Hide Idle (>14 d.) Chans


← 2022-03-26 | 2022-03-28 →
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2022-03-26#1089375 << irecall this lul from last (?) yr, and still dun get, with what is the victim supposed to pay a tax on ~unrealized~ 'gain' ?
dulapbot: Logged on 2022-03-26 22:49:22 mats: https://archive.ph/niT1F "President Joe Biden will propose a minimum 20% tax rate that would hit both the income and unrealized capital gains of U.S. households worth more than $100 million as part of his budget proposal to be released on Monday."
asciilifeform: ... if it's with money -- after coupla iterations, he'll have none; if with 'asset', and expected to sell it -- ditto, and nonlinearly, most paperola is only worth ~any when ~nobody sells
asciilifeform: seems moar humane to shoot & dekulakize a la 1917.
dulapbot: Logged on 2022-03-26 22:51:40 verisimilitude: It's nice to see most white people worth more than $100 million thanks to his policies.
asciilifeform: ... via the obv. path.
dulapbot: Logged on 2022-03-25 16:05:41 asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2022-03-25#1088692 << can't speak for other folx, but in asciilifeform's corner of the barbed wire, 'sandwich' seems set to win the race for 'new zero', well ahead of btc-usd exchrt
signpost: the lul is going to be generating unrealized losses and scraping gambling money *out* of the government with this.
signpost: they'll likely pay with loans, if this passes.
verisimilitude: Apparently, Putin has compared criticism of Russians to criticism of JK Rowling.
signpost: we're going to be nuked by a harry potter fan, verisimilitude.
mats: jk rowling is great
mats: lost count of the number of dipshits i used to know, that have instagrams picturing themselves in culturally appropriate dress and black children during a month long fakework trip
mats: religious people are the worst
signpost at a certain point trashed all social media accts populated with folks he grew up with.
signpost: ended up a nice filter, plenty of high quality folks remained.
signpost: but yeah, totally know what you're talking about, posing on a mission trip, etc
verisimilitude: I've a joke for this occasion.
verisimilitude: What is a similarity between major religions and major corporations?
verisimilitude: After they saturate domestically, they begin making massive appeals to foreigners.
mats: no
mats: just stop
signpost: that's just a statement of fact, nothing unexpected, nor unusal path to the expected.
signpost: *unusual
verisimilitude: Some of the best jokes are offensive statements of fact.
mangol: god has a sense of humor
mangol: i've been working on a plan to scrape good stuff from the web, process it into more convenient formats, and store it for offline archival / hash based sharing
mangol: essentially a graph where the initial node is a program that downloads from www and spits out static files
mangol: those files are ideally immutable (the initial program can take ephemeral www content, e.g. rss or irc, and chop it into daily/weekly/monthly static files)
mangol: the immutable files are then fed to other programs as desired, to produce derived files
mangol: immutable files and derived files are hashed to ensure integrity / detect changes. programs can be hashed
verisimilitude: archive.today
mangol: i'm gathering a simple subset of scheme so the filters can run in a portable, reproducible environment
mangol: typically scraping is done with no attention paid to reproducibility
mangol: archive.today == deplorable emulation of archive.org?
mangol: btw just downloaded uncle al 's site from the latter, on recommendation of asciilifeform
verisimilitude: Alright.
verisimilitude: Would my website be scraped so, mangol?
mangol: if you want
verisimilitude: Know that's foolish.
verisimilitude: My Gopher hole is much better for this purpose.
mangol: gopher fine as well
mangol: i'm preoccupied with making a simple scheme-based framework for reproducible filters
verisimilitude: Alright.
mangol: the more filters that can be plugged into it, the better
mangol: the idea is to have a clean separation between teh unruly internets and the clean, reproducible filter world
mangol: i have a logotron scraper done
mangol: i.e. download http://logs.nosuchlabs.com/static/log_db.gz and turn into daily tab-separated-values files
mangol: and a filter to turn the daily files into json
mangol: (i have a draft spec for representing irc logs as json, can post for review soon)
mangol: archive.today seems to have "download .zip" link for each page, but the ones i tried are not currently working
mangol: to get stuff from archive.org, https://github.com/hartator/wayback-machine-downloader is wonderful
mangol: it's a ruby script
mangol: i noticed you all wrote irssi, znc, etc. scrapers for logotron. i can adapt some of those to scheme
mangol: should probably write a scheme V as well
mangol: asciilifeform: if you're not allergic to JSON and the above IRC JSON spec, a URL to download last full day's logs (or last 30 full days, etc.) would be great
mangol: or if you _are_ allergic, something like tab-separated values or comma-separated values will do as well (though the latter is _more_ complex than JSON)
mangol: when i downloaded http://logs.nosuchlabs.com/static/log_db.gz it took like half an hour or something
mangol: ^ that's the output of running `wayback_machine_downloader https://www.mazepath.com/uncleal/`
mangol: almost 1000 html files, he was quite prolific
verisimilitude: JSON sucks.
mangol: i have a plan for civilizing html markup too, but will explain later. ties easiy into filter scheme plan.
mangol: *easiyl
mangol: *easily
mangol: verisimilitude: why?
verisimilitude: It's a retarded form of S-expressions.
verisimilitude: It's perfect for the WWW, however.
verisimilitude: Some stupid asshole takes some syntax from a shitty programming language made over a weekend and starts fellating himself for inventing a new data format.
mangol: there are few bigger fans of S-expressions than i am, and i hold JSON in high regard
mangol: one of the few standards done right
verisimilitude: Store a large number with JSON.
mangol: a large integer or fraction, i presume
mangol: how does that differ from doing same with s-expressions?
verisimilitude: It won't work with JSON.
verisimilitude: It's clear now that, between us, only I know of this issue.
mangol: conventionally people explain themselves when that's the case
verisimilitude: The JavaScript numerical nonsense interferes.
mangol: JSON is not tied to JavaScript. JSON fractions don't even have to be parsed into floats
verisimilitude: Sure, but many implementations will mangle numbers.
mangol: so will many lisp implementations. check github for over9k flavors of the month
mangol: conflating languages (or formats) with their implementations is useful if you gotta ship a product asap
mangol: but unsound in principle
verisimilitude: It's harder to parse than S-expressions.
verisimilitude: JSON has no symbol type.
verisimilitude: JSON is far uglier.
verisimilitude: JSON doesn't encode programs effectively.
mangol: (fwiw my IRC JSON spec doesn't use the JSON number type - only strings, "arrays", "objects")
mangol: depends on what you wanna do
mangol: JSON isn't human writable. IRC logs don't need to be human written
mangol: the reason i chose JSON instead of sexp for this project is 1) it has a standard way to encode "dictionaries", aka "maps" or "objects"; 2) https://modern.ircdocs.horse/ - a great irc spec site - is already using it; it may be used in future version of irc protocol; 3) over9k ready made libraries; 4) most importantly, can hope to get other people to use it, so in the future i could download some logs
mangol: that others readily provide as JSON, don't have to write own scraper for everything
mangol: as a spec, JSON has one big caveat, and it's that it assumes unicode
crtdaydreams trying to get mp-wp up on freebsd
mangol: one of the most desirable items for above outlined "filter scheme" contraption would be a wordpress extractor
crtdaydreams: I'm also using nginx+FPM as opposed to apache while having to source install php56 (EoL). Should be ``fun''.
mangol: does mp-wp use the same HTML structure to the extent that poasts and comments can be found via standard wp HTML ID attributes
crtdaydreams: If anyone has already done so (x doubt) and has a guide. Would def like. Otherwise might end up hacking sauce and writing own guide.
crtdaydreams: **or** if someone has downdated mp-wp to an older version of php i.e. php70.
crtdaydreams: s/older/shittier/
mangol: does #a use old versions of php?
crtdaydreams: not sure. current ver. I'm working off for mp-wp as per billymg's guide is <=php56
dulapbot: Logged on 2022-03-27 05:13:32 crtdaydreams: does mp-wp run with anything other than mysql?
crtdaydreams has yet to read compress mp-wp genesis
crtdaydreams: **and associated Vtree
mangol: verisimilitude: `lynx gopher://verisimilitudes.net` finds your site
dulapbot: Logged on 2022-03-26 18:57:30 verisimilitude: I've seen a few retards lately arguing that orthogonal persistance is bad, because then they can restart the machine to ``fix'' an error.
mangol: HN started to be unreadable sometime during the last decade
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2022-03-27#1089385 << what reason is there to think it'll work in reverse gear ?
dulapbot: Logged on 2022-03-27 00:52:32 signpost: the lul is going to be generating unrealized losses and scraping gambling money *out* of the government with this.
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2022-03-27#1089443 << this is nifty. here's a warezball of his old Official floppy dist. for completeness (could've sworn posted yrs ago, but couldn't turn up in log)
dulapbot: Logged on 2022-03-27 03:45:52 mangol: https://misc.lassi.io/2022/uncleal/
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2022-03-27#1089440 << there's a raw-log knob, which yields up to 500 ln, e.g.
dulapbot: Logged on 2022-03-27 03:16:01 mangol: or if you _are_ allergic, something like tab-separated values or comma-separated values will do as well (though the latter is _more_ complex than JSON)
mangol: talk radio as text files on a floppy :D interesting experience
asciilifeform: mangol: talk radio ?
mangol: in spirit
asciilifeform: mangol: 1st time ever heard uncleal described as 'radio in spirit' lol
mangol: maybe something doesn't translate over to this side of the pond but i get the american conservative talk radio vibe from his floppy & site. the rush limbaugh type. not normal?
mangol: raunchy and unabashed, mixing high brow and low brow
billymg: crtdaydreams: you won't have much luck with php7 out of the box but you could in theory patch it and create a fork that's php7 compatible. easier imo to build php5.6 from source or find an old .deb build file for it (that's what i do for gentoo, never used debian's package manager)
billymg: http://logs.bitdash.io/asciilifeform/2022-03-27#1089484 << for comments, yes, standard across themes as it calls wp_list_comments() inside wp-includes/comment-template.php to generate the html
bitbot: Logged on 2022-03-27 09:28:20 mangol: does mp-wp use the same HTML structure to the extent that poasts and comments can be found via standard wp HTML ID attributes
billymg: post content might vary slightly across themes if you're relying on ids or classnames to find the content
mangol: billymg: thanks. is the variation meaningful or just a historical accident?
billymg: actually, i take that back, the class "hentry" is defined in wp-includes/post-template.php, so that would be a theme-agnostic way of finding the post content
mangol: cool
bitbot: Logged on 2022-03-24 15:13:47 billymg: http://logs.bitdash.io/asciilifeform/2022-03-24#1087693 << "The trouble was that none of the 9 were properly documented and none were bug free. Basically each person had implemented his own solution and it worked for him so that was fine. This is a BBM attitude; it works for me and I understand it. It is also the product of not needing or wanting anybody else's help to do something."
bitbot: Logged on 2022-03-24 15:16:12 billymg: imo this is something we could work on
mangol: all of those plans will take a long time tho :-/
billymg: imo the one nice thing about having a fuhrer (back in the mp era) was someone to say "we're doing x and using y to do so, get to work" vs. 100 splintered efforts. however, great collaboration on e.g. pest without fuhrer
billymg: i have a feeling there will be a natural pull back to joint efforts. at the moment is like video of vase dropping to floor and flying into a thousand pieces, soon tape will stop and be played in reverse with all the pieces reassembling back into one. i think we're almost at that point
billymg: or like a tree recursive algo, each of has had to go and discover our own answers first before being able to come back with something useful
asciilifeform: billymg: pest is necessarily a joint effort
asciilifeform: rather difficult to test properly w/out multiple players
billymg: asciilifeform: yeah that's why i mentioned as good example of collaboration sin fuhrer
asciilifeform had hoped folx would offer vpatches to spec, esp. given as asciilifeform has vehehery limited cycles to work on subj. but even as it is, moving along
asciilifeform sadly hasn't the resources to do much 'fuhrering'
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2022-03-27#1089508 << fella born in '50s, watched the opening shots of usa collapse process in '90s. hence the flavour.
dulapbot: Logged on 2022-03-27 11:30:07 mangol: maybe something doesn't translate over to this side of the pond but i get the american conservative talk radio vibe from his floppy & site. the rush limbaugh type. not normal?
asciilifeform: $ticker btc usd
busybot: Current BTC price in USD: $44900.67
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2022-03-27#1089520 << plans w/out resources tend to take, lol, ~inf. time.
dulapbot: Logged on 2022-03-27 11:48:05 mangol: all of those plans will take a long time tho :-/
asciilifeform: many things worth doing don't yield all that well to 'nights & weekends' approach
asciilifeform: (and not even speaking of irons projects which require serious dough)
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2022-03-27#1089429 << much better imho to stand up yer own copy of the logotron which gives yer preferred format ( atm not clear to asciilifeform for what you need json. ) or use the existing knob.
dulapbot: Logged on 2022-03-27 03:02:37 mangol: i.e. download http://logs.nosuchlabs.com/static/log_db.gz and turn into daily tab-separated-values files
dulapbot: Logged on 2022-03-27 11:13:43 asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2022-03-27#1089440 << there's a raw-log knob, which yields up to 500 ln, e.g.
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2022-03-27#1089476 << asciilifeform aint much concerned w/ 'future versions of irc' as has 0 plans to use'em for anyffin. in fact plans to retire the dulapnet irc at some pt when pest 100% mature
dulapbot: Logged on 2022-03-27 04:22:20 mangol: the reason i chose JSON instead of sexp for this project is 1) it has a standard way to encode "dictionaries", aka "maps" or "objects"; 2) https://modern.ircdocs.horse/ - a great irc spec site - is already using it; it may be used in future version of irc protocol; 3) over9k ready made libraries; 4) most importantly, can hope to get other people to use it, so in the future i could download some logs
asciilifeform: and fwiw on a pestnet erry participant station has complete logs 'for phree'
mangol: http://logs.bitdash.io/asciilifeform/2022-03-27#1089537 << i'll keep archiving logs for chans that will stay on irc indefinitely - the lisp ones, maybe others
bitbot: Logged on 2022-03-27 16:58:19 asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2022-03-27#1089429 << much better imho to stand up yer own copy of the logotron which gives yer preferred format ( atm not clear to asciilifeform for what you need json. ) or use the existing knob.
dulapbot: Logged on 2022-03-27 03:02:37 mangol: i.e. download http://logs.nosuchlabs.com/static/log_db.gz and turn into daily tab-separated-values files
dulapbot: Logged on 2022-03-27 11:13:43 asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2022-03-27#1089440 << there's a raw-log knob, which yields up to 500 ln, e.g.
mangol: wasn't aware of knob, will use for now, thanks!
mangol: log_db.gz isn't a standard feature of logotron?
mangol has a general interest in schemifying scripts, so could look into a irc/pest bot at some point but can't promise a timeline
asciilifeform: mangol: nope, is cron job
mangol: http://logs.bitdash.io/asciilifeform/2022-03-27#1089533 << the worst thing is that resources don't always help. i've thought about getting lisp back on track a lot, and came to the conclusion that no amount of money will do it in <5yrs without botching up the technical apects
bitbot: Logged on 2022-03-27 16:03:00 asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2022-03-27#1089520 << plans w/out resources tend to take, lol, ~inf. time.
dulapbot: Logged on 2022-03-27 11:48:05 mangol: all of those plans will take a long time tho :-/
signpost: mangol: while you're archiving things, this will give you all the pubkeys deedbot knows. http://wot.deedbot.org/keys.asc
mangol: good design is just slow to do. i produce new stuff weekly, but it's a snail's pace.
mangol: can parallelize somewhat, but can't produce good insights faster - i'm at max speed
mangol: all the enthusiastic projects to do a lot of lisp work fast (racket being latest iteration) end up with impressively big stuff where individual technical details are high quality, but the contraption doesn't really fit together as a whole
mangol: clojure was similar. first-rate people, great work, but at that pace you cut corners re: "does system make sense as a whole?"
mangol: this is like some higher-level, grown-up version of CADT model where the coders are genuinely wizardly, but process has the same fundamental flaw
mangol: "let's throw away CL and make clojure" -> "let's throw away clojure and make racket" -> (to be continued)
mangol: Perlis: "Around computers it is difficult to find the correct unit of time to measure progress. Some cathedrals took a century to complete. Can you imagine the grandeur and scope of a program that would take as long?"
mangol: no cathedrals, only a succession of CADT-edrals
mangol: http://logs.bitdash.io/asciilifeform/2022-03-27#1089554 << thanks! that's a lot of keys! you can decode these to get cleartext email addresses, right?
bitbot: Logged on 2022-03-27 17:41:06 signpost: mangol: while you're archiving things, this will give you all the pubkeys deedbot knows. http://wot.deedbot.org/keys.asc
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2022-03-27#1089551 << to be concrete, resources help in the elementary sense where somebody can actually keep project loaded into his head fulltime. beyond that -- not particularly
dulapbot: Logged on 2022-03-27 13:39:20 mangol: http://logs.bitdash.io/asciilifeform/2022-03-27#1089533 << the worst thing is that resources don't always help. i've thought about getting lisp back on track a lot, and came to the conclusion that no amount of money will do it in <5yrs without botching up the technical apects
dulapbot: Logged on 2022-03-27 12:03:35 asciilifeform: many things worth doing don't yield all that well to 'nights & weekends' approach
signpost: mangol: yeah, whatever folks had in the key they imported to deedbot is there.
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2022-03-27#1089558 << asciilifeform was not the least bit impressed w/ subj, either the product or the people ftr
dulapbot: Logged on 2022-03-27 13:43:50 mangol: clojure was similar. first-rate people, great work, but at that pace you cut corners re: "does system make sense as a whole?"
asciilifeform: mangol: to 'get on track', even e.g. asciilifeform + ice40k8 board could prolly do it in ~yr. but 'if wishes were horses'. hasn't a yr, or even a month, of continuous cycles, and aint aboutta
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2022-03-27#1089562 << era of napkin sketches of cathedrals. most of which turn out to be escherian on close look, lol
dulapbot: Logged on 2022-03-27 13:47:47 mangol: no cathedrals, only a succession of CADT-edrals
mangol: asciilifeform: largely agree with your criticism of clojure, but they come from corner-cutting and the standard pitfalls of "idea people" personalities.
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2022-03-27#1089557 << the contraptions fall apart because nobody dared to even try to get the proper foundation built.
dulapbot: Logged on 2022-03-27 13:42:55 mangol: all the enthusiastic projects to do a lot of lisp work fast (racket being latest iteration) end up with impressively big stuff where individual technical details are high quality, but the contraption doesn't really fit together as a whole
mangol: have read that post
mangol: i disagree about where the problem lies. lisp is about language oriented programming. that's not "bottom up" or "top down" - means you can pick any convenient level of abstraction and start there
mangol: you can build up from atoms, but tons of useful stuff can be done simply by assuming the environment is conjured up by magic
asciilifeform: mangol: ... and will end with the same garbage as prev. existed.
dulapbot: Logged on 2022-01-18 13:31:33 asciilifeform: it's sad but troo, well-behaved guism on existing os is in fact over9000x harder than e.g. writing a kernel and compiler.
dulapbot: Logged on 2022-02-24 16:26:41 asciilifeform: warned loong ago that the days of lin. kernel being buildable on sane gcc are numbered
asciilifeform: unless you start at bottom.
asciilifeform: because there is no 'magic'.
asciilifeform: the existing env. is liquishit, and seeps through immediately.
dulapbot: Logged on 2020-07-03 19:27:09 asciilifeform: apparently such proggy can only be written if write 100% of graphics stack (incl. font renderer) from 0. as in e.g. emacs (before anyone asks, i ~specifically~ did not and still do not want this kinda thing as elisp progggy -- it'll 100% guaranteed choke on 200MB+ of indexable db.. )
mangol: ah, gui saga
mangol: i have similarly sordid experiences with the topic
asciilifeform: grim fact, but is the reason erry single attempt at 'serious lisp' on pc went to exactly same place when anyone attempts to do serious i/o (vga, network, etc)
mangol: i suspect you're talking about lisp as in lambda calculus on a computer
mangol: to me, lisp is language oriented programming. source code is lists, you have lisp.
mangol: lambda calculus is incidental. it's generally the best thing since sliced bread, but also easily leads to wankery. there are places i wish people didn't assume it as a model
asciilifeform: to asciilifeform , the sexpr aspect is only '1 part of balanced diet' of bolix-like machine complying with 7laws.
mangol: sure, hard to disagree with your goals. but we'll always be dealing with other kinds of machines too, and software for them
mangol: to me, the draw of language-oriented programming is general applicability
mangol: you can make a DSL that describes only what you need, and port it anywhere
asciilifeform: mangol: is what the bedrock article was about. i.e. can you successfully keep the water outta yer submarine and actually abstract over the physical mechanisms, or do they leak.
mangol: i agree, but use an approach where there's a different bedrock for each thing designed
asciilifeform: 'port it anywhere' is the approach that gave us the extant garbage.
asciilifeform: in practice, ~nuffin is actually portable in the sense claimed.
asciilifeform: e.g. witness the titanic pile of #ifdef crud in ~any nontrivial unix proggy.
mangol: there's a pervasive culture of virtually no attention paid to reusability in the field
mangol: a large part of why is that we use programming languages where it's almost impossible to make a good DSL
asciilifeform: folx use what's avail., a la 'junkyard wars'.
mangol: yep. the attitude is easy to see in action around most programming languages
mangol: the scheme community is very good re: principled design. i've been trying to drag them in the direction of even fewer environmental dependencies
asciilifeform: whole field since mid-'90s sums to various attempts to 'reach the moon by stacking chairs'.
asciilifeform: that it doesn't matter what kind of chairs, remains unclear somehow to the participants.
mangol: to be fair, i don't think anybody knows how to build a software "cathedral" of the perlis kind
asciilifeform knows. and aint the only one
asciilifeform: it isn't any kind of mystery.
verisimilitude: There's a means to port software anywhere, but it requires focussing on the highest level only, and that requires there to be a suitable high level.
asciilifeform: verisimilitude: the suitable ~low~ level is absent.
asciilifeform: or did verisimilitude sleep through the gui thrds etc
verisimilitude: One man can build a cathedral, mangol, because computers are meant to automate.
verisimilitude: The suitable high level requires a suitable low level, yes.
asciilifeform: for that matter, verisimilitude's experience recently w/ the unix ip stack oughta have taught sumthing
mangol: asciilifeform: what method do you suggest?
mangol: or rather, what foundation
asciilifeform: mangol: asciilifeform has entire www re subj...
verisimilitude: Oh trust me, when I release the library, which is effectively finished, it will be amusing.
mangol: i can't figure out how to build a cathedral from the ingredients in loper
asciilifeform: sane iron, a la bolix; on the latter, sane os. adding up to 7laws compliance.
verisimilitude: I'm working on the documentation now.
mangol: the key thing is to come up with a way to build software where complexity is not a problem. nobody has that.
verisimilitude: Before building a cathedral, give the carpenter somewhere nice to live.
asciilifeform: mangol: read the scheme-79 materials ?
mangol: not familiar with scheme-79, but suspect it deals with lambda calculus on a chip
asciilifeform: entire machine, incl. cmos cad & tapout generator, represented by scheme proggy.
mangol: sounds really cool
verisimilitude: I forget, was the garbage collector there in hardware, or represented as a Scheme program?
asciilifeform: this was before peak bolix, too
asciilifeform: verisimilitude: hardware -- represented by scheme proggy (compiled to gatelist on tapeout)
verisimilitude: Alright.
mangol: but you're talking about foundations again
asciilifeform: mangol: correct. asciilifeform is interested in sane foundations, thought this was 100% clear historically
asciilifeform: sane foundations is how you actually build working airplane; as opposed to 'flap arms harder'
mangol: a cathedral is a really big, really long lasting building. software equivalent is big and complex, something like mathematica / wolfram alpha
verisimilitude: Yes, the African Space Program.
mangol: nobody knows what foundations we should use to build systems so big
asciilifeform: mangol: asciilifeform's position, informed by scheme79, bolix, etc. is that almost all of the apparent complexity is illusory
asciilifeform: again with this 'nobody knows'
mangol: asciilifeform: see also alan kay's VPRI which tried (and succeeded to some extent) to minimize size of "desktop OS" implementation
asciilifeform: or e.g. wirth's 'oberon'
mangol: yep. have emulated it, good stuff
mangol: but this is all 80s-90s
asciilifeform: the closest to 'the bottom' these started, the moar interesting from asciilifeform's pov
verisimilitude: It's simple, mangol: Correctly compose correct components.
asciilifeform: mangol: from asciilifeform's pov, absolutely nuffin of any interest happened in system arch. since '90 or so
asciilifeform: ~all sums to attempts to speed up winblows
mangol: or linux, nowadays
mangol: well known that "systems research is dead"
verisimilitude: The murderers keep saying that, yes.
mangol: i still don't believe people know what abstractions to build big, long-lived systems out of
mangol: as evidenced by myriad of incurable problems in applications today
verisimilitude: We can attack the rotten foundations of computing in other ways, mangol. Consider Elision.
asciilifeform: the 'incurable problems' come 100% outta 'stack chairs' approach.
verisimilitude: I questioned representing all text as characters, and realized it was stupid.
mangol: but if you stack something else, how do you know it's not just the chair of the future?
verisimilitude: By being right.
mangol: i think nobody's right at the moment
asciilifeform: mangol: moreover, the entire economic underpinning of the field as it exists currently is based on studiously avoiding actual solutions to anything
dulapbot: (agriculturalsupremacy) 2020-02-27 asciilifeform: black holes that can absorb ~infinite effort and still suck -- sell. 'job-creating technology'(tm)(r).
mangol: sure, but there are millions of hobbyists who don't seem to have better ideas for future generations of systems
asciilifeform: mangol: consider, as early as '70 cdc co was selling irons on which buffer overflow was impossible. (before bolix et al). today there are 0 commercially available. because, see, 'computer seekoority field' 'needs' to exist, and could not exist in anyffin like present form if such iron were avail.
mangol: buffer overflow is a comparatively trivial problem. bignums likewise
verisimilitude: It's fundamental.
mangol: how do you structure big systems out of parts that are good to think with?
asciilifeform: mangol: all of these problems were 'trivial' in the sense of having been 100% solved in '70s. then ended up 'unsolved', via mechanism described above.
mangol: sure, no dispute
mangol: but we're still talking about different things
mangol: e.g. non annoying approach to concurrency. have yet to hear of one
mangol: nobody knows what to do about concurrency
asciilifeform: mangol: the bolix approach ~worked~. if carried out on current-day vlsi, would work over9000x. would dispute this in some concrete aspect, or simply pretending it 'unhappened' and doesn't merit discussion ?
mangol: closest i've seen is the synchronous programming guys
verisimilitude: Implicit concurrency works well.
verisimilitude: APL is one example.
mangol: asciilifeform: what's bolix?
asciilifeform: the 'non-annoying approach to concurrency' is to implement entire machine as fpga fabric, and let independent processes actually run on physically independent circuits. but most of what people do w/ computer in practice doesn't require any serious degree of concurrency.
mangol: the hardware part of concurrency can be worked out one way or another (challenging, but not conceptually)
mangol: the problem is we don't know how to think about concurrent stuff. those problems are not a good fit for the brain
mangol: the "reactive" and "synchronous" crowds have promising stuff, but not clear it can do everything
asciilifeform: mangol: there's very little 'conceptually challenging' about it when you've proper foundations. here's an illustration of concurrent system with coupla dozen indep. parts where 0 possib. of deadlock or whatever von neumann ills.
asciilifeform: naturally you won't have this on pc or similar von neumann tarpit where memory bottleneck.
mangol: is that Verilog code?
mangol: since describes circuits, basically is synchronous and reactive?
asciilifeform: has both synchronous & asynchronous components
mangol: right
asciilifeform: ( the analogue entropy source aint clocked )
mangol: i know very little about electronics. is synchronous the norm?
asciilifeform: in digital circuits? very much the norm
asciilifeform: ( most off-the-shelf components, e.g. rams, assume a central clock )
mangol: does async tend to be at the periphery (i.e. I/O)
asciilifeform: typically
mangol: have you read about COSA?
asciilifeform: mangol: mentioned this on my www on various occasions, but there in fact is a (1950s!) scheme for 100% async. logic, 'muller's gate'
asciilifeform: read. fluff
mangol: fluff as in not enough specifics?
asciilifeform: ~0 concretes
mangol: i don't know if anyone from that project ever produced an implementation. last i heard, its author was thinking about AI
mangol: (incidentally just came across these: [https://devzendo.github.io/parachute/][The Parachute Project
mangol: ])
mangol: oops
mangol: emulator from 2019
shinohai: What is a transputer, an AI that generates Codes-Of-Conducts ?
asciilifeform: shinohai: lol, nope
dulapbot: Logged on 2022-03-01 19:14:37 asciilifeform: mats: whatever incidentally happened to 'transputer' ? entirely fine 'iron nn' arch of 1980s.
asciilifeform: '80s connectionmachine-like thing.
asciilifeform: mangol: c-element, aha
signpost: impressive.
dpb: funny anagram
asciilifeform read at 1st 'aids' lol
asciilifeform: ... meanwhile cement testbed at 719+k. likely will sync some time in the wee hrs.
asciilifeform: !w poll
watchglass: Polling 15 nodes...
watchglass: 94.176.238.102:8333 : Could not connect!
watchglass: 205.134.172.6:8333 : (172-6.core.ai.net) Alive: (0.041s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Return Addr=0.0.0.0:8333 Blocks=729303
watchglass: 205.134.172.4:8333 : (172-4.core.ai.net) Alive: (0.056s) V=70001 (/therealbitcoin.org:0.7.0.1/) Jumpers=0x1 (TRB-Compat.) Blocks=729303
watchglass: 54.39.156.171:8333 : (ns562940.ip-54-39-156.net) Alive: (0.112s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Blocks=729303
watchglass: 205.134.172.26:8333 : Alive: (0.081s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Return Addr=0.0.0.0:8333 Blocks=729303
watchglass: 71.191.220.241:8333 : (pool-71-191-220-241.washdc.fios.verizon.net) Alive: (0.158s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Blocks=729303 (Operator: asciilifeform)
watchglass: 205.134.172.28:8333 : Alive: (0.083s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Return Addr=0.0.0.0:8333 Blocks=729303 (Operator: whaack)
watchglass: 208.94.240.42:8333 : Alive: (0.144s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Blocks=729303
watchglass: 205.134.172.27:8333 : Alive: (0.343s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Blocks=729303 (Operator: asciilifeform)
watchglass: 54.38.94.63:8333 : (ns3140226.ip-54-38-94.eu) Alive: (0.249s) V=88888 (/therealbitcoin.org:0.8.88.88/) Jumpers=0x1 (TRB-Compat.) Blocks=729303
watchglass: 82.79.58.192:8333 : (static-82-79-58-192.rdsnet.ro) Alive: (0.328s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Blocks=729303
watchglass: 75.106.222.93:8333 : Could not connect!
watchglass: 103.36.92.112:8333 : (terebe.ns01.net) Alive: (0.539s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Blocks=729303
watchglass: 103.6.212.28:8333 : Alive: (0.494s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Return Addr=0.0.0.0:8333 Blocks=385157 (Operator: whaack)
watchglass: 143.202.160.10:8333 : Busy? (No answer in 100 sec.)
asciilifeform: i.e. within 40days of kickoff.
asciilifeform: !q seen cgra
dulapbot: cgra last seen here on 2022-02-28 12:26:43: http://logs.nosuchlabs.com/log/asciilifeform/2022-02-27#1081208 << close enough to ponder whether finland's next, or what was going on again
asciilifeform: !q seen whaack
dulapbot: whaack last seen here on 2022-03-21 12:11:46: !e view-height
asciilifeform: loox like his going even faster
dulapbot: Logged on 2022-03-27 17:45:42 watchglass: 103.6.212.28:8333 : Alive: (0.494s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Return Addr=0.0.0.0:8333 Blocks=385157 (Operator: whaack)
dulapbot: Logged on 2022-03-08 19:54:49 whaack: asciilifeform: plox to add 103.6.212.28 to watchglass
dulapbot: Logged on 2022-03-27 17:49:37 asciilifeform: loox like his going even faster
asciilifeform: cementism does in fact work as printed on the tin; at no pt moar than coupla sec. b/w end of eating a block and start of next.
asciilifeform: still oughta make cgra's suggested corrections before adding to flagship vtree.
asciilifeform: his pt re 'The cement mode, and mining mode should be mutually exclusive' defo valid imho.
crtdaydreams: asciilifeform: ftr would like to mention when I first heard about FPGAs and the ability to program hardware back in 2k20, a month or so before I actually ~read~ chunks of you blog with great intrigue. This was *exactly* what I thought of.
crtdaydreams: The basis would be the FPGA HDL conversion into new lisp fundamentals (lower than either cons or lambda). The core feature would be macros that write the hardware, all HDL code would be in an equiv. "BLOCK" statement. All the prev. features are preserved (homoiconicity, everything can be modified realtime, 7laws etc.) For sanity, hardware-defined code would be generated (for the most part) by macros. Then the lisp would have
crtdaydreams: to be bootstrapped onto hardware written in itself (the hard part).
crtdaydreams: Biggest roadblock would not just be dough, but having sane FGPA fundamentals; if you were to say use Xilinx over Alterra completely diff lang to interface with, resulting in a different core for the lisp.
crtdaydreams: You'd need direct access to FPGA and not the shitware chip(s) in-between.
crtdaydreams: Important steps like circuit allocation would have to be programmed into the bootstrap.
dulapbot: Logged on 2022-03-27 18:21:50 crtdaydreams: The basis would be the FPGA HDL conversion into new lisp fundamentals (lower than either cons or lambda). The core feature would be macros that write the hardware, all HDL code would be in an equiv. "BLOCK" statement. All the prev. features are preserved (homoiconicity, everything can be modified realtime, 7laws etc.) For sanity, hardware-defined code would be generated (for the most part) by macros. Then the lisp would have
asciilifeform suspects that crtdaydreams has rather perverse notion of subj; forgivable for a noob, but srsly oughta rtfm if wants to discuss meaningfully.
crtdaydreams: http://logs.nosuchlabs.com/log/asciilifeform/2022-03-27#1089762 << good links. will definitely read through. admittedly I largely forgot iCE40 was freeware, but I have a suspicion that to bake a lisp CPU you'd need more than 8k. The z80 as it stands has 8.5k~.
dulapbot: Logged on 2022-03-27 20:06:57 asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2022-03-27#1089757 << asciilifeform recs to read into basics of subj. and into detail.
← 2022-03-26 | 2022-03-28 →