Hide Idle (>14 d.) Chans
← 2022-03-15 | 2022-03-17 → ↓
scoopbot: New article on A Syndication of Verisimilitudes: A Review of ``The Codeless Code''
mangol: hello. this might be a good time to ask whether you wanna talk about programming.
mangol: i've been reading the loper blog and the logs, and am impressed. would like to know what kinds of coding projects the people here are interested in pursuing, and what programming languages are in your favor
mangol: my background is mostly lisp and scheme
mangol: plan to stick with these and study the ML family in depth
mangol: would be interesting to get some pointers on Ada, which i know next to nothing about. it seems to be highly regarded here and i'd get an in-depth picture of why
verisimilitude: Hello, mangol.
mangol: in particular, do you think Ada intrinsically fills an important niche which the Lisp and ML families can't handle, or do you consider Ada more of an interim measure -- a better C or C++, say
mangol: (ATS http://www.ats-lang.org/ seems like the ML family equivalent of Ada)
verisimilitude: It's the former, mangol.
verisimilitude: Ada gives me something Lisp never can, rigidity.
verisimilitude: When I already know exactly what I want, Ada usually beats Lisp, although not always.
verisimilitude: Lisp is more suited to very small things than Ada is, usually.
verisimilitude: Know I'm the only APL hacker here, mangol.
mangol: thank you for the helpful pointers. i'll read them in depth.
verisimilitude: It's no issue.
mangol: so Ada 2012 is the latest standard, and Programming in Ada is an introductory book for it by one of the best-known people in the community?
mangol: (is there such a thing as "the Ada community"?)
verisimilitude: I wouldn't claim there's much of a community, however, no.
mangol: well over 95% closed source code, probably
mangol: and often confidential
mangol: verisimilitude: what's your opinion of type inference?
verisimilitude: I don't think about it much.
crtdaydreams: mangol: have you considered joining the wot? It helps to have a continuity of identity.
mangol: i followed the instructions and deedbot accepted my key. it isn't yet shown on the website. maybe there's a delay.
mangol: re: Elision, my first impression is it sounds like a project which would have the same strengths and weaknesses as [Cyc][https://en.wikipedia.org/wiki/Cyc]
PeterL: mangol: have you been following the Pest development?
mangol: hello PeterL. Pest seems like a good design. i've read most of the spec (appreciate the prettyprinted version) but not yet familiar with all the details.
mangol: the main unanswered question in my mind is, how scalable is the design -- how hard is it to handle a pestnet of 100 or 1000 stations
mangol: i'll implement it in scheme if i can find the time
PeterL: Do you think you will ever have 100 people using it? So far we have only had about 10 stations on our test net, mostly using a prototype written in python.
mangol: (Scheme vs Common Lisp is one topic i have a lot of interest in and would be happy to explore, with people who wanna be focused on production and not arguing)
PeterL: I've never really used either, asciilifeform or signpost could probably talk to you about that
mangol: PeterL: i assumed asciilifeform has a plan for scaling Pest, but maybe that's not a goal to begin with
mangol: there are many irc channels with hundreds of people
PeterL: what do you imagine it needs to scale more than it already has?
PeterL: what makes you think it will not scale as it is?
mangol: flood routing + encryption. i'm not well versed in either so take this as the hunch of an uninformed guy.
asciilifeform: welcome to #a, mangol
mangol: asciilifeform: thank you very much!
mangol: you're one of the most interesting people on the net IMHO
asciilifeform: why ty mangol
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2022-03-16#1084540 << in principle you can have any # of stations you like on a pestnet if you set a reasonable bounce cutoff. however asciilifeform's official pov is that a group of mutual ~peers~ must be <= 'dunbar' number. the design reflects
dulapbot: Logged on 2022-03-16 08:30:43 mangol: the main unanswered question in my mind is, how scalable is the design -- how hard is it to handle a pestnet of 100 or 1000 stations
dulapbot: Logged on 2021-11-29 13:30:15 signpost: the correct social topology is human-sized, e.g. the number of reasonable pest peerings.
asciilifeform: !!key mangol
asciilifeform: !!rate mangol 1 noob
asciilifeform: !!v 058E1004DC05D59DC0E10C4027EA78538CE8219F33C04977669A0CF79269BA9D
deedbot: asciilifeform rated mangol 1 << noob
dulapbot: Logged on 2022-03-16 06:08:00 mangol: i followed the instructions and deedbot accepted my key. it isn't yet shown on the website. maybe there's a delay.
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2022-03-16#1084526 << if you know ru, this'll be the best book on subj; otherwise barnes 2012
dulapbot: Logged on 2022-03-16 04:11:12 mangol: so Ada 2012 is the latest standard, and Programming in Ada is an introductory book for it by one of the best-known people in the community?
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2022-03-16#1084545 << it scales precisely as well as it needs to. see above.
dulapbot: Logged on 2022-03-16 08:36:31 mangol: PeterL: i assumed asciilifeform has a plan for scaling Pest, but maybe that's not a goal to begin with
PeterL: I guess it helps to think of a pest network more like a single channel on traditional irc, where there are a handful of people talking, rather than e.g. freenode which could have thousands. It is easy enough to set up multiple pest stations, so you can have disjoing sets of peers
PeterL: for learning ada, there is the rationale and specification documents which are long and detailed, but worth reading if you have the time
mangol: makes sense. so for pest nets that >dunbar people are interested in reading, the idea is to publish the logs on a separate medium.
PeterL: right, like the log
mangol: and having >dunbar people with write access to the same channel will create a mess, so that's not an important use case
PeterL: this is all assuming the current use. If signpost gets his file transfer protocol working on top of pest then we might need to look at other ways for optimization, but at this point that would be premature
dpb: mangol, check out http://atruechurch.info/ and don't go to hell like the rest of the world!
mangol: wow, i didn't know they have door-to-door preachers on irc!
PeterL: dpb, have you tried standing up a blatta yet?
dpb: no
mangol: some believe that pr0n should be the first use case on which to test all new protocols
mangol: but why not religion
mangol: subversive message for a subversive protocol
PeterL: pest-sexchat when?
mangol: HexChat -> hexadecimal -> sexadecimal -> SexChat
mangol: sexed vs hexed
PeterL: so mangol, where are you from? you got a blog or anything that you have written out on the internet you could point us to?
PeterL: do you identify as a millenial, gen-z, gen-x, or boomer?
dulapbot: Logged on 2022-03-16 10:37:47 mangol: wow, i didn't know they have door-to-door preachers on irc!
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2022-03-16#1084574 << the correct mental model for a pestnet imho is a party in a large room. if errybody's voice carried to entire room, would be impossible to speak. however people normally stand in groups and can speak, sometimes moving from one to another
dulapbot: Logged on 2022-03-16 10:30:54 mangol: and having >dunbar people with write access to the same channel will create a mess, so that's not an important use case
asciilifeform: the current pestnet is ~dozen people 'sitting around coffee table' which worx ideally.
mangol: PeterL: .fi
mangol: PeterL: i identify as an x-millennial
asciilifeform: mangol: btw we even have another finn - cgra
mangol: torille!
mangol: my pronouns are let/lambda
signpost: wahahaha
signpost: welcome mangol
signpost: http://logs.nosuchlabs.com/log/asciilifeform/2022-03-16#1084543 << to my mind there are productive kinds of argument, though likely not the kind you're referencing.
dulapbot: Logged on 2022-03-16 08:34:49 mangol: (Scheme vs Common Lisp is one topic i have a lot of interest in and would be happy to explore, with people who wanna be focused on production and not arguing)
signpost: "flamewars" are usually driven by particularly stupid people who ought to stick to watching sports.
signpost: re: the particular subj, I admit to not having tried the more recent schemes, but find most of what I need in a lisp in CL, and lack of moving target a plus.
dulapbot: (trilema) 2019-02-04 asciilifeform: outside of unix planet, typical example of type 3 failure is r5rs scheme (which is ~impossible to write nontrivial program in without reintroducing good % of what makes it seem 'light' vs commonlisp )
dulapbot: (trilema) 2015-08-07 asciilifeform: http://log.bitcoin-assets.com/?date=07-08-2015#1229675 << complicated. scheme is an example of a 'suckless movement' that didn't catch on. that is, 'throw out all the legacy crud', but after that the thing has to be inhabited. if not inhabited, then dies young, while leaving, yes, beautiful corpse
mangol: signpost: asciilifeform: thanks for the thoughtful remarks. tbh, this is exactly the kind of discussion i was hoping to find here
mangol: the problem with most debates is not that they're debates, but that most of the debaters 1) have a shallow understanding of the topic; and 2) are more interested in tearing things down than supporting existing things and building new ones
mangol: shallow people can easily generate lengthy and controversial debates where their arguments seem impressive to other shallow people, and sometimes turn into netwide memes that go on for a decade or more
mangol: 1) is rarely a problem without 2). a shallow understanding, in someone more interested in creation than destruction, quickly develops into a deeper understanding
mangol: CL vs Scheme makes for a great flamebait topic like that. Naggum had at least one insightful remark on it (i.e. that Scheme is incompatible with CL and so dilutes Lisp's brand). i consider it partly obsolete.
signpost: belief that there's one shared reality, that one human's rarely sufficient to see all of it, and in the iterated solution to the prisoners' dilemma gets one pretty far imho.
mangol: sounds like a decent approach
mangol: IMHO most fruitful to see people as processes. people have very different personalities, and when you put a particular kind of personality in a specific kind of formative environment, a peculiar information-gathering process develops
mangol: which has peculiar strengths and weaknesses
mangol: collect enough personalities and you slowly learn who to go to for what kind of topic or perspective
mangol: the culturally approved method is to gather "objective facts". the main reason i seek crackpot forums like this is that people don't do that here.
mangol: and the force of personality is multiplied when creating new things instead of simply learning about existing ones. (not that learning isn't a small kind of creation in its own right)
mangol: so ultimately i'd like to build a bunch of cool stuff, and am missing many of the personalities required to do so :)
mangol: i'm reminded of asciilifeform 's "fungible cogs" essay. cogs don't create new kinds of things.
signpost: nope, what they create is slave empires, in aggregate. old as history.
signpost must bbl
mangol: signpost = trinque?
shinohai: http://logs.nosuchlabs.com/log/asciilifeform/2022-03-16#1084545 <<< The fact that it doesn't "scale" is a feature and not bug imho.
dulapbot: Logged on 2022-03-16 08:36:31 mangol: PeterL: i assumed asciilifeform has a plan for scaling Pest, but maybe that's not a goal to begin with
shinohai: $vwap
busybot: The 24-Hour VWAP for BTC is $ 40236.68 USD
shinohai: $ticker btc usd
busybot: Current BTC price in USD: $40210.12
verisimilitude: http://logs.nosuchlabs.com/log/asciilifeform/2022-03-16#1084536 My model is primarily concerned with form, mangol. When I resumed studying Latin, I realized I could get a higher-level Elision with much less work, albeit not without difficulties in entering existing Latin texts into the system.
dulapbot: Logged on 2022-03-16 06:31:41 mangol: re: Elision, my first impression is it sounds like a project which would have the same strengths and weaknesses as [Cyc][https://en.wikipedia.org/wiki/Cyc]
signpost: mangol: yep that's correct
mangol: signpost: (and why not verisimilitude and others): what languages do you use currently for hobby projects?
mangol: i have a ton of ideas, especially around the lisp family, so would be interesting to compare notes
signpost: verisimilitude: been a pet hunch of mine that english is a barrier to machine inference over language because it's syntactic goop.
signpost: mangol: typically CL over here, python "in anger" sometimes.
mangol: neat. is deedbot written in CL?
signpost: it is
signpost: http://trinque.org/src << can be found here
shinohai: http://logs.nosuchlabs.com/log/asciilifeform/2022-03-16#1084583 <<< Wouldja believe Ms Shinohai already asked that question? lol
dulapbot: Logged on 2022-03-16 10:41:40 PeterL: pest-sexchat when?
signpost: shinohai: wot-tronic sex work is a when-not-if
signpost: obvious application imho, good for all involved
shinohai had the problem that only women far South of his current local had half a brain to grok the benefits ....
PeterL: I use python, and have done some small projects with ada in the past
shinohai: "But my cashapp Bitcoin wallet only gives me `3` addys, y can't you pay me?"
signpost: myup
mangol: signpost: thanks for the src link, will peruse
signpost: world getting more dangerous will solve these pretenses, and starve the north american ones down to a respectable size to boot.
signpost: questionable whether that solves anything regarding the lifetime of imitation food
shinohai: Nah not unless you get one of the aforementioned non-americanized grils that cooks fresh daily.
mangol: PeterL: i will probably never be useful for Ada. my low-level programming goal is to take existing large C codebases and auto-translate them into a better type system. not yet clear whether this is feasible. finding it out will take years of work.
shinohai inhales the aroma of fresh arepas as we speak .....
signpost: shinohai: found myself a lovely blend of redneck and well-read. don't expect the experiment is repeatable, lol
PeterL: mangol: why don't you think you will get ada? I find it much easier to understand when reading it than c/c++
mangol: PeterL: will probably get ada, but won't be able to invest the time to write useful code
mangol: signpost: congrats!
signpost: girl can just as well discuss european literature as rig a makeshift still
shinohai: Sadly signpost the only reading redneck girls do in my locale is the back of boxes of cold medicine to see if contains right ingredients for meth-making.
mangol: lol
signpost: texas is a special place, or was!
mangol: meth-making and matchmaking are mutually exclusive
signpost: tbh I credit the proximity of mexican culture.
shinohai: Well I mean I have Colombian girls in the house but we don't process freebase coke.
signpost dated a colombiana
mangol: freebase and FreeBSD also don't fit in the same community, or so i imagine
shinohai: But they do know how to enter a chroot which is nice.
signpost: very smart and batshit crazy
signpost: was fun!
signpost: mangol: what do you keep yourself busy with in meatspace?
signpost assumes programming
shinohai rather likes the fiery temper part of these girls
mangol: signpost: you're not wrong
mangol: signpost: asciilifeform: re: Lisp and Scheme, my current thinking is to summarize them as follows:
mangol: Lisp = representing code as lists
mangol: CL = Lisp + closures + generic functions
mangol: Scheme = Lisp + closures + tail calls
verisimilitude: I use Ada, APL, Common Lisp, and the other main project of mine regards hardcore machine code hacking, mangol.
mangol: everything else is incidental, unless i forgot something
signpost: mangol: TCO is nice, but SBCL and iirc several other implementations have it
signpost prefers to think recursively
verisimilitude: It's not guaranteed.
mangol: signpost: me too
mangol: some Scheme implementations also have generic functions
mangol: and some C compilers can do TCO
signpost: verisimilitude is correct I believe, certain forms around the tail call, or lack of optimization flags thwart tail-call optimization in sbcl
mangol: the essence of a language is to embody a particular programming style (which some other langs/compilers can also do, but it's not the main thing there)
mangol: generic functions are the essence of CL (as distinguished from Maclisp and Scheme)
mangol: tail calls are the essence of Scheme. (contrary to popular belief, first-class continuations are not)
mangol: verisimilitude: re: Ada, APL, CL -- sounds like you're the language connoisseur of the channel
mangol: interestingly, a lot of CL programmers (myself included) don't use GFs very much, sticking closer to the trad. dialect. like many C++ programmers stick to C style
signpost: mangol: in that deedbot code you'll even see me naming generics in a java-OOP influenced way.
signpost: it was earlier in my practice with the language.
signpost: e.g. ircbot-connect vs connect
mangol: i started out that way too. (am by no means an OOP wizard, and would love to talk to somebody who is to find out more about the philosophy)
mangol: Smalltalk (and Ruby = Smalltalk + Perl) is class-centric. CLOS (and Dylan IIRC) is method-centric
mangol: a Smalltalk-80 implementation in CLOS would be very enlightening to compare the two approaches
signpost: right, my sense is that my naming should've described classes of object-interaction, not the hierarchy of objects acted upon.
signpost: "object protocols" is probably the right term
mangol: exactly. (the idea, that is; not sure about the terminology -- not enough OOP cred)
mangol: it's my impression that people who build big, long-lived systems invariably find that hierarchies are too constraining
mangol: whence sprung the increasingly popular "composition, not inheritance" meme
mangol: fortuitously, method-centric OOP demotes inheritance into almost an implementation detail
mangol: a lot of people, myself included, find thinking about class hierarchies almost impossible, whereas thinking about composition (protocols) is natural
signpost: yep, had that one thumped into my head at last $megacorp job, almost to excess
mangol: http://trinque.org/src/ -- ircbot-* and logbot-* together encompass several different but related bots?
signpost: nah, failure of comoposition!
signpost: latter adds a db connection, former just connects to irc
verisimilitude: I don't care too much about CLOS, but use it when I see opportunity. The programming style I champion most is done easily in most any language.
verisimilitude: Whereas CLOS has :BEFORE, :AFTER, and :AROUND methods to deal with irregularities, my style supports them so easily.
mangol: verisimilitude: "The programming style I champion most is done easily in most any language." --> a common sentiment among good programmers
mangol: verisimilitude: in my experience, people who arrive at this sentiment flock around Scheme and ML a lot
mangol: i've heard at least two other schemers explicitly say the same. these are "lambda calculus assembly language", tongue only half in cheek
mangol: tabular programming?
mangol: aka table-oriented programming ("TOP")
verisimilitude: Yes and no.
verisimilitude: Read it.
mangol: verisimilitude: to a useful approximation, CL is CLOS, and CLOS is generic functions (i.e. method-oriented OOP)
verisimilitude: Yes, but the error in that approximation is so painful.
mangol: the parts of CL that aren't CLOS (or CLOS-like) are trad. lisp with some cruft cleaned up, and excellent libraries.
mangol: Emacs Lisp is approx. what you get when you take CLOS out of CL
mangol: verisimilitude: i read your essay but i can't quite parse what is the extra ingredient on top of TOP
verisimilitude: Link me to an article on table-oriented programming. We may have read the same thing.
verisimilitude: Using JavaScript and cookies, I still can't view what garbage into which he turned that wiki.
verisimilitude: I found the raw JSON.
verisimilitude: I've several answers, but will finish reading first, so let us discuss this in a few hours, mangol.
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2022-03-16#1084654 << autotranslation of proggies is generally a dead end. and not only because produces in ~all cases ~unreadable soup. but e.g. how wouldja 'automatically' remove reliance on pointerism from c liquishit ? in vast majority of cases, braindamage which 'comes with' a lang aint removable except by vehehery careful rethought of entire proggy
dulapbot: Logged on 2022-03-16 12:44:05 mangol: PeterL: i will probably never be useful for Ada. my low-level programming goal is to take existing large C codebases and auto-translate them into a better type system. not yet clear whether this is feasible. finding it out will take years of work.
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2022-03-16#1084658 << can't resist to ask, have you read asciilifeform's ffa series ? imho entirely usable 'intro to ada'
dulapbot: Logged on 2022-03-16 12:45:21 mangol: PeterL: will probably get ada, but won't be able to invest the time to write useful code
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2022-03-16#1084664 << 'meth-maker, meth-maker, make me a ...'(tm)(r)(c)
dulapbot: Logged on 2022-03-16 12:47:09 mangol: meth-making and matchmaking are mutually exclusive
dulapbot: Logged on 2022-03-16 13:55:32 mangol: Emacs Lisp is approx. what you get when you take CLOS out of CL
dulapbot: Logged on 2020-07-03 20:17:15 asciilifeform: cuz elisp has the efficiency & reliability of a trs-80 soldered from dumpster scraps
dulapbot: (trilema) 2018-06-19 asciilifeform: mircea_popescu: elisp is a gnarly surviver of dark age 1970s lisps, suffers -- unfixably -- from problems long ago solved even in 'scheme' (e.g. 'funarg problem')
dulapbot: Logged on 2022-03-16 13:55:32 mangol: Emacs Lisp is approx. what you get when you take CLOS out of CL
dulapbot: Logged on 2020-07-03 20:17:15 asciilifeform: cuz elisp has the efficiency & reliability of a trs-80 soldered from dumpster scraps
dulapbot: (trilema) 2018-06-19 asciilifeform: mircea_popescu: elisp is a gnarly surviver of dark age 1970s lisps, suffers -- unfixably -- from problems long ago solved even in 'scheme' (e.g. 'funarg problem')
asciilifeform: ... meanwhile, cement testbed at 645+k
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2022-03-16#1084626 << lolyes. asciilifeform would like to meet the fella who thinks he has over9000 folx worth exchanging keys & directly peering with
dulapbot: Logged on 2022-03-16 12:31:39 shinohai: http://logs.nosuchlabs.com/log/asciilifeform/2022-03-16#1084545 <<< The fact that it doesn't "scale" is a feature and not bug imho.
verisimilitude: He has a name, abbreviated by U and S.
shinohai sees pest as appendage of wot, where one doesn't keep garbage keys around in personal keyring ...
mangol: http://logs.nosuchlabs.com/log/asciilifeform/2022-03-16#1084735 << sorry, i wasn't being precise. auto-translating into a language with a better type system. the system needs an "unsafe" escape hatch for pointers. it may be possible to mechanically translate some pointer code into safe types with the help of hints from the programmer; haven't looked into it.
dulapbot: Logged on 2022-03-16 16:20:43 asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2022-03-16#1084654 << autotranslation of proggies is generally a dead end. and not only because produces in ~all cases ~unreadable soup. but e.g. how wouldja 'automatically' remove reliance on pointerism from c liquishit ? in vast majority of cases, braindamage which 'comes with' a lang aint removable except by vehehery careful rethought of entire proggy
verisimilitude: Ada uses access types very sparingly, mangol.
verisimilitude: Ada encourages a programming style very different than the C language.
mangol: the idea is that one can incrementally translate unsafe code into safe code. C code is translated into "safe" C++ in this way, but C++ is broken beyond repair so it'd be nice to have something simpler and better
verisimilitude: This is a fundamentally mistaken way.
mangol: what's the mistake?
verisimilitude: The only way to do something similar well is completely rewriting it.
mangol: that's my main motivation -- nobody can rewrite all our legacy C code from scratch. even if you take the capabilities of a modest Unix system
verisimilitude: Who would want to do that?
verisimilitude: UNIX is vile shit.
mangol: i'd hazard a guess you're typing that on a unix-like OS
verisimilitude: No, in Emacs.
mangol: running on unix
verisimilitude: Listen here and well, mangol, popularity is no indication of quality, or why would UNIX be better than Windows?
verisimilitude: Windows is more popular.
verisimilitude: Oh, but Windows is obviously shit, so that doesn't count, right?
mangol: well, you have to run either of those if you wanna do anything practical
verisimilitude: Should it be so forever?
mangol: both are written in C (Windows probably quite a bit of C++ as well)
verisimilitude: People here are interested in escaping this nonsense, not embracing it, mangol.
mangol: countless people have volunteered to write a better OS, yet virtually no people can use the results
verisimilitude: They existed once, and, in the words of asciilifeform, no one can uninvent them.
verisimilitude: UNIX and Windows are already on the way out, for the next generation of this attitude, the WWW with JavaScript.
mangol: this is not a question of right vs wrong ways to do computing -- it's a question of whether or not incremental approaches have any value
verisimilitude: Sure; I'd argue they usually don't.
mangol: that's where we differ
mangol: i put virtually all of my effort into incremental stuff
verisimilitude: I've recently written a binding to the underlying UDP interface for Ada, mangol, and I'm taking pains to contain it so that it won't spread to everything which it touches.
mangol: nothing wrong with from-scratch efforts, but virtually all of them should be considered "research"
dulapbot: Logged on 2022-03-16 14:46:33 verisimilitude: I've several answers, but will finish reading first, so let us discuss this in a few hours, mangol.
verisimilitude: Those approaches are concerned with programming as in a database.
verisimilitude: I'm concerned with representing a function as static data, as much as feasible.
verisimilitude: The only variable data in my Roman Numerals program are the indices.
verisimilitude: I then show how the approach makes it easy to prove software correct, and how I can even encode the inverse function easily.
verisimilitude: I want to reduce the code, the actual code, to its minimum.
verisimilitude: The differences should be clear now.
mangol: where others would use code, you prefer to have data?
verisimilitude: It's compression; the program could be represented as a collection of all possible answers.
verisimilitude: By ``tabular'', I mean any arrangement of static data; one design for a later program will use a trie in one part, as an example.
mangol: sure
mangol: i assume you've come across the branch of information theory where the amount of information in some data = length of shortest known program to generate that data
mangol: chaitin, kolmogorov, etc.
mangol: re: #asciilifeform and incremental approaches -- Naggum was a staunch proponent of incrementalism. notably, he lambasted young people who think they can improve upon CL by starting from scratch
signpost: verisimilitude: I worked under a gent once upon a time who espoused the same data fundamentalism.
signpost: probably good lisp code tends toward what you describe as well, where what was imperative code ends up a declarative DSL for the problem-space.
signpost: see a lot of merit in it, actually.
signpost: mangol: heh, this is not meant as a throwaway comment, but Naggum ulcered himself to death too.
signpost: man not separable from works.
mangol: AFAICT the main cause of naggum's untimely death was financial stress, which in turn was caused by some weird principle whereby he refused to pay his taxes
mangol: he was conscientious to a fault, so no doubt was stressed about other stuff too, but that was the clincher
signpost: then not through-and-through incrementalist, or why did paying his taxes bother him?
signpost hates the state in all forms, but worthwhile question.
mangol: nobody is a through-and-through anything
signpost: still these incoherent outputs, and the creatures that produce them, interest me perhaps more than anything.
signpost: sure the thing can operate in a state of contradiction just fine. yet it knows it does, and still... does!
mangol: i'm pretty sure anyone who says humans are incoherent, is using a less effective definition of coherence than what nature is using
signpost: nature's mighty comfortable with eventual consistency, surely.
verisimilitude: I'm aware, mangol.
signpost: mangol: still that's a large shadow in which to hide things, that because nature allows for conflict, also conflict within oneself is productive.
mangol: nature doesn't seek consistency in the things where inconsistency is easiest for us to notice
signpost: probably wouldn't argue with that, still naggum's noteworthy both in his works, and downfall.
mangol: yeah :)
mangol: brilliant people tend to have inexplicable weird habits and beliefs
verisimilitude: I just checked; gravity still applies; nature is consistent.
signpost: some of that I suspect is a hobbling meme
signpost: nerd cannot also talk to girls; that sort of thing.
verisimilitude: Rumors start in truth, signpost.
verisimilitude: I'm certainly touched in the head.
signpost: yes, but do not consummate truth.
signpost wiggles eyebrows
shinohai: Quid est veritas?
verisimilitude: VĒRITĀS EST VĒRITĀS
signpost: shinohai: I suppose that's what the eventually consistent algo mangol mentioned is evaluating.
verisimilitude: VĒRITĀS QVOQVE ERAT VĒRITĀS
signpost: give it time.
signpost: as a side note, it's one of those magnificent days in Austin, approx 75F, sun shining.
mangol: signpost: "conflict within oneself is productive" >> IIRC this was the main point of Minsky's book Society of Mind
signpost: pouring caffeine into my head at a favored spot
mangol: neat. snow here :D
signpost: mangol: oughta unpack conflict, unfamiliar with the book
signpost: mangol: that's any given day in Finland, no?
mangol: i wish :D
signpost: there is the productive kind of conflict, where one divides himself into a sort of honest representation of uncertainty regarding a given problem.
signpost: and the unproductive, where the truth has been long known, yet ignored.
signpost sees both the former and latter in the character of naggum.
signpost: (well and in us all, which is why it's even interesting to say)
signpost: lemme give an example. on the subject of one's own death, cannot be divided.
mangol: this is a very personal opinion that i don't expect anyone to share, but in my experience, a pronounced concern with ethics is a red flag for long-term failure of the naggum variety
signpost: most of the failures of our contemporaries have to do with division towards that one.
signpost: mangol: expand, really depends.
signpost: maybe he wanted to "joe stack" that way, and here we are still talking about it.
signpost: perhaps he's satisfied.
signpost: *would be
mangol: a lot of the stuff we do in civilized societies is luxury afforded by idleness -- stuff that disappears as soon as there's something actually at stake. ethics tends to be at the top of that list (i.e. one of the most expendable things humanity has invented)
mangol: people who spend time thinking about ethics are spending time thinking about something that virtually all people will just throw by the wayside the moment they stand to lose or gain something
signpost: skeptical. are we going to trust each-other enough to survive in the ancestral environment if I'm ever-waiting to stab you in the side?
signpost: the luxury is in never encountering ethical dilemma, not in considering it
mangol: IMHO morality is very real, ethics is wishful thinking on top
signpost: problem of definitions then
signpost: interpersonal honor is the only thing that matters when the luxuries run dry
mangol: morality comes from the animal part of us. group cohesion is impossible without it. ethics is some codified fiction about how someone thinks morality will work
signpost: perhaps say what you consider to be the difference
mangol: morality doesn't require thinking (i.e. rationality) to work. even animals can do it (e.g. prioritize their kin or group)
signpost: sure, derived from god, or instinct, as you like
mangol: whereas ethics consists entirely of thinking, abstractions
signpost: dunno that ethics are expendable like this, even in the hard "outside".
signpost: suppose your brother lay dying, shot, as you've been ambushed. the attackers are headed to your town.
signpost: do you render aid to your brother, or rush back to warn.
signpost: reason might cause you to do something which avoids the pitfalls of crude heuristics in the moment.
verisimilitude: The moral is survival.
signpost: of whom
verisimilitude: It could be the individual, or his children, or his friends.
mangol: signpost: i'd be very surprised if someone makes an ethical (e.g. trolley problem style) decision in that kind of scenario. probably the animal part of their body will drive them toward a decision, and the concious mind will confabulate some explanation if needed
signpost: verisimilitude: the lack of weights assigned to the people involved is certainly a problem. supposes human fungibility as a given.
signpost: mangol: how many such situations do you suppose it takes to desensitize one to the immediate feeling?
signpost would not be surprised if harsher living conditions result in *more* rational decisionmaking, not more "animal" urge.
signpost: even worthwhile to consider the political usefulness of creating space within which "I would've done otherwise, but urge overwhelmed" is allowable.
verisimilitude: Oh, we may see less aposematism, signpost?
signpost: another item to unpack at length
verisimilitude: That's why I insult the stupidity of those who take the trolley problem seriously, signpost.
verisimilitude: Such philosophical nonsense is worthless to me.
signpost figures the ancients had the dispensative / receptive sexual categories correct.
signpost: verisimilitude: yeah, it's constructed to get you to say "I would of *course* save the side with more voters!"
mangol: signpost: rational decision-making doesn't have to be ethical decision-making
verisimilitude: Why stop there? Why not include sexual responsibilities and attitudes towards the unknown?
signpost: mangol: pull 'em apart for me
mangol: ethics is a specific kind of rationality -- one that involves moral "shoulds"
mangol: rationality is just any anticipation of consequences by reasoning, doesn't need to have a moral aspect, and definitely doesn't have to be morally "correct"
signpost: hmmm
mangol: ethics is about putting morality into a self-consistent system. rationality can just take morality as is, without passing judgment on it
signpost considers certain outcomes of reasoning inevitable given correct definitions of being
signpost: but that's by no means straightforward or settled.
mangol: concretely, people like Naggum and RMS e.g. expressed ethical concerns about US presidents, as if presidents were selected on ethical grounds
mangol: that kind of over-applied and mis-applied conscientiousness would eventually drive most people mad
mangol: people who aren't brilliant-and-peculiar outliers tend to notice signs that they're going mad, and stop being so concerned about ethics
signpost: why should I go mad by knowing US political outcomes are ethically indefensible?
signpost: I do not also have such narcissism that I consider this my doing
signpost: (and I suspect that latter bit is actually what's happening)
mangol: you shouldn't, because you aren't an outlier of the naggum/rms variety :)
signpost: democracy lies by telling each special atom that they're a fungible unit of sovereignty
mangol: normal people don't feel a compulsion to follow principles to their conclusions
mangol: even most abnormal people (such as this chan) don't feel such a compulsion very strongly. it's really quite rare
signpost: exceptional people can still be infantile in the ways they refuse to inspect.
signpost: see: Mircea Popescu
signpost: it's this, not a consequence of their being exceptional.
mangol: yep. another good example
signpost: just that nothing says being a towering intellect means you learn to wipe your arse correctly
signpost: or, you know, swim with, not against.
signpost: (or without rum)
dulapbot: Logged on 2022-03-16 17:41:26 mangol: AFAICT the main cause of naggum's untimely death was financial stress, which in turn was caused by some weird principle whereby he refused to pay his taxes
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2022-03-16#1084860 << honor is real, if 'not evenly distributed'. plenty of folx when rubber hits the road prioritize something other than own skin. whether you grow up hearing about'em depends heavily on yer culture, the anglo space is particularly rotten in this respect
dulapbot: Logged on 2022-03-16 18:01:52 signpost: interpersonal honor is the only thing that matters when the luxuries run dry
asciilifeform: some cultures -- 'split'. modern ru for instance by asciilifeform's analysis largely divides into descendents of a) folx who did right thing in ww2 b) scum who 'looked after own skin' , and each behaves moar or less as one'd expect
signpost had the benefit of a (retired) italian culture on one side, largely evident by "something used to be here" than "something is", and mil on other.
asciilifeform: sometimes entire cultures succumb to 'psychology of the traitor' (ru concept w/ very little exposure in engl., outside of #t/#a lol) -- e.g. south kr (100% of elite traceable to collaborators w/ japanese then american colonizers), ukristan
signpost: former rotted within a generation of exposure to american air.
signpost: latter into "what, we were men, for this?"
asciilifeform: the closest engl. term afaik -- 'hamstering'
asciilifeform: alternates b/w 'it was inevitable', 'foreigners were simply Better Than Us', 'we Just Wanted to drink bavarian(tm)' and a good % of time 'mno it didn't happen whatcha talking about' when rubbed nose 1st into the act of betrayal
signpost: one can see something of this, perhaps prelude, of, in silicon valley.
signpost: welcoming e.g. chicom culture into the states and getting paid obscenely
asciilifeform: cultures where the traitorous scum 'lost' (such 'loss' sadly always temporary) -- north kr, yugoslavia ( both ruled after the war by partisan commanders who stood up for native soil ), to partial degree su
asciilifeform: china notably
dulapbot: Logged on 2022-03-16 18:26:21 mangol: even most abnormal people (such as this chan) don't feel such a compulsion very strongly. it's really quite rare
signpost: *prelude of
asciilifeform: (on cn tv still one can see historical dramas about partisans tortured by japanese gestapo etc, and the collaborationist scum all were pushed off to their scum island, nao usg's 'unsinkable carrier'(tm))
verisimilitude: The Japanese almost destroyed the communists, apparently, but had to defend against the US.
asciilifeform: at any rate nato reich took the angle of simply destroying errybody's very concept of 'native soil'
asciilifeform: 'why should i die for this hotel' will answer today's usa, uk, fr, de, etc inmate.
verisimilitude: I'm sure they also have programs about how the four pests campaign went wonderfully, if they even mention it, asciilifeform.
asciilifeform: verisimilitude: afaik not
signpost is no friend of CN, but they're inarguably one of the few over there that don't round to american colony.
asciilifeform: there's a 'clef' to the 'roman a clef' imho.
verisimilitude: We had this discussion a week or so ago.
asciilifeform: some cultures 'alive' (if ill), others -- had 'heart cut out'
verisimilitude: I can believe North Korea is independent; China obviously isn't.
asciilifeform: verisimilitude: what's it dependent on ?
signpost: bud CN will probably be one of the great outmaneuvers of all history.
verisimilitude: They're called ``lizards'' here, asciilifeform.
signpost: they are positioning themselves for world empire.
asciilifeform: verisimilitude: what do the lizards supply to cn that it can't do without ? green toilet paper?
signpost: j00z better get moving.
verisimilitude: I don't know.
signpost: asciilifeform: one thing that I question is whether putin believes the chicoms are his friend.
asciilifeform: signpost: fella's got moar brain than that
signpost: just looks like somebody else about to stick a straw in the slavs
signpost: I would assume.
asciilifeform: 'only friends of ru are its army and navy'(tm)(r)(alex the 1st)
verisimilitude: Israel has sold plenty of weapons to China.
verisimilitude: We'll see what else China gets, and from whom.
dulapbot: Logged on 2022-03-03 00:43:16 asciilifeform: if continues on that trajectory, eventually result -- 'eurasian reich'.
signpost: seems both RU and CN hadn't much issue demonstrating working hypersonics
signpost: US less so.
asciilifeform: together they add up to a su circa 1960s, w/ modern irons.
signpost: iirc j00z also lacking in these
verisimilitude: Intel has plants in Israel, right?
asciilifeform: the 1 natural resource abundant in ru & cn, and not muchly elsewhere, that aint listed in any 'cia fact book', is 'willingness to die for motherland'
asciilifeform: it can't be poured into barrels , bought&sold etc. but is a thing.
verisimilitude: That really helped them during the Opium Wars, asciilifeform.
verisimilitude: Of course, now they have the white man's weapons, so it's different.
asciilifeform: verisimilitude: it arguably ~generated~ by -- not least -- the opium wars.
signpost: asciilifeform: don't get much of this pouring unwanted peasants into factories on another continent, and eventually rewarding some of them with "congrats! yer white nao!"
verisimilitude: That really helped them during the Japanese occupation, asciilifeform.
asciilifeform: verisimilitude: current-day cn in fact 'celebrates' a 'national humiliation day'
signpost: verisimilitude: the pretense that CN can only steal "white man's" weapons is going to get white man killed
verisimilitude: I doubt it, signpost.
signpost: I would welcome your doubt being proved justified.
asciilifeform: verisimilitude: national humiliations have a kind of 'maxwell's daemon' effect on folx.
asciilifeform: (how -- see above thrd)
verisimilitude: Where are the Chinese-made weapons that aren't disease and famine?
asciilifeform: some % -- become a nation. some % -- taiwanize.
verisimilitude: If the US fights China, those playing both sides will win anyway.
asciilifeform: verisimilitude: lol just about erry plastic widget on yer desk, is a lethal cn weapon.
asciilifeform: to the extent few others needed.
signpost figures these are weapons so fearful they have to not exist, or be explained away.
signpost: because otherwise gotta confront that... we don't have 'em
signpost: iirc our last attempt broke up over pacific
verisimilitude: China's known for accurate reporting, just like the USSR.
asciilifeform entertained that folx still seem to think that 'who whom' decided by 'who has longer rocket'
signpost: asciilifeform: at some point enemy can hit faster than you can launch, no?
signpost: but sure, rot, far better, more effective.
signpost does not argue otherwise
verisimilitude: What's a nuclear submarine, signpost?
asciilifeform: signpost: phunphakt, mao was in favour of getting it over with and having the nukefest
verisimilitude: The glorious half will live, and the evil half will die, right, asciilifeform?
asciilifeform: 'kill 90% of us , we're still cn, kill 90% of you, light out'
signpost: asciilifeform: not an indefensible notion
asciilifeform: may've heard mao's trademarked 'the bomb is a paper tiger'. is what ref here.
verisimilitude: That paper tiger ripped through paper Japanese houses.
signpost: verisimilitude: USA's utter lack of a traditional culture can't be denied.
signpost: if the shelves ran dry, couple weeks of cannibalism and then zilch
verisimilitude: I agree, signpost. I won't die for it.
asciilifeform: verisimilitude: at some pt will rip through the even paperier houses in asciilifeform's shithole, what of it
verisimilitude: I suppose nothing of it, then.
asciilifeform: (modern usa seems dead set on replicating hiroshima architecturally, for sumreason)
signpost doesn't discount that something can emerge from actual suffering imposed, but what is unclear
signpost: and birth of nations takes longer than a signpost interval
signpost: so doesn't help me much.
verisimilitude: Think about what would happen when the police no longer protect certain kinds of people, signpost.
asciilifeform: the current great kingdoms emerged this way. is 'geological' observation, you can't exactly take it to the bank & cash it
signpost: verisimilitude: right now I'm sitting on a coffee shop deck, and there's a "wild" rabbit sitting not 3ft from me.
asciilifeform: and imho knowing the fact aint useless, despite this. is interesting to know that e.g. not 1 weapon in reich larger than rifle is made w/out chinesium parts. and not 1 factory.
signpost: he's so desensitized to me that he approximately lacks the instincts needed to avoid me if I later decide he's lunch.
signpost: the people that populate your country are the same.
asciilifeform: cn philosophy 'sit by riverside and wait for enemy's corpse to float by' is in action as we speak.
signpost: were danger to re-emerge they'd be decimated, and sure, perhaps what's left contains some that can function, as randomness would have.
verisimilitude: Sure, signpost.
verisimilitude: We agree too many people are alive today.
signpost: well and, once the real is encountered, too few.
asciilifeform: btw modern cn party line is that they're ~under~ populated
asciilifeform: (prev. 'party line' 'worked too well')
verisimilitude: All of that infanticide must show their ability to think ahead.
dulapbot: Logged on 2022-03-07 13:43:08 asciilifeform: ru demographic loss in '90s exceeded ww2 losses by solid margin. (and this not even counting the seceded territories)
verisimilitude: At least the infanticide in the US has a clear purpose.
signpost: US also under replacement rate iirc
asciilifeform: (no small part of why ru loox like such a delicious pie to invaders)
asciilifeform: at any rate, upstack, 'country' that becomes 'hotel', where at 1st sign of trouble errybody runs to wherever legs will carry'em -- is already 'lost'
asciilifeform: in such shithole, even long before war, 'no, won't build factory here, will build in cooliestan and buy 3 yachts w/ the difference' is tolerated, encouraged, until hollow shell
asciilifeform: 'hotel'.
signpost: that RU does not appear to be susceptible to USSR memberberries, but Tsar-era, oughta be instructive.
signpost: US has no tsar era to which to return.
signpost: the utopian mistake is all there ever was.
asciilifeform must bbl
signpost gonna find some food also, cya
signpost: verisimilitude: will say, the desire for a solution to this dead end was core to my involement in tmsr.
signpost: even considered mp one of the most likely folks able to resurrect being european contemporarily.
signpost: gods said otherwise.
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2022-03-16#1085028 << in ru folx view history as continuous whole, roughly from 1k a.d. up. to some degree this pov imposed 'top-down' but quite large afaik 'grass root' component to it.
dulapbot: Logged on 2022-03-16 19:02:56 signpost: that RU does not appear to be susceptible to USSR memberberries, but Tsar-era, oughta be instructive.
asciilifeform: in e.g. jp similar pov. is almost definitionally a thing in actual nations.
mangol: http://logs.nosuchlabs.com/log/asciilifeform/2022-03-16#1084913 >> agree. seems humans (esp. men) have honor hardwired. it's expressed to varying degrees. not sure whether non-human animals have any vestige of it.
dulapbot: Logged on 2022-03-16 18:30:47 asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2022-03-16#1084860 << honor is real, if 'not evenly distributed'. plenty of folx when rubber hits the road prioritize something other than own skin. whether you grow up hearing about'em depends heavily on yer culture, the anglo space is particularly rotten in this respect
asciilifeform: mangol: depends whatcha count. e.g. insect kingdom has 'die for motherland' 100% covered
verisimilitude: The US was founded by extremists and a new society can be built by the same.
asciilifeform: all new things built by 'extremists', verisimilitude
verisimilitude: I'll find a counterexample.
mangol: below replacement fertility is considered a 'moderate' political position in many places
mangol: among many other examples
verisimilitude: Does food count?
mangol: http://logs.nosuchlabs.com/log/asciilifeform/2022-03-16#1084911 << i got my info from the same comment as you, asciilifeform. did i misinterpret it? seems he was right to object to the tax, but couldn't let go of the matter
dulapbot: Logged on 2022-03-16 18:27:55 asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2022-03-16#1084807 << mno, not imho accurate summary at all; naggum's tax for ref.
verisimilitude: Let's change topic. Care to see an example of a koan from that collection I dislike?
← 2022-03-15 | 2022-03-17 → ↑
Random(asciilifeform) | Download hourly DB snapshot | Get Source Code