Hide Idle (>14 d.) Chans
← 2018-08-24 | 2018-08-26 → ↓
deedbot: http://trilema.com/2018/and-today-in-fetlife-updates-sad-news/ << Trilema - And today in Fetlife updates, sad news.
deedbot: http://qntra.net/2018/08/san-francisco-to-pay-six-figure-usd-compensation-to-members-of-seperate-feces-and-needle-clean-up-teams/ << Qntra - San Francisco To Pay Six Figure USD Compensation To Members Of Seperate Feces And Needle Clean Up Teams
Mocky: http://btcbase.org/log/2018-08-24#1844275 >> why would you need initial guess for final 2 piles when you can compute them directly from given input?
a111: Logged on 2018-08-24 17:49 asciilifeform: all you gotta do is to take { floor(sum(mass) / sum(qty)) - 1 , floor(sum(mass) / sum(qty)) } as the initial guess for the split of qualities of the final 2 piles
Mocky: well, maybe my maffs is wrong and someone can correct my work
Mocky: http://btcbase.org/log/2018-08-24#1844275 >> let A = sum(mass) / sum(qty), final pile qualities will floor(A) and ceil(A)
a111: Logged on 2018-08-24 17:49 asciilifeform: all you gotta do is to take { floor(sum(mass) / sum(qty)) - 1 , floor(sum(mass) / sum(qty)) } as the initial guess for the split of qualities of the final 2 piles
trinque: BingoBoingo: *separate
BingoBoingo: trinque: ty fxd
Mocky: size of piles are computed by solving the two equations for x and y 1) floor(A)*x + ceil(A)*y = sum(mass) 2) x + y = sum(qty)
mircea_popescu: Mocky alf's mysterious disappearance promises he saw the problem with his idea :p
Mocky: other than not needing the guess the final state, I'm still trying to find a problem with it
Mocky: i think that it can work, i don't see how it would work better than A* if heuristic was found, but I don't see how worse either
mircea_popescu: not deterministic as claimed, the ~only problem.
mircea_popescu: otherwise sure works.
Mocky: I'm not clear on how you can backtrack piles directly though, as if they were from isolated branches, when possibly final 2 piles from mixed descent and not from isolated mixing
mircea_popescu: that's the problem!
Mocky: ok, lol
mircea_popescu: the addressable quanta here is the single unit, of a specified quality. these are presented in groups (of the same q), and you're asked to reform the "minimal groups", ie, 2 qualities. whether any individual strand belongs in final-group-A or final-groub-B is however ambiguous until you've made the last move (recursive pov, the last move is tha first move, recursion flows backwardsd, like dream states).
mircea_popescu: hence my http://btcbase.org/log/2018-08-24#1844303 question, it quashes the "deterministic" notion.
a111: Logged on 2018-08-24 18:19 mircea_popescu: specifically, how do i pick among alternatives 1. "all stacks with odd index, left, all stacks with even index, right" and 2. "all stacks under 75 or over 125 left, all stacks within 75 to 125, right" ?
asciilifeform: http://btcbase.org/log/2018-08-25#1844413 << ha yea i sat down to write the proggy and realized
a111: Logged on 2018-08-25 17:41 mircea_popescu: Mocky alf's mysterious disappearance promises he saw the problem with his idea :p
asciilifeform: it can be made to work ( stuff all the whole piles possib into the two lumps , then solve for linear combinations of the excluded q's) but dun win anyffing
asciilifeform: http://btcbase.org/log/2018-08-25#1844412 << canhaz proof ? ( i can buy it, but suspect that can construct counter ex )
a111: Logged on 2018-08-25 17:40 Mocky: size of piles are computed by solving the two equations for x and y 1) floor(A)*x + ceil(A)*y = sum(mass) 2) x + y = sum(qty)
asciilifeform: http://btcbase.org/log/2018-08-25#1844421 << i had it pictured in exactly these terms, looked for some cheat, but yet found none
a111: Logged on 2018-08-25 17:47 mircea_popescu: the addressable quanta here is the single unit, of a specified quality. these are presented in groups (of the same q), and you're asked to reform the "minimal groups", ie, 2 qualities. whether any individual strand belongs in final-group-A or final-groub-B is however ambiguous until you've made the last move (recursive pov, the last move is tha first move, recursion flows backwardsd, like dream states).
mircea_popescu: asciilifeform what do you mean ? that's exactly how it goes, atomically. all mixed piles will be of size sum sizes and quality w-a qualities.
asciilifeform: mircea_popescu: q-1 and q are not the only possible immiscible piles
mircea_popescu: q of diff parities are ; but in ~practice~ it is ~almost always~ the case that a "large enough" starter set is presented to that q, q-1 occur.
asciilifeform: there exist qa,qb that are only miscible in quantities that are not available on a given game board
mircea_popescu: but you have a point, in that investigating the "large enough" condition might shed some lights.
asciilifeform: i'ma post a proggy when i get home, currently in town
mircea_popescu: they have a town where you live ?!
Mocky: lolcat suggest some 'large enough' conditions, no?
asciilifeform: lol sorta
mircea_popescu: Mocky yes, kinda.
mircea_popescu: also, it is imo a great mistake to limit the modelling in terms of 2 pile starter conditions. this is very rarely the case.
mircea_popescu: anyway, let it be pointed out how lulzy and innerving the recurse-dream relation is. "can't make the first move till you made the last move" etc.
mircea_popescu: (and, for the esthlos of the present and they of the future, the "lie-tician" was a reference to lie groups, which is a sort of we-wanna-galois-too but with more symmetry, less comprehensibility and possibly much powerful-er overall. i half-expected he's going to start throwing manifolds at me or who knows what nonsense, take me out of commission for the remainder of the season.)
ave1: And to continue, It's possible to do this one in 20 steps: http://btcbase.org/log/2018-08-22#1843598, see http://p.bvulpes.com/pastes/o2Gzz/?raw=true
a111: Logged on 2018-08-22 16:49 a111: Logged on 2018-08-22 05:48 Mocky: http://btcbase.org/log/2016-08-20#1525884 ftr my algo does it in 40 steps preserving full value of 1,639,142: http://p.bvulpes.com/pastes/bvsMD/?raw=true
asciilifeform: ave1: observe also that your final state is exactly example of http://btcbase.org/log/2018-08-25#1844434 ! and counter-ex to http://btcbase.org/log/2018-08-25#1844412
a111: Logged on 2018-08-25 18:06 asciilifeform: there exist qa,qb that are only miscible in quantities that are not available on a given game board
a111: Logged on 2018-08-25 17:40 Mocky: size of piles are computed by solving the two equations for x and y 1) floor(A)*x + ceil(A)*y = sum(mass) 2) x + y = sum(qty)
ave1: Yes, the whole thing is not guaranteed to finish (I used different sorting algorithms and even got cycles)
mircea_popescu: how the heck can you get a cycle ? "quality" here approximates enthalpy, it only moves one way
mircea_popescu: (or rather : for something to be a move, quality-diversity in the universe decreases)
asciilifeform: mircea_popescu: you move into an immiscibility zungzwang, that's how
mircea_popescu: zungzwang, eh ?
mircea_popescu: (it's zug-zwang, but honestly the onomatopoetic value justifies the grammar breach)
asciilifeform: i prefer my grandfather's corrupted ver lol
mircea_popescu: "oh, it's so difficult to make computer games -- you spend millions and months to generate content, and then the players consume it in minutes"
mircea_popescu: fucktards. NOT ONLY do they do it wrong, but they ACTUALLY attempt to pass their failure for "how it is" and "move on".
ave1: lol
ave1: btw, retrying what I did and so far could not regenerate cycle, so for all could have been a bug
mircea_popescu: i expect.
asciilifeform: mircea_popescu, ave1 , Mocky , et al : http://p.bvulpes.com/pastes/JJtFX/?raw=true
asciilifeform: ^ simple plaything for evaluating moves
asciilifeform: ( does not solve the game, only displays possible next move from given state )
Mocky: asciilifeform, ave1's above is not in fact example of immiscible quantities: ave1's final state can be stacked down to q, q-1 by my algo in 2 moves
asciilifeform: Mocky: plz post the moves ?
asciilifeform: cuz my proggy dun see any moves from his final state
asciilifeform: oh hm nm
Mocky: http://p.bvulpes.com/pastes/0JK0r/?raw=true , elementarily, they are both odd qual...
asciilifeform: i was working from old coad
asciilifeform: (find-variants '((7846 . 173) (1592 . 177))) gets me (((0 . 1) ((1592 . 175) (4776 . 174)))) .
mircea_popescu: !!key ave1
asciilifeform: would still like a proof that 'in the end always 1 or 2 piles'
ave1: Aha, I refactored the code to end as soon as possible, turning that back I get the same answer too.
asciilifeform: ave1: oughta end only when 0 valid moves remain
Mocky: 3 piles must contain odd,odd or even, even pair therefore stackable
asciilifeform: ok sold
ave1: It does not matter if the two piles are mixed in the end (the loss will always be the same). But I'll have to start playing to get to why you'll want to this specific end state.
asciilifeform: ave1: contemplated puzzle asks for 0 loss
mircea_popescu: ave1 eulora is very slim margin, losing as much as 1% could take you out entirely. think in terms of closed systems, such as "what is acceptable oxygen loss in space station" or i guess http://trilema.com/2016/what-lasts-forever/ etc discussions of prices and value
Mocky: mircea_popescu, does ave1's example of stacking down to 7846x173q 1592x177q, actually violate any criteria, despite that it 'could' be stacked further, it could not be stacked down into fewer slots
mircea_popescu: i don't see a problem with it, from a gameplay perspective.
mircea_popescu: but it also is at variance with what was de-facto specified, though not seriously ever reviewed.
asciilifeform: i dun know enuff about the game, but is it possible that player might prefer e.g. 173/177 to 174/175 etc
mircea_popescu: i can't right off conjure a reason why he'd care.
Mocky: the fewer server roundtrips would be preferable, assuming they can be had
mircea_popescu: i mean, from the trader pov, debalanced count pairs are arguably slightly better (so 1 - 9999 perhaps better split than 5000-5000), as mocky points out ( http://trilema.com/2018/algorithmics-problem-seeking-experts/#comment-126524 , and as i say, this advantage is indeed very slught)
mircea_popescu: then again, from the crafter pov, debalanced stacks not preferable, because of considerations of interplay between item q and bp q in the final product, it'll be more work to calc proper bp q for two stacks apart than for two stacks consecutive
mircea_popescu: but this is also very slight indeed.
mircea_popescu: so in short, i don't see a practical reason to say ave1 's 7846x173q 1592x177q is not a legitimate final state, even if it doesn't match what we've been saying final state should be.
Mocky: from my pov, stacking down to *any* 2 stacks is the bulk of the value. but in my mining the stacks are going to converge to q q-1 eventually. and in fact large piles of q q-1 are sitting around waiting to eat incoming small stacks
mircea_popescu: Mocky right, this is a recurrent problem with extremely meta-stable attractors, i don't expect in the end it makes ANY difference that in an isolated state it produced weird double-evens.
mircea_popescu: or w/e, double-odds.
asciilifeform: mircea_popescu: puzzler might have all sortsa interesting applications, possibly all the way to e.g. defraggings in a hypothetical trbi
mircea_popescu: ask honest questions, and the value of the answers will never cease to amaze.
asciilifeform: mircea_popescu, Mocky , et al : can it be shown that in e.g. 173x+174y=S , x and y are unique ?
mircea_popescu: two consecutive numbers are always coprime!!1
asciilifeform: ( observe that elementarily if gcd(a, b) != 1 and | S , they aint )
mircea_popescu: asciilifeform that's the criteria, that 173 and 174 share no factors. if they do -- then x * factor +y and x +y * factor both work.
mircea_popescu: but the smallest possible factor being 2, two consecutive numbers are always coprime.
jurov: should be fixed
lobbesbot: jurov: Sent 17 hours and 27 minutes ago: <mod6> We're getting 504 & timeouts from the ML
asciilifeform: jurov: still seems to hang
asciilifeform: is the box under ddos or wat
jurov: no
asciilifeform: then odd, why would it hang
jurov: mailman was stuck, after restart it went fine from here
asciilifeform: it still dun load from my end
trinque: negative
mod6: just tried it 2x, each from a different spot. worked good for me.
mod6: lemme try a 3rd & forth
mod6: *fourth
asciilifeform: mircea_popescu, Mocky , et al : here's a lol : given as there is finite range of Q (minq .. maxq from initial board) and finite range of quantities (1 .. totalcount) , it is possible to enumerate all possible mixes, in a space-for-time trade, to speed up exhaustive search
mod6: yeah, seems to be ok. let's keep an eye on it.
asciilifeform: still dead from here.
mod6: ok weird.
mod6: now i've got one location where it's responding fine, and another that i'm getting 504.
jurov: no idea, will wait and see. ml should be eventually moved to the foundation box
mod6: yeah, i want to move the website to one of the foundation boxes as well.
mod6: i just gotta get the time available to do it. with any luck the current configuration will hold up for another month or so to allow me to get freed up a bit.
jurov: mod6 let me know when it's set up to serve static files, i'll at least copy the archive there
mod6: the website?
jurov: the /ml/ archive
mod6: it can serve static files
mod6: i have bunch of tarballs and things on there now that one can just easily download or whatnot.
mircea_popescu: asciilifeform that is true ; meet my friend mr ! and also http://btcbase.org/log/2018-07-22#1837153
a111: Logged on 2018-07-22 16:55 mircea_popescu: (something very similar was the original impetuus for mp asking ben_vulpes for candi [where is she btw ?] only to then discover after a whole day of basic-training-with-admirals that in fact the memory footprint for what i had in mind was ~infinite)
jurov: I'll install/configure apache on the pizarro-colocated foundation box. Want to prepare for mp-wp, anyone has handy their notes pls?
mod6: Ah, thanks jurov!
mod6: Oh you're gonna do wp-mp on there too?
mod6: Lemme dig up the one script that I have from ben_vulpes, the rest of the notes I think BingoBoingo has.
asciilifeform: mircea_popescu: store as '3d cube' of 1bit 'mix? yes/no'
asciilifeform: well, 4d.
mod6: script ^
mircea_popescu: 4d. think, just looking through my storage right now, i have a dozen or so SIX FIGURE stacks.
asciilifeform: mircea_popescu: will be also interesting to represent problem geometrically, as paths in this space.
mircea_popescu: and i have items with six figure qualities right now in my inventory.
asciilifeform: store sparsely ( the mixability is periodic across qty axis )
mircea_popescu: im not even the biggest, iirc danielpbarron got 100s of millions of whatever from his remarkable claim
asciilifeform bbl,meat
mircea_popescu: the question of HOW to store sparsely is a re-implementation of the original question asked, in haskell.
jurov: mod6: thanks, and what apache/mysql versions are used?
mod6: on there? I don't even think he has either setup on there.
mod6: afaik it's just running trb
mod6: lemme see
jurov: i know, but which ones should I emerge.
mod6: oh, you don't mean which are on stormy, but which to use for wp-mp
mod6: i have no idea.
mod6: why do we need that anyway?
mod6: or is that what is needed for mailman stuff etc?
jurov: We need at least apache.
mod6: Yeah, sounds good. apache2 should be fine.
mod6: BingoBoingo ben_vulpes trinque anyone else ^ ?
mircea_popescu: jurov mysql 5.560 / apache 2.4.16 know to work.
jurov: But i don't want to install newest one, it may pull gcc 6 and what not, does cuntoo has an apache version?
mircea_popescu: !!key nicoleci
jurov: mircea_popescu: aok
mircea_popescu: !!rate nicoleci 1 i own her.
trinque: mod6: what was the question?
mod6: curious as to what apache2 version is suitable for just normal website or one that would include wp-mp
mod6: and i think stormy, the foundation box was a gentoo setup by ben.
mod6: but I don't know if it's cuntoo-ish, or early cuntoo-ish.
mircea_popescu: jurov this was gcc 4.4.7
trinque: yep, I host with you, so whatever it is you provide worx.
mircea_popescu: trinque i don't think they know what they provide :D
trinque: lawdy
mod6: it's not the same box.
mod6: that's a different box, i'll check and see there on the shared environment. one sec.
jurov: no rush
mod6: ok shared environment is:
mod6: Apache/2.4.29
mod6: mysqld Ver 5.5.59
jurov: mod6 good. Please copy over /usr/portage from the shared box to a new dir (don't overwrite anything) and I'll go on from there.
jurov: bbl
mod6: ok will do.
mod6: wait
mod6: /usr/portage ?
mod6: you want the entire thing?
mod6: is that a good idea?
jurov: yes, whole. why, it's in the same switch, no?
mod6: yeah, it's on the same network. im not as worried about that, but i just thought that it might be bad to mix and match the portage files from what we have existing, to whatever the shared environment has on there.
mod6: maybe that's not really an issue.
mod6: dunno
jurov: no, mixing custom version of few select packages like apache is usually not an issue
mod6: jurov: alright, will have to take your word for it here.
mod6: i will throw the tarball of the shared environment's portage into your ~/
mod6: then you can just unroll it there (not overwriting the current one) and do what you need.
mircea_popescu: nicoleci : ssh-keygen -t rsa -b 4096 -C "nicole@MP's whorehouse"
mircea_popescu: it'll take a while to run. when it's done...
jurov: mod6 you can throw in /etc/make.conf and /etc/portage for good measure
mircea_popescu: !!key mod6
mod6: jurov: ok
mod6: jurov: make.conf lives in /etc/portage on shared env. will just grab the entire /etc/portage dir
mircea_popescu: nicoleci when it's done curl http://wot.deedbot.org/027A8D7C0FB8A16643720F40721705A8B71EADAF.asc | gpg --import ; and then item=`cat ~/.ssh/id_rsa`; echo $(curl -Ls -o /dev/null -w %{url_effective} -X POST -F "pastebox=$item" http://wotpaste.cascadianhacker.com -w %{url_effective}) ; it'll spit out a pastebin url, say mod6 <url> when it does.
nicoleci: yes Master
mircea_popescu: in the meanwhile, once you learn italian you could read http://aretino.letteraturaoperaomnia.org/aretino_la_cortigiana.html ; but until then i guess http://btcbase.org/log/2018-08-21#1843362 should make ok replacement for... whatever was the stupid shit you were reading ?
a111: Logged on 2018-08-21 15:00 mircea_popescu: http://www.michaelhaldane.com/kleinzaches.htm << shall link it.
nicoleci: gang leader for a day..
mircea_popescu: kids today!
mod6: ben_vulpes: I need to ask you a question about your excise-hash vpatch, I'll do so in #trilema-mod6
BingoBoingo: Most comprehensive mp-wp write up at the moment appears to be from lobbes http://blog.lobbesblog.com/2018/06/mein-kompendium-mp-wp/
BingoBoingo: Includes links to the other guides
mod6: jurov: ok both tarballs are now in your ~/ : The entire /usr/portage ~and~ the /etc/portage (which also included a sybolic link to /usr/portage/profiles/default/linux/amd64/17.0/no-multilib which I also dropped in there incase you needed it -- all of this is from the shared environment, UY1.
mod6: They've been chowned for ya, so you should be g2g. Let me know if otherwise.
mod6 looks
mircea_popescu: mod6 do me a favour and add the slut to the shared.
mircea_popescu: how does she get to her mp-wp admin from teh ssh ?
mod6: ok Sir, will get started on that now.
mod6: mircea_popescu: actually not sure. i'm pretty unclued on the entire thing.
mod6: my guess, ssh in, and do the what is needed from her ~/
mircea_popescu: you mean install one ?
mod6: once I have her account and other things set up from the script, BingoBoingo is going to take over and do the wp-mp setup.
mircea_popescu: that's nice.
mod6: he's gonna take some notes for us. this way im not so unclued next time.
mod6: alright, was able to decrypt nicoleci's message, contents look normal. about to try to run the script here...
mod6: mircea_popescu: what domain is she going to be using for wp-mp? (the script needs to know upfront)
mod6: brb, I gotta pull these ribs out of the smoker and eat 'em.
mircea_popescu: mod6 a damn, forgot. 1sec
mircea_popescu: mod6 what ns ? was it ns1.qntra.net ns2.qntra.net ?
BingoBoingo: mircea_popescu: It is indeed.
BingoBoingo: Let me know what the domain is and if any additional records are desired for the domain
mircea_popescu: "There was a problem with one or more of the NameServers you entered.This is likely caused by the entered NameServers not being created yet."
mircea_popescu: usure BingoBoingo ?
mircea_popescu: im getting bimbo.club ; apparently this is free ?!
BingoBoingo: mircea_popescu: Very sure ns1.qntra.net and ns2.qntra.net Problem could be unhelpful error message intending to complain the nameserver at present lacks a record for bimbo.club
mircea_popescu: jesus fuck, check out these idiots : advertise it at $1.89 ; charge $15. da fuck.
mircea_popescu: now i got the choice of either chasing a ~nothing or making some scammer's day ? pshaw.
mod6: ive got an emergency over here, and i have to leave.
mod6: BingoBoingo: nicoleci's keyfile is in /home/mod6
mod6: please go ahead and get started on that for me, i don't know when i'll be back.
BingoBoingo: mod6: ack
mircea_popescu: don't stress out about it, chick's slow anyway.
mircea_popescu: not to mention has a raftload of shit to do.
BingoBoingo: I've got some time before the Peruana gets off work
mircea_popescu: how's that coming along ? wedding soon ?
BingoBoingo: Not soon, but not out of the question
mircea_popescu: fancy that wonder!
BingoBoingo: It's taken some time adjusting to "apartment != alt-hotel", but time is a sort of one way valve
mod6: ok I'm back. my apologies.
mod6: so am I correct in that we're going to set up the account for 'bimbo.club' ?
← 2018-08-24 | 2018-08-26 → ↑
Random(trilema) | Download hourly DB snapshot | Get Source Code