00:02:50 -!- augur has quit (Remote host closed the connection). 00:04:25 -!- hppavilion[1] has joined. 00:08:46 <\oren\> anyone else watchin the handegg? 00:11:12 <\oren\> otherwise known as "armored rugby"? 00:11:35 <\oren\> you don't need a cable thruough the earth if we can do it with a neutrino beam <-- i think bandwidth / energy might be a problem there. 00:12:11 because you only detect a tiny fraction 00:13:15 (this concludes my logreading.) 00:14:17 * shachaf generates more logs. 00:14:18 > pi/2 00:14:20 1.5707963267948966 00:14:42 also, the time saving isn't that big compared to just going around... 00:15:06 shachaf: your scheme fails on account of me being on the channel hth 00:15:52 logs are logs 00:17:21 <\oren\> glog glog glog 00:31:25 -!- tromp_ has joined. 00:35:48 -!- tromp_ has quit (Ping timeout: 248 seconds). 00:50:58 -!- jaboja has joined. 00:51:18 -!- augur has joined. 00:51:48 -!- augur has quit (Remote host closed the connection). 00:52:33 -!- augur has joined. 00:54:34 -!- lynn has quit (Ping timeout: 256 seconds). 00:59:35 > pi/sqrt 2 00:59:37 2.221441469079183 00:59:42 @messages- 00:59:42 boily asked 9h 48m 20s ago: since when are we misleading the topics? everything makes sense, eh? 01:00:23 @messages 01:00:23 You don't have any messages 01:00:25 @tell boily I didn't set the "misleading topics" thing 01:00:25 Consider it noted. 01:00:38 @tell boily I just added "esoteric" for consistency 01:00:38 Consider it noted. 01:01:02 We should have the topic on revision control or something 01:01:20 -!- jaboja64 has joined. 01:02:35 -!- jaboja has quit (Ping timeout: 264 seconds). 01:05:08 -!- variable has joined. 01:09:05 Elronnd: GitHub for IRC? 01:10:18 Elronnd: Speaking of which, I should probably GitHub my personal server and client soon so people can actually visit my personal network 01:11:03 isn't github really expensive though? 01:12:17 Elronnd: ... 01:12:17 No 01:12:26 GitHub is the free one, unless you want private repos 01:12:29 Which I don't 01:12:49 Or if you want their backend to run on your own servers, I think that's also expensive 01:12:51 Or at least big 01:13:25 How would you put github on your server if you didn't get "their backend to run on your servers"? 01:15:30 Elronnd: No, I was going to put the server/client on the github website 01:15:36 Elronnd: I see the confusion 01:15:39 ohhh 01:15:48 that kind of personal server 01:15:54 I meant publish to github 01:16:06 Elronnd: It's like HTTP/BBS, but if you use it you look awesome 01:16:07 yeah, I get it now 01:16:23 but if I use it no one else can look at it 01:16:26 Because it's more a terminal application than anything else 01:16:28 almost no one 01:16:34 bitbucket doesn't everything github does, but better 01:16:46 even better though: host your own git/hg repo 01:16:54 it's not that hard 01:17:00 diginet: Yes, but I want exposue 01:17:03 *exposure 01:17:08 it's hilarious how github has managed to centralize DVCS... 01:17:16 diginet: DVCS? 01:17:18 lol 01:17:24 distributed version control 01:17:26 Ah 01:17:28 xD 01:17:29 hppavilion[1]: Decentralized Version Control System 01:17:30 system 01:17:52 I cannot tell you how many programmers I've met who don't know what to do when github goes down 01:17:54 Elronnd: Do you think the idea for a terminal-based command line-accessed server sounds cool? 01:17:54 the irony is lost on them :/ 01:18:02 What's centralized about it? 01:18:06 server of what? 01:18:14 Things like bug tracking are centralized, but those are centralized with other systems too. 01:18:27 diginet: Just a little server that spits out user-created content 01:18:39 I do not use GitHub for my own projects 01:18:44 so...some kind of social network? 01:18:50 diginet: Paired with a Qt-based pseudoterminal client 01:18:53 -!- tromp_ has joined. 01:19:03 hppavilion[1]: why not just use a regular terminal? 01:19:05 diginet: Sort of, but it's not a website, so you look cool to everyone else when you use it 01:19:06 that seems like a lot of work 01:19:10 Can you connect to it with a real telnet client? 01:19:14 diginet: Because I like doing GUI 01:19:20 zzo38: Probably not yet xD 01:19:28 zzo38: It's my own protocol 01:19:28 ...but you're making a GUI to clone a CL UI? 01:19:31 Such thing should be added on then 01:19:35 that seems redundant 01:19:37 zzo38: Perhaps I will 01:19:50 diginet: The GUI allows me to add formatting and images down the line 01:19:51 I host my own hg repo on my VPS that I was already paying for, so in effect I have a "private" repo for free 01:20:12 hppavilion[1]: plz tell me this isn't a clone of that godawful termkit thing 01:20:21 diginet: What's termkit? 01:20:36 So no, I guess. If I don't know what it is, I can't clone it xD 01:20:40 hppavilion[1]: a really stupid project, the creator got all buttmad when people didn't fawn over it 01:20:46 Even xterm supports colors and pictures and so on though 01:20:59 diginet: I don't expect people to use it, but it'd be cool if somebody did 01:21:14 diginet: Part of the point is that you can easily engineer your own client or server from scratch 01:21:24 hppavilion[1]: basically imagine a terminal which inexplicably uses Chrome for rendering thumbnails and stuff...it's like terminal but with web crap 01:21:26 or whatever 01:21:28 it was stupid 01:21:29 It literally uses string.split() for lexing AND parsing commands 01:21:30 and immoral 01:21:40 diginet: Ah, yes, that's stupid. And immoral. 01:21:40 that sounds nasty... 01:21:50 diginet: It is, but it works for now 01:21:51 what language is this? 01:21:59 diginet: I sued python 01:22:01 *used 01:22:08 I would sue python if I could 01:22:10 for being awful 01:22:32 (I actually typoed "used" to "sued", realized my mistake, fixed it, then realized it was funny and unfixed it) 01:22:36 diginet: How so? 01:22:40 hppavilion[1]: personal bias 01:22:43 Ah 01:22:53 diginet: Do you at all like my idea? 01:23:01 I don't know enough to comment 01:23:05 OK 01:23:07 hppavilion[1]: actually, my main issue with python is that it is too bloated 01:23:12 I myself happen to prefer JavaScript over Python too though 01:23:13 I can send you a session example 01:23:23 zzo38: JS is far far far far worse 01:23:30 there's not really a clear ly defined "core" of the language 01:23:35 porting it is an utter nightmare 01:23:50 diginet: http://pastebin.com/xRnvGfJk is an example of a fairly boring session 01:24:13 new idea: drop python and Qt and whatever else and use FreePascal 01:24:28 hppavilion[1]: so it's like a BBS? 01:24:29 diginet: I'm planning to expand the client/server to allow powerful things like Dwarf Fortress-style game graphics 01:24:39 diginet: Yes, it is, I mentioned that 01:24:54 ahh, I didn't read the backlog 01:24:57 diginet: I'm trying to modernize the concept of BBS in a way. 01:25:03 I see 01:25:10 well, that could be interesting I suppose 01:25:24 ".porn" 01:25:26 diginet: Also, if someone sees you using it you look REALLY awesome from a non-programmer's perspective 01:25:26 wat 01:25:32 diginet: Yes. .porn file. 01:25:44 diginet: Remember, it's part of the internet, so of course there's porn. 01:25:45 TIL porn uses special container formats 01:25:51 Yep xD 01:25:54 hppavilion[1]: so make it not internet 01:26:00 no pr0n allowed 01:26:13 diginet: How do you not make something internet while still using sockets? 01:26:33 hppavilion[1]: use twine and tin cans 01:26:34 diginet: Correct answer: You don't, sockets use the internet 01:26:51 porn violates the categorical imperative 01:26:59 diginet: The wat? 01:27:09 You could use a local network I suppose 01:27:14 and wait..how are you using complex numbers to refer to users? 01:27:18 zzo38: intranet? 01:27:21 Therefore it can use socket without needing internet 01:27:26 you realized complex numbers only have partial order right? 01:27:29 diginet: That was a joke, though a client could do that if they liked 01:27:39 hppavilion[1]: the categorical imperative. Do you even Kant? 01:27:43 diginet: It's largely a thing implemented by the server 01:28:02 diginet: I don't Kant. I Kant. 01:28:04 diginet: Do you know ES6? 01:28:11 http://cow.org/csi/ 01:28:25 zzo38: I'm familiar with it, it's just as shit as ES1-5 01:28:52 hppavilion[1]: useful 01:28:59 hppavilion[1]: do you even Deontology? 01:29:00 Well I think it is good, except, it doesn't have "goto", and "!yield*" is not acceptable 01:29:21 JS is like a braindead Lua, IMO 01:29:28 diginet: Thank you! 01:29:36 However a few ES6 feature are not supported in Node.js yet I think 01:29:37 diginet: Lua: Glorified configuration 01:29:51 Node JS is cancer 01:30:09 hppavilion[1]: well considering that was one of its original purposes...I odn't see how that's bad 01:30:23 diginet: Part of the server is that much calculation is done by the server; the only real required feature is that it accepts text input and that "METHODS" gives you a list of methods 01:30:31 Object Pascal is the one of the best languages ever designed 01:30:45 And the text part is really optional 01:31:40 this sounds over engineered 01:31:45 I like C programming too though, I also make programming in C and also 6502 assembly language and BASIC and a few others 01:31:47 diginet: All the server needs to do is accept information from the socket and return `HEAD:body` information 01:31:56 zzo38: C is awful 01:32:00 And if you've done that, then you have a server 01:32:11 The client then connects to it and sends user input 01:32:32 it's one of the worst languages ever written. If you measure badness by costs wrt loss productivity due to hacking or loss assets, it may be the single most desctructive language ever 01:32:41 why the hell are buffer overflows still aproblem in 2016? 01:32:48 `%conn [port]` connects to the server in my client 01:32:52 /home/hackbot/hackbot.hg/multibot_cmds/lib/limits: line 5: exec: %conn: not found 01:32:54 Then any input you sent the server receives 01:32:58 Sorry, HackEgo 01:33:21 (I think we should start apologizing to HackEgo whenever we accidentally send an invalid command) 01:33:23 diginet: Because of bad programmers I expect 01:33:52 zzo38: even good programmers fall victim to oversites, esp in million+ SLOC projects 01:34:04 that's why you'd have to be braindead not to use a memory safe and type safe language 01:34:20 diginet: Rust? 01:34:25 I think one of my favorite feature of JavaScript is that you can make "function-oriented programming" with it. 01:34:32 hppavilion[1]: rust is a kitchen sink language 01:34:48 diginet: All software sucks, all hardware sucks 01:34:53 This is a 6502 code: http://wiki.nesdev.com/w/index.php/User:Zzo38/Famicom_Z-machine 01:34:57 it's a convoluted mess that adds no conceivable benefit over languages like object pascal 01:35:00 hppavilion[1]: no, not really 01:35:14 zzo38: how is that specific to JS though? 01:35:15 diginet: I would say yes 01:35:26 it's equally true of Lua...and Python...and Perl...and Lisp...and countless others 01:35:47 diginet: It isn't specific to JS of course 01:36:01 zzo38: nor is it esp well odne in JS 01:36:09 theory: functional programming is the OOP of the 2010s 01:36:17 diginet: Haskell. Haskell is nothing but functional programming. 01:36:18 Yes you are right, it is better done in such thing as Haskell 01:36:29 Though closures are useful 01:36:36 I've used them before in serious code 01:36:36 we see the same ridiculous overpromies we saw in the 90s with OOP 01:36:48 "bug free code, OOM less lines" etc etc 01:36:49 Granted, that serious code was to implement a CL interpreter, but... 01:37:11 and yet, Cabal is one of the buggiest pieces of software I've ever used 01:37:14 but w/e 01:37:21 Do you like assembly language programming? 01:37:31 diginet: What features would you like to see in my server, OOC? 01:37:40 What would be a cool thing? 01:37:50 zzo38: that's too broad a category, which asm? 01:37:50 Perhaps a built-in language? xD 01:37:55 zzo38: VAX...sure 01:37:56 diginet: ELK ASM 01:38:00 zzo38: x86, not so much 01:38:05 hppavilion[1]: what's ELK? 01:38:12 hppavilion[1]: OOC? 01:38:29 diginet: It's a VM/ASM/Infrastructure I'm making as my own .NET-like system 01:38:40 "Out Of Curiosity" 01:38:43 ahh 01:38:46 Modern x86 is especially terrible, the old one was not quite as bad. 01:39:03 What's so bad about x86? in particular? 01:39:03 However, the assembly language programming I know best is 6502 assembly language programming, which I sometimes use. 01:39:04 pre-286 was tolerable at best 01:39:13 I'm not saying it's good, but I don't know much x86 01:39:13 diginet: Yes, that is what I meant. 01:39:15 hppavilion[1]: it's a nightmare 01:39:21 diginet: Such as? 01:39:47 In my opinion 6502 assembly language is good 01:39:56 zzo38: Perhaps you should port 6502 to modern machines xD 01:39:57 hppavilion[1]: you have to memorize all kinds of weird, implicit rules about which instructions use which registers implictly 01:39:58 Somehow 01:40:01 paucity of registers 01:40:09 diginet: Ah, yes, that sounds pretty shit 01:40:10 weird, non-orthogonal addressing mode restrictions 01:40:24 needlessly complicated booting procedure 01:40:41 IMHO, there should be instructions to do things on ANY register, and the assembler should simplify that to the builtins if applicable 01:40:52 hppavilion[1]: interesting, I've been developinng my own VAXoid ISA 01:41:00 hppavilion[1]: congratualtions, you just invented RISC 01:41:15 along with a compiler for a Pascal/Oberon like language 01:41:28 porn violates the categorical imperative <-- of course not. you just have to stream all your own sex too hth 01:41:41 So if ZA1 (set Auxiliary 1 to 0) exists, and you call SET %A1 0, it'll simplify to the former 01:41:50 THe 6502 is ported to modern machines, there are several emulators! 01:41:58 oerjan: it violated the second formulation, that you never treat another human as a means to an end 01:42:06 *violates 01:42:08 zzo38: Yes, but are there any that compile to x86? 01:42:21 I think I have read of such a thing once 01:42:25 hppavilion[1]: not worth it 01:42:29 Not an emulator, mind you, but a compiler from 6502 to x86 01:42:40 hppavilion[1]: this has been done from x86 to ARM 01:42:44 I figured something out yesterday 01:42:49 hppavilion[1]: actually, wait 01:43:07 hppavilion[1]: enjoy: http://andrewkelley.me/post/jamulator.html 01:43:20 too bad it uses piece of shit language like Go 01:43:36 -!- jaboja64 has quit (Ping timeout: 250 seconds). 01:43:43 You can model data moving instructions as ordered pairs where s >= 1 and d >= 0, where s is the length of the reference chain to the source and d is the reference chain to the destination 01:44:16 So <1, 0> is the SET instruction, <1, 1> is MOV, <2, 1> is IMOVL, etc. 01:44:17 I have designed a VM too, called QUACKVM and have written a few programs in it (a minesweeper game and a robot find kitten game are some) 01:44:37 zzo38: What makes for an optimal VM? 01:44:39 zzo38: what kind of ISA? 01:44:44 stack or register? 01:44:53 diginet: I know that his had 32 instructions, which was impressive to me 01:45:02 what did? 01:45:10 diginet: QUACKVM 01:45:16 diginet: I've also been attempting to design a VM based on digraphs 01:45:23 ..huh? 01:45:29 a graph reduction machine you mean? 01:45:31 diginet: zzo38's VM 01:45:35 oh 01:45:51 QUACKVM has memory-mapped stack and registers 01:46:01 diginet: No, like an actual machine code that could- in theory, given completely different models- be run on bare metal 01:46:34 I...don't kinow what you mean bu that 01:46:44 The program counter is memory address zero, address one is the default stack pointer, and address two is a "quick access" register, instructions that take operands from cell two can use a shorter encoding than those at later addresses 01:47:51 zzo38: Is JMP not a builtin then? 01:48:13 Yes that is correct, JMP is not a builtin (although the assembler provides a macro) 01:48:22 diginet: I mean that it's a VM that uses a digraph w/ pointer instead of registers 01:48:25 (RETURN is not a builtin either) 01:48:36 zzo38: It's not a builtin in ELK either 01:49:01 zzo38: Just a macro for POPTO %1 \ CJMP %1 01:49:27 RETURN that is, not JMP 01:49:31 JMP is builtin (for now) 01:49:56 In QUACKVM, RETURN is a macro for "PUT ,,STACK" 01:49:58 zzo38: Though ELK has nearing on 128 instructions because I wanted floats to be builtin. 01:50:03 zzo38: Yes, you mentioned that 01:50:10 I don't really understand it xD 01:50:12 (Which encodes into a single memory cell) 01:50:23 Wait, I think I get it now 01:50:32 , separates instructions, right, right 01:50:52 No it does not separate instructions 01:50:59 I mean arguments 01:51:00 Sorry 01:51:01 Yes 01:51:05 That is correct 01:51:17 I meant arguments the first time 01:51:28 OK 01:52:14 `CALL line` in ELK is a macro for `PUSH \ JMP line` 01:52:15 /home/hackbot/hackbot.hg/multibot_cmds/lib/limits: line 5: exec: CALL: not found 01:52:22 Sorry again, HackEgo 01:52:31 And CCALL is t a macro for the same, but with CJMP 01:53:58 diginet: THings get weird in that notation if you allow negatives 01:55:24 QUACKVM does have CALL as a builtin, although all instructions use the same encoding (the instruction set is orthogonal), so a computed call is the same CALL instruction. 01:55:56 zzo38: You can do CALL as a builtin in QUACKVM if QUACKVM happens to have MOV, IIAC 01:56:24 Wait, no, that's CCALL 01:57:24 zzo38: At what point does a VM cease to qualify as a RISC? 01:57:34 I've read RISC doesn't imply less instructions that CISC 01:57:51 In fact, RISCs tend to have /more/ opcodes than CISCs 02:00:18 -!- AlexR42 has joined. 02:02:25 I don't know how it works 02:02:36 It seems to me people decide it arbitrarily 02:04:36 zzo38: Ah. 02:04:43 oerjan: it violated the second formulation, that you never treat another human as a means to an end <-- according to wikipedia, you're missing an important "merely" there hth 02:07:35 that still applies 02:07:42 porn clearly violates the second formulation 02:07:50 as told to me by an actual Kantian 02:07:52 so.. 02:09:05 so does buying chocolate and now i should stop. 02:11:19 -!- Hoolootwo has left ("Leaving"). 02:11:22 *stop trolling. 02:11:28 -!- Hoolootwo has joined. 02:16:52 zzo38: Perhaps I should just have conditional SKIP instructions and make JMP and its conditional variants macros... 02:18:12 In QUACKVM, all instructions are conditional skip instructions. 02:18:38 So `JMP.Z condreg line` macros to `IFZ condreg %cond \ SKIPIF %cond \ SET %pc line` 02:19:04 zzo38: Interesting... 02:35:27 zzo38: OK, there are now 0x60 instructions in ELK 02:35:48 And 0x23 macros 02:36:06 Wait, 0x60 isn't counting SKIPIF 02:37:05 0x61 03:05:40 -!- Phantom_Hoover has quit (Read error: Connection reset by peer). 03:08:25 -!- AlexR42 has quit (Quit: My Mac has gone to sleep. ZZZzzz…). 03:08:35 [wiki] [[Generic 2D Brainfuck]] https://esolangs.org/w/index.php?diff=46372&oldid=43414 * 216.58.126.185 * (+1) "d moves pointer counter down and r does too? I don't think so!" 03:11:19 -!- boily has joined. 03:15:42 <\oren\> RS419236755NL 03:17:37 <\oren\> why is everything i order from china being shipped from rotterdam 03:19:49 \oren\: Could it be because it ships through the Netherlands/ 03:19:51 *? 03:20:21 <\oren\> Tracking number 03:20:36 <\oren\> RS375708670NL 03:20:44 <\oren\> that one too 03:20:57 he\\orenederlands\. 03:21:03 @massages-loud 03:21:03 hppavilion[1] said 2h 20m 37s ago: I didn't set the "misleading topics" thing 03:21:03 hppavilion[1] said 2h 20m 24s ago: I just added "esoteric" for consistency 03:21:22 hppavilion[1]: hppavellon[1]. I was mislead. tdh. 03:22:59 <\oren\> oh, my other package is actually in china 03:23:20 <\oren\> the tracking status is 【广州互换局】已出口直封 which seems legit 03:24:11 <\oren\> RI853168115CN 03:24:33 <\oren\> oh, i see 广州 = Guangzhou 03:25:01 -!- boily has quit (Quit: TAPPED CHICKEN). 03:44:22 -!- lleu has quit (Quit: That's what she said). 03:52:02 -!- Elronnd has changed nick to fnurgle. 03:52:05 -!- fnurgle has changed nick to fnurggle. 03:52:10 -!- fnurggle has changed nick to Elronnd. 04:16:07 -!- mad has joined. 04:16:27 https://www.youtube.com/watch?v=tK50z_gUpZI <- guy trash talks garbage collection 04:19:38 Hi, mad! 04:19:59 I'm making a VM called ELK that is meant to be an alternative to CLR! I'm a bit of an idiot! 04:22:18 oh? 04:23:15 mad: It's made to work like the common language runtime, but not be microsoft 04:23:22 Literally, my design principals are: 04:23:34 * Have the cool interaction-between-language feature of the CLR 04:23:48 * Note to self: Don't be Microsoft 04:25:00 <\oren\> AAAAAAAAAAAA 04:26:17 <\oren\> mad: the problem is these people reek of fake expertise 04:27:42 \oren\ : what makes you think that 04:28:40 <\oren\> just a general feeling, plus, seriously, if the Ouya is open source, then there is absolutely nothing that can prevent them from releasing and running native executables? 04:29:32 I'm not familiar with the ouya, I guess it did have the native executable fixes that android got 04:29:56 though to be fair I think Java in android was a stupid idea and should never have been done 04:32:26 and they are game programmers so of course they're going to trash talk java 04:33:23 sheesh scott aaronson's blog theme reset again 04:33:39 this is the third time, i think. he needs a more permanent fix to that bug. 04:34:14 (which is technically a "we don't support your php any more" type of bug iirc) 04:34:17 \oren\ : also, if the ouya store only takes java games, you're screwed 04:35:10 <\oren\> bah, some of the most memorable games I played as a kid were written in ActionScript 04:35:42 and didn't run at 60fps :D 04:35:46 <\oren\> yeah 04:36:50 funny, some of the most memorable games I played as a kid were written in 65816 assembly 04:36:57 and ran at 60fps 04:37:19 on a 3mhz system with barely any ram to talk of 04:39:34 *any more, so we sometimes delete all of it 05:09:58 -!- variable has quit (Quit: 1 found in /dev/zero). 05:18:33 I can program in 6502 assembly, not 65816 though 05:21:10 One of my pet projects is looking at the 6502 to see if you could 32bit-ify modernize it and stuff 05:23:05 mad: by 60fps do you mean 60fpm? 05:23:45 snes games run at 60fps 05:24:14 nobody had an apple2gs which I admit typically runs at 60fpm :D 05:24:52 yay 05:26:51 I find 6502 is OK as it is, although I use the stable unofficial opcodes too (they can be used only on NMOS 6502 or on other implementations of the NMOS 6502 instruction set) 05:29:17 * oerjan isn't sure what is happening in today's girl genius but he doubts it's proper handling of books 05:34:51 zzo38 : anu useful ones? 05:36:27 Do you mean useful unofficial opcodes? They can be useful for different purposes, I have used LAX, ARR, ANC, and others too 05:46:34 mad: Yes, several of them were used fairly often back in the day. 05:48:03 lax loads both lda and ldx at the same time? 05:48:26 mad: Yes. 05:48:34 Yep. 05:49:00 I can see how that could be useful 05:49:06 (However it does not work correctly with the immediate addressing mode, so LAX should be used only with non-immediate addressing modes.) 05:51:59 Here is the explanation: http://wiki.nesdev.com/w/index.php/Programming_with_unofficial_opcodes 05:53:23 One .NSF file I released starts with the instruction "EOR #0" in order to create a patch point, so that the multi-song NSF can be converted into a single-song NSF of any of the songs, without having to recompile the entire program. 05:56:20 However I noticed that the instruction that would be "STA immediate" actually does nothing and does not overwrite the instruction. 05:57:01 (If it did, it would probably be convenient for fast bankswitching) 05:58:28 it's hard to make a faster version of the 6502 is what I mean 06:01:08 <\oren\> why? 06:01:29 <\oren\> can't you just make the chip smaller and the clock faster? 06:01:57 yes but you're still reading opcodes directly from DRAM 06:02:07 with a full access cycle every time 06:04:29 like, if you put it in competition with a MIPS or an ARM it will lose badly 06:05:10 because they use their memory cycles much more efficiently 06:06:45 Still I do not like the automatic reordering and automatic stalling and automatic caching and so on of many newer instruction sets, and I believe they should be done explicitly, although also some instruction might just not return the result right away (such as multiplication register for example) and you can do other stuff in between 06:09:25 -!- XorSwap has joined. 06:10:15 actually what I like about the idea of a modernized 6502 is the potential for new ways to do automatic reordering and automatic stalling and automatic caching 06:13:36 provided that you take out the stuff that tends to create stalls and doesn't help throughput too much 06:19:07 In my ELK ASM, I was at first going to make conditional JMPs macros (using SKIPIF), but I was told that it's better to make the important ones builtin for efficiency 06:19:54 mad: zzo38: everyone else: We should team up as the Council of Eso and construct the world's best ISA 06:20:06 I have some ideas 06:20:11 mad: Such as? 06:20:26 ok you know how a RISC is like 06:20:31 mad: Also, ICYC, ELK is documented at https://docs.google.com/document/d/1ZFHwxE0AqpW9nRqTjLMPKZ68RUubLsYGRozKAwTI68s/edit# 06:20:33 Yes 06:20:35 add r0, r4, r5 06:20:44 (add r4 to r5 and store in r0) 06:20:46 Yeah, that's pretty much ELK 06:21:05 Where instructions take one (or at most 2, but rarely) CPU cycles to complete 06:21:14 ok I'd look into doing an architecture where instead you go 06:21:47 Not necessarily less instructions, but less complicated instructoins 06:21:50 *instructions 06:22:04 actually I'm looking into "less retirement" 06:22:21 so instructions that don't write to any register / flag / memory are GOOD 06:22:26 Ah 06:22:32 because you have one less register to rename 06:22:40 one less write port to the register file 06:23:05 the problem with a sequence, like 06:23:14 mad: So the instructions don't have to have a target? 06:23:17 add r4, r4, r5 06:23:21 mad: I am confus 06:23:25 shr r4, r4, 16 06:23:29 shr? 06:23:33 shift right 06:23:34 Don't know what shr does 06:23:38 Ah! 06:23:40 I prefer RSH 06:23:48