit: wa5qjh (Remote host closed the connection) 03:16:24 --- join: wa5qjh (~quassel@175.158.225.216) joined #forth 03:16:24 --- quit: wa5qjh (Changing host) 03:16:24 --- join: wa5qjh (~quassel@freebsd/user/wa5qjh) joined #forth 03:40:20 --- join: ncv__ (~neceve@90.207.216.74) joined #forth 03:40:20 --- quit: ncv__ (Changing host) 03:40:20 --- join: ncv__ (~neceve@unaffiliated/neceve) joined #forth 03:43:10 --- quit: ncv_ (Ping timeout: 252 seconds) 04:03:17 --- quit: KipIngram (Quit: WeeChat) 04:05:44 --- quit: nighty- (Quit: Disappears in a puff of smoke) 04:16:56 --- quit: wa5qjh (Quit: http://quassel-irc.org - Chat comfortably. Anywhere.) 04:18:57 --- join: KipIngram (~kipingram@185.149.90.58) joined #forth 04:18:57 --- mode: ChanServ set +v KipIngram 04:19:26 --- join: wa5qjh (~quassel@175.158.225.216) joined #forth 04:19:26 --- quit: wa5qjh (Changing host) 04:19:26 --- join: wa5qjh (~quassel@freebsd/user/wa5qjh) joined #forth 04:21:07 Changing from direct to indirect is strangely buggy still, will retry again on the weekend 04:21:16 Even on the "minimal" Forth 04:27:10 --- join: jedb_ (~jedb@199.66.90.113) joined #forth 04:31:05 --- nick: jedb_ -> jedb 04:35:43 What sort of bugs? 04:36:09 And is it "unstable" or "doesn't work"? 04:36:26 Doesn't work. 04:36:42 Also, I lose TOS, which means I have to rewrite most of the CODE words 04:36:52 I suppose the gain is that I can get to have kilobytes of space instead 04:37:17 --- join: nighty- (~nighty@s229123.ppp.asahi-net.or.jp) joined #forth 04:37:42 As it stands, I still have a measly 400 bytes of space after HERE, which means loading larger programs won't work. 04:38:13 I'm not clear yet on how it cost you a register - you said you already had a W. 04:38:22 But I need an X, right? 04:38:36 What are you calling X? 04:38:43 Another spare register? 04:39:10 Ok, walk me through the registers again, of your direct-threaded Forth. 04:39:17 Or, let me start. 04:39:21 Right 04:39:24 You start. 04:39:24 You have an IP (instruction pointer). 04:39:27 --- join: dddddd (~dddddd@unaffiliated/dddddd) joined #forth 04:39:49 So tell me how your direct threaded NEXT operates, step-by-step. 04:40:51 (IP) -> W 04:40:56 IP+2 -> IP 04:40:58 JP (W) 04:41:04 That's it 04:41:07 Ok. 04:41:31 DOCOL is like this: 04:41:37 PUSH IP 04:41:41 W+2 -> IP 04:41:45 NEXT 04:41:59 Ok, so you can avoid having to use another register. 04:42:11 And yes, I feel like indirect threaded should consume an extra register. 04:42:15 but you can do this: 04:42:20 NEXT: 04:42:29 (IP) -> W 04:42:34 IP+2 -> IP 04:42:40 (W) -> W 04:42:44 JP (W) 04:42:52 Now - you no longer have W pointing at your word. 04:43:09 So when you need that (non-primitive), you have to reocover it from IP again. 04:43:20 (IP-2) -> W 04:43:34 That's not fun, but it's FAR better than losing TOS register. 04:43:35 The problem is that I can't do (W) -> W 04:43:42 Oh. Why? 04:43:46 Z80 04:43:55 I can't do it directly, anyway 04:44:27 Also, the only register that can do JP (W) is the 16-bit HL register 04:44:30 Well, that sucks. But ok - yeah; those old processors had a lot of limitiations there. 04:44:52 So, this means I need to do LD HL, (HL) (dest, src) 04:45:01 What's your whole list of registers? 04:45:05 Which means, I need to deference HL into something else 04:45:15 DE, HL, BC, IX, IY mainly 04:45:24 IX is for the return stack, IY is used by the OS for interrupts 04:45:27 So I'm left with DE HL and BC 04:45:37 Technically AF as well, but A is the accumulator 04:45:56 Maybe I can salvage IY 04:46:13 * KipIngram drumming fingers on desk... 04:46:15 Hmmm. 04:47:31 Which register is your data stack pointer? 04:47:38 PSP 04:47:43 I mean SP 04:47:48 Oh, ok, the regular stack. 04:47:58 I could save/restore it, but that would add extra overhead 04:48:06 But in DOCOL you said PUSH IP. 04:48:15 Yes 04:48:16 That makes that one seem like the return stack. 04:48:19 OH sorry 04:48:25 I mean PUSH_IP_TO_RS 04:48:30 push IP to the return stack 04:48:35 "push" 04:48:41 Ok. 04:48:55 exit is: POP_IP_FROM_RS \ NEXT 04:49:03 So the problem here is that you are starved for registers that can address RAM. 04:49:25 Right. 04:49:43 Losing TOS in a register hurts. But I think anything you do here is going to hurt. 04:49:56 Oh, well. 04:49:59 Do you have a register-to-register xchg instruction? 04:50:15 ex de, hl 04:50:20 Is the only valid exchange 04:50:26 Otherwise it has to go through the 8-bit accumulate 04:50:30 Urgh. 04:50:30 accumulator* 04:50:50 e.g. to do BC -> HL you have to do: 04:51:10 ld a, c \ ld l, a \ ld a, b \ ld h, a 04:51:16 Yeah. 04:51:40 And each load takes around 4 cycles, but NEXT is something that I would be running a lot 04:51:41 Well, shoot. I don't really see any way out that wouldn't probably impact you as much as not having registered TOS. 04:51:47 You have several ways out, but they all cost. 04:51:52 Right. A full rewrite is needed. 04:52:03 Just the primitives, the WORD words don't need to be 04:52:36 Well, isn't just add a pop at the beginning and a push at the end of all the words? 04:53:04 Yeah 04:53:20 Oh becaues the register allocations will be different 04:53:27 Oh. Ok. 04:53:35 So in your test version, how is it misbehaving? 04:53:42 Or can you tell yet? 04:53:48 It just crashes. 04:54:10 My current strategy is an unusual one. To accomplish (W) -> W then JP (W) 04:54:32 I do this: (W) -> W then perform self-modifying code to change the destination of the JP 04:55:26 Hey, if it works... I've never shared the professions aversion to self-modifying code. I understand the general discomfort, but basically if something WORKS it works. 04:55:45 i.e. I deference W into itself, then deference W again into the two bytes after JP 04:55:50 Theoretically it should work, I'm not sure why 04:55:52 Right. 04:55:58 why it's not * 04:56:35 So basically you're re-writing NEXT on every word. 04:56:45 Yeah 04:57:17 In an effort to try to not change my current register allocations 04:57:30 Right. 04:57:55 Exactly the sort of thing I'd do. So no chance you have the endieness of that jump target wrong? 04:58:26 I thought that it might, but I really don't think so. 04:58:36 Should be JP [lowest byte] [highest byte] 04:58:56 Ok. 04:59:12 And JP is an absolute jump? 04:59:14 Not relative? 04:59:19 Absolute. 05:01:42 I guess you can't push a memory word, can you? 05:01:48 push (w)? 05:02:10 Nope. 05:02:21 If you could do that then push (w); ret would get you there. 05:02:21 Only PUSH reg_16 or PUSH index_reg 05:02:36 Ah, if I could, yes. 05:03:18 Just as a test, have you tried push ix; w->ix; (ix)->w; pop ix ? 05:03:22 One thing could be to blame: my macros. 05:03:57 Huh there's EX (SP), HL 05:04:33 I'll try that 05:04:45 I tend to avoid the index registers though, they're slow. 05:05:08 Interesting. 05:05:19 Well, "slow" is relative. 05:05:29 Takes microseconds but they add up 05:05:30 I have to keep reminding myself how old this tech is. 05:05:38 We were still "toddlers." 05:06:23 This would never be a problem if I was able to execute RAM above 0xC000 05:06:32 Yeah. 05:06:39 So that's deliberately built into the calculator? 05:06:49 The operating system, I believe 05:07:12 The assembler really isn't helping me with this too. I'll need to write my own later. 05:07:17 I think it must be in the hardware somehow - the OS shouldn't know what code you're executing. 05:07:48 Ah, it's probably a hardware thing then. 05:07:56 If OS code is executing, you IP value isn't in IP. 05:08:17 yeah - they must or the top bits of the program counter and send that into an interrupt. 05:08:54 Maybe you could mask interrupts for the brief time you're passing through you direct threading preambles? 05:09:21 I haven't learned interrupts yet. Maybe that's what I'm missing. 05:09:22 NEXT would mask just before it does the jp (w)? 05:09:32 I'll check my guide on that 05:09:46 Ok. Problem there is that there usually is a "non-maskable interrupt." 05:09:55 If they're using that one then you can't beat them that way. 05:10:08 There seems to be an interupt mode called mode 2: "The most fun. In mode 2, the CPU can theoretically jump to any address in memory. This is the interrupt mode we are interested in." 05:10:26 But there's usually an instruction or a flag in the flag registers for masking maskable interrupts. 05:11:16 NEXT could mask just before it does the JP (W), and then you'd unmask at the start of all the routines that you get to from the preambles. 05:11:44 I'm assuming those bits of code that JP (W) send you to are themselves jumps? 05:11:50 Jumping you back down to low RAM? 05:12:03 I.e., you have jp docol before your definitions? 05:12:17 Wow, a bunch of instructions I haven't used yet: 05:12:26 EXX EI IM EI etc 05:12:45 I have a CALL DOCOL before my word definitions 05:13:11 Ah, so that puts the address you need to load into IP on your stack. 05:13:41 So yeah, figure out how to mask interrupts, mask just before JP (W) in NEXT, and then unmask as the first thing in docol. 05:14:03 And dovar, and do. 05:14:18 Sorry, I don't completely understand the use of interrupts. 05:14:19 I'll read more on it. 05:14:46 You know, it's interesting. With your direct threaded thing you have 05:15:05 CALL DOCOL; etc. 05:15:12 And indirect threaded just needs 05:15:21 DOCOL; etc. 05:15:24 Right. 05:15:40 So that's really already there - you'd just be incrementing all of the cells of your definition by one byte. 05:15:49 Point just after the call opcode instead of at it. 05:15:57 Should work even if it was still there. 05:16:14 But you have your headers back there too, I think. 05:16:44 Usually people don't CALL DOCOL. 05:16:48 They JP DOCOL. 05:17:04 You have W pointing there, so incrementing it past the call instruction gets you the thing you need for IP. 05:17:15 You're using macros as well, right? 05:17:21 Yes. 05:17:22 Heavily. 05:20:22 This is perhaps the most constrained target I'm going to write for, right? 05:21:48 I think so, yes. 05:21:57 In its day the z80 was considered a big advance over the 8080. 05:22:04 But I doubt you'll wind up needing to program it. 05:22:34 Every other assembly language seems to be so high-level 05:22:37 Like the x86 05:22:41 Wow 05:23:08 To bad this isn't a 6809 - I think of it as pretty much the best of those early generation units. 05:23:13 Very "orthogonal." 05:23:41 My very first Forth was on a 6809 - it was the brains of the TRS-80 Color Computer. 05:24:38 Around how old were you when you did that? 05:26:26 Oh, 20? 05:26:38 It was horrible. 05:26:40 It worked. 05:26:47 Horrible how? 05:26:49 But I had no real knowledge of Forth internals when I did that. 05:27:11 Later, when I got hold of a book outlining Forth "guts," I was blown away by how simple it all was when done right. 05:27:17 O.o it's possible to implement a Forth without understanding its internals? 05:27:37 Sure. You're just likely to overcomplicate a lot of things. 05:27:48 I need to find a good book on implementing Forths 05:27:57 A lot of "firsts" are almost always the most sub-optimal 05:28:12 Forth Fundamentals volume 1 is the book I used, and it's excellent. 05:28:16 McCabe. 05:28:20 It's hard to find. 05:28:41 It's not even on library genesis 05:28:51 I've got a copy, and at some point I'm going to find a way to get it online. 05:28:57 But I don't want to ruin the book in the process. 05:29:02 Just haven't tackled it yet. 05:29:04 So what are some possible reasons why Forth isn't as popular as it used to? 05:29:11 It's too good to be "lost" to the community. 05:29:30 Lisp has exploded in popularity and reverence 05:29:59 ^it also helped that there was an MIT course called Structure and Interpretation of Computer Programs, that made Lisp shine 05:30:00 Well, I think a couple of possible reasons are 1) one of Forth's biggest strengths is that it's low-RAM-consuming. 05:30:04 But that hardly matters anymore. 05:30:16 Ah, it's niche. 05:30:48 2) It's not "intellectual" in the sense of obeying some formal set of syntax rules. It's fairly arbitrary, in actual use. 05:31:00 Programming languages shouldn't have ambigious order of operations. Those come from algebra but don't fit the world of computing 05:31:02 I consider that an advantage, but I think people like following rules. 05:31:18 Finally, I think Forth people tend to alienate other parts of the community. 05:31:26 How so? 05:31:34 We have a tendency to go around talking about how Forth is just clearly better than anything they're doing. 05:31:41 We act kind of superior about it, and that doesn't win friends. 05:31:59 Aren't Lispers the same way? 05:32:10 And also Forth really just doesn't play well with other languages. It likes to be in charge. 05:32:24 (unfortunately, I haven't had anyone to advocate Forth to in real life) 05:32:39 I've always found that I can do good work with a system that's Forth all the way to the hardware. It just doesn't fit into our modern "write massive libraries" mentality. 05:32:52 Plus it's HARD. Easy to learn, but hard to become *good*. 05:33:02 It takes a lot of work to understand an appication well enough to write good Forth. 05:33:06 Right. You need complete control in Forth. 05:33:11 Also it has no security model 05:33:17 Python, say, makes it a lot easier to hack something out quick and dirty. 05:33:33 Right. 05:33:38 It's security model is "none." 05:33:48 Because any security = slower 05:33:49 Here's the universe - you're God. 05:34:29 Yeah, I was thinking as I drove to work yesterday about the comparitive speed of interrupts and polling. 05:34:47 Really, interrupts should be at least about as fast as the best polling loop you could write. 05:34:52 Often should be a hair faster. 05:34:59 Frankly, "god mode" can be dangerous for some embedded systems. 05:35:07 But invariably in modern systems an interrupt means a security level change - a context switch to the kernel. 05:35:09 I wouldn't want to open a REPL on a pacemaker 05:35:11 And that's expensive. 05:35:22 No doubt. 05:35:47 But I can't help but look at embedded systems through the eyes of a developer. 05:35:53 And as a developer, I want to be able to do anything. 05:36:16 Right. But one question is _should_ you be able to do anything? 05:36:20 Especially on life-critical systems. 05:36:31 Certainly not, to a real pacemaker in a live human. 05:36:42 But that part comes later, after I've done my thing. 05:37:25 I don't understand the absurd limitation on mobile phones not being very programmable 05:37:38 All of our big-time operating system sare designed with the idea that they will be populated with adversarial users, simultaneously. 05:37:44 So the thick security layer is required. 05:37:47 There's really no harm in that, just use sandboxes, tc. 05:37:48 etc 05:37:56 Have you heard of Qubes OS? 05:38:00 No one ever uses my Mac notebook except me, but I bear that layer anyway. 05:38:05 No. 05:38:24 It goes all the way by allowing you to have applications in specific "workspaces", which are equivalent to virtual machines 05:38:37 So your browser would be completely isolated to your personal information, for instance. 05:38:50 A single failure is just a single failure 05:38:58 Yes. Containers. 05:39:17 I haven't been able to get it working on my Mac, sometime later when I switch to a PC 05:39:48 I'm writing my Forth with the view that it will be completely in control of any machine it's running on (eventually), and that 100% of the software on the machine is friendly and cooperative. 05:40:32 https://blog.invisiblethings.org/2015/12/23/state_harmful.html 05:40:42 "State considered harmful - A proposal for a stateless laptop" 05:40:52 Pretty good paper 05:40:58 Cool. 05:41:16 Though of course what a computer IS is a state machine. 05:41:25 So they must be using "state" in a particular way. 05:41:46 A good world would be one in which you carry your data with you but share laptops and so on. You can verify that the laptop is secure and plug in your data and use it as if it was yours. 05:42:06 Then when you're done you save the image onto your drive and leave no trace on the hardware 05:42:25 I do think that would be a good world, but it involves a lot of trust of the hardware. 05:42:38 --- join: rdrop-exit (~markwilli@112.201.162.180) joined #forth 05:42:49 You heard of "digital dead drops?" 05:42:50 A lot of old computer systems that underly our world will have to change 05:43:01 I've heard of those, but they sound like a good way to get viruses 05:43:07 Exactly. 05:43:22 They're the flip side of what you just described: bring your own hardware, here is data. 05:43:37 But wow - the analogy with STDs is striking. 05:43:50 The fundamental problem with security is that it's harder to tell when you're safe _digitally_ than physically. 05:43:59 Haha 05:44:32 I don't know if my laptop was compromised by an evil maid attack, but I can tell if someone broke my lock 05:44:43 Right. 05:44:59 We erred when we started building computers so that digital maid attacks were even possible. 05:45:18 Research in encrypted computing is still in its infancy 05:45:20 But the concept is fascinating. 05:45:21 To respond to something from the outside in some heavy way, without your user authorizing / requesting it... 05:45:28 Yes it is. 05:45:37 --- quit: z0d (Quit: leaving) 05:45:49 I feel the same way about "active content" on the wb. 05:45:51 web 05:46:01 Browsers should fetch data and render it on your screen. 05:46:03 End of story. 05:46:31 We've softened the boundary between "code" and "data" too much. 05:46:34 active content? 05:46:49 I have javascript disabled by default 05:46:51 Like Javascript and other such things that "execute" on the client PC. 05:47:09 Around 80% of the time it works fine except for when I need to sign in/use interactive content 05:47:20 I use uMatrix, great add-on to Firefox 05:47:32 We probably don't really need it as much as we did. 05:47:45 There was a period in there where networks were slower, and it enabled very high-performance web experiences. 05:47:51 I enable JavaScript for some websites like Wikipedia, DuckDuckGo 05:47:54 A cleaner model would probably work in today's world. 05:47:54 Right 05:47:59 Now a lot of JavaScript is junk 05:48:25 Also not to mention that JavaScript is so slow and inadequate as a language people are compiling TO JavaScript 05:48:37 Or bypassing it completely with webassembly 05:49:05 It was just an example. I meant to refer to anything where you hand control of your processor over to something embedded in the web content. 05:49:20 Right 05:49:34 --- quit: ashirase (Read error: Connection reset by peer) 05:49:52 Just viewing a website should not make your computer vulnerable - that was a bad step. 05:51:24 Query injection attacks are similar - the dbms is allowing data to be treated as code. 05:51:36 It should have a strong wall between "this is query" and "this is data." 05:52:15 Not to mention that JSON is often eval'd because it is compatible with JavaScript syntax 05:52:27 Yes. :-) 05:52:47 One thing you can count on the world for: if you open a door, someone will walk through it. 05:53:11 One problem is that when a lot of our network infrastructure was first developed the community was mutually trusting. 05:53:21 Scientists, military, etc. - all a team. 05:53:31 I sometimes wonder what an alien civilization's computing would look like 05:53:32 Then we dropped it on everyone without fixing those things. 05:53:40 +KipIngram> I'm writing my Forth with the view that it will be completely in control of any machine it's running on (eventually), and that 100% of the software on the machine is friendly and cooperative. 05:53:42 How much of it is really fundamental? 05:53:43 :-) 05:53:50 outside of an embedded environment, you know you'll never get that, right? 05:54:05 I just laughed when they uploaded a virus to the alien network on Independence Day. 05:54:09 even if you replace the OS, there's still the firmware, the microcode, etc. 05:54:14 Wow - they use the same protocols we do... 05:54:37 Ha 05:54:46 Definitely impossible 05:54:56 yeah, what I primarily meant by that is that there won't be other software on the system being run by other users. 05:55:09 Binary is fundamental, right? 05:55:24 no 05:55:26 NASA attempted to exploit that in their Voyager golden discs 05:55:27 siraben: They could have at least put in a few lines about how they'd had that alien ship at Area 51 for decades, and had studied their computers. 05:55:34 They could have made it ok, with just a few lines of dialog. 05:55:46 Perhaps 05:55:56 --- join: ashirase (~ashirase@modemcable098.166-22-96.mc.videotron.ca) joined #forth 05:56:01 It still would have been ridiculous for someone who'd known nothing about it all until a few hours prior to have done the viral attack. 05:56:13 It would have been a staffer who'd been studying the stuff for years. 05:56:34 Then again, Hollywood never strives for accuracy 05:56:42 No, they don't. 05:56:58 And that's ok, I guess, but I think we have hoardes of terribly misinformed people out there. 05:57:15 But anyway, that movie was worth it for Bill Pullman's speech. 05:57:34 Talk about the (mis)representation of programming in movies 05:57:38 It's crazy, right? 05:57:41 Yes. 05:57:43 Black Hat. 05:57:55 Oh my God - I felt physical pain during that movie. 05:58:34 https://en.wikipedia.org/wiki/Gell-Mann_amnesia_effect 05:59:38 (perhaps relevant to the Hollywood discussion) 05:59:56 Yes. 06:00:20 I'm an expert - I watched ." 06:01:24 "We need to start turning away from media." 06:01:25 Amen. 06:01:36 The media has long since stopped being mere "distributors of information." 06:01:44 They actively try to steer our thinking. 06:02:17 Computing is especially misrepresented because 1. most people don't understand it 2. it looks cool and magical 06:02:26 Right, exactly. 06:02:36 and 3. the people who perform the tasks are wizards 06:02:37 It *is* cool, but Hollywood needs it to be cooler. 06:02:47 To appreciate it's actual coolness, you need some knowledge of it. 06:02:51 And the audience doesn't have that. 06:03:07 Just last week, I was taking notes on Emacs and it was mirrored to the screen and there was always the occasional glance like "how did you do that"? 06:03:27 For instance, I did the equivalent of 20 undos in a couple of keystrokes 06:03:43 Yes. 06:03:44 "Any sufficiently advanced technology is indistinguishable from magic" 06:03:52 Clarke rocked. 06:04:03 I have felt that sensation when looking at videos of other users 06:04:25 Watching someone who really knows how to fly vim or emacs code is... entertainment. 06:04:28 Especially before getting to know my editor 06:04:30 It can be quite impressive. 06:04:47 I don't hold a candle to a few of the whiz kids in my office. 06:05:15 Google Docs is painfully slow, I can't believe people tolerate the latency 06:05:29 yeah - it's crazy. 06:05:31 Eclipse. 06:05:32 ~ 5 seconds to just get from URL to being able to edit the file 06:05:33 Can't stand it. 06:05:36 Or more 06:05:52 There's really no excuse to make things fast, right? 06:06:06 You mean not to? 06:06:16 Right, not to make them fast 06:06:21 That's how I feel. 06:06:35 We have these machines that are just absolutely remarkable compared to what was avaialble a few years ago. 06:06:57 But we've bloated our OSes and done other things that just throw that power out the window. 06:07:56 We should have a rule, that at least 90% of the power of your processors has to reach the end user application. Or something like that. 06:08:01 And I'm joking, I don't believe in such rules. 06:08:08 But wow - we're so far from that it's crazy. 06:08:50 Unfortunately on macOS you can't replace the window manager with something else resource intensive. 06:09:01 We've sort of done it to ourselves, though. People drool and go "oh, ah" over glitzy graphical effects in the os and so on, instead of focusing on what the thing lets them get done. 06:09:14 I have used i3 on Linux before, wonderful software. 06:09:32 i3? 06:09:49 A window manager 06:09:53 ah 06:10:14 If you peek into a restaurant kitchen, you don't see polished marble tables and oak closets, you see efficiency, and it ain't pretty. 06:10:20 That's how I feel about it 06:10:34 RIGHT - good analogy. 06:10:38 Of course, aesthetics matter but only in the final product, that's why I use LaTeX for paper writing 06:10:49 But peek into an upscale suburban home kitchen. 06:10:53 You don't see efficiency. 06:10:56 You see "show." 06:11:03 Ugly but efficient writing process but beautiful output 06:11:30 My wife and I have strongly different opinions on kitchens. 06:11:35 I love that quote from Starting Forth 06:11:36 She likes the swanky looking status symbol stuff. 06:11:41 I want a WORKING kitchen. 06:11:53 I stay out of the kitchen :) 06:11:58