log☇︎
72000+ entries in 0.497s
asciilifeform: here we have another piece where i used a nail instead of a hinge.
a111: Logged on 2017-02-28 12:31 mircea_popescu: a seventh and final problem : you now constructed a chain of casks, on top of the blockchain. consider what happens if node fails to keep his ENTIRE history of casks EVER issued : http://btcbase.org/log/2017-02-14#1613884
mircea_popescu: but bitcoin is a lot more promiscuous about it. you want marriage, bitcoin does the "handjob in theatre / quick fuck alley behind bar" thing
asciilifeform: 'what if i know no nodes' was a problem there as well.
a111: Logged on 2017-02-28 12:24 mircea_popescu: a sixth, also major problem is that the system offers no serious guarantees to the user. think of some situations : a) i mined some bitcoin back in 2011, today i clean out the closed and boot up old laptop, to check out the tits and bits of ex gf who was back then hot and heavy into me. i jack off, and then snoop around and find a bitcoin wallet with ~500 btc in it. i go to spend it... and discover... that i know no nodes...
mircea_popescu: i think the notion is generallyt accepted, under the broad heading "bitcoin is a prototype"
asciilifeform: but to demonstrate to woman that she is unfit for husbands, and ought to proceed to cat accumulation, can be a valuable gift.
mircea_popescu: showing woman another light in which her current husband appears a shitbag does not give her a new husband.
mircea_popescu: yes, well, all it needs to show is that it is a problem. you don't get to unproblem it.
asciilifeform: or that it is a solvable problem any more so than heat death.
asciilifeform: so it is not clear to me that 'dust' would become a problem at any rate separate from the rate of heat death.
asciilifeform: aaah yes. wasn't this the earlier thread, where we determined that we already have a 'proton decay' going, and that the usable lifespan of a blockchain is -- likely -- finite ?
mircea_popescu: the (false!) theory that the bitcoin unit of account is the 0.00000001 btc or satoshi is based on the batshit insane theory that a system could ever be devised that would track 21 million hundred million individual 512byte outputs. ☟︎
asciilifeform: this'd be the 'deflation' thread again, neh. if it ever turns into a palpable problem -- add moar decimal places.
mircea_popescu: but the fixation on collecting the dust is quite akin to the fixation on "collecting the water" in a spaceship, or for that matter with the fixation on "retaining the atmosphere" on planet earth. venus, see, didn't so much care about its dust.
asciilifeform: unrecoverable dust is a fact of life, not only in bitcoin.
mircea_popescu: this is a very valid point.
mircea_popescu: asciilifeform suppose i have a wallet full of dust. how do i resolve the problem ? count dust / 2 txn in sqrt(count dust) tranches ?
asciilifeform: the fixed-width-tx is a provable component of any long-term-sane trb-i. regardless of what other parts are included or excluded. without it, you get rapid rot.
asciilifeform: as a 'proper ring signature'.
mircea_popescu: but it would ~seem~ that if you do this, then you should put in the work and make a ~proper~ ring signature scheme.
mircea_popescu: forcing all txn to be 2 in 2 out is a poor man's ring signature i guess.
mircea_popescu: there's a reason i said 2 in 2 out. anything else is either nonmoving (1-2) or else tree pruning (2:1 ends up with a single coinbase ; 1:2 ends up with an infinity of satoshi sized coinbases you might as well stop lying about and issue outright)
mircea_popescu: but not in a fixed width.
a111: Logged on 2017-02-28 12:08 mircea_popescu: a third, and rather major, problem is that you will have serious trouble creating fixed width transactions in general. the reason is that the amount of information itself varies ; you can pretend to push this inconvenience all the way to the user ("hey, always use two inputs and two outputs and fu!") but it's somewhat unlikely to ever stick.
a111: Logged on 2017-02-28 12:08 mircea_popescu: a third, and rather major, problem is that you will have serious trouble creating fixed width transactions in general. the reason is that the amount of information itself varies ; you can pretend to push this inconvenience all the way to the user ("hey, always use two inputs and two outputs and fu!") but it's somewhat unlikely to ever stick.
a111: Logged on 2017-02-28 12:14 mircea_popescu: a fourth, minor point, is that you have your market primitives ass-backwards. no price formation is ever envisaged to have "asks may depend on bids" as an edulcoration of "asks are fixed ; bids are fixed" because it is always possible to produce a pricing function for the product (based on cost) whereas it's never possible to produce a pricing function for the demand (demand is a psychological, not a physical phenomenon). the
mircea_popescu: see, a market is a thing of itself, not a sort of air that fills crevices you leave, but a sort of magnetic field, goes where it goes.
mircea_popescu: you can still have a market. "cost of spill vs cost of fill"
mircea_popescu: same way you have a market with price lists for restaurants.
asciilifeform: how do you have a market, with fixed fee ?
asciilifeform: a tx, recall, is finalized (from egg to larva) ~upon receipt of rolled cask~, which means that the immediate parent immediately knows if the fee suffices.
a111: Logged on 2017-02-28 11:59 mircea_popescu: there's a different, much less notable problem wrt to what constitutes "a rotten fill". if i go to the stock market (i mean the old, gentlemany, pit of hand gestures thing) and bid "sentiments" for a certain share, i am in fact engaging in retarded behaviour, on the level of sjwing, and will, and should be kicked out.
asciilifeform: http://btcbase.org/log/2017-02-28#1619756 << the 'rotten fill' thing was simply a restatement of 'a tx is not a tx unless it is valid, and a block is not a valid block unless it consists wholly of valid tx.' cribbed straight from classical bitcoin. phrasing was, evidently, confusing. ☝︎
mircea_popescu: understand, the principle is naturalia non sunt turpia. what this means is, that if the girl MUST go on her knees and cover her own face in semen while looking up at you, then THEREFORE it CAN NOT be "a shameful act" for a woman on her knees to cover her own face in semen while looking up.
asciilifeform: the design specifies that the rate of promise-breaking is measurable. which is a prerequisite to any wotronic approach to incentivizing promise-keeping.
mircea_popescu: this is not a merit but a dismerit of the design, as such.
asciilifeform: there is a reason why i specified 'fib' as a fact of life.
asciilifeform: he can self-fill a %.
mircea_popescu: i must ensure i get 1k filled casks for it ; but if nodes can return casks at any time, i will ALWAYS have a % of casks returned too late.
mircea_popescu: FOR ME however, and unbeknownst to them, this is becoming a crunch.
asciilifeform: at whichever level of the pyramid where the '4th roll' happens - is where a child's head rolls.
mircea_popescu: everyone gets a cask, thinks it fresh.
mircea_popescu: there isn't a "this is a 4th roll of this cask - must fill" field in your protocol.
mircea_popescu: (because as a miner you got this cask returned for the 5th time, whose head do you cut.)
asciilifeform: (cask'd be sent down to a child for an interval, if it fails to float, gets reissued)
a111: Logged on 2017-02-28 11:56 mircea_popescu: to demand miner makes a certified statement to the effect of "offset 0xa to 0xb in block y is reserved for txn 0xc" is putting a constraint on this p2p system the likes of which it can't conceivably bear.
asciilifeform: http://btcbase.org/log/2017-02-28#1619755 << later realized that the protocol ought to permit returning an unfilled cask to the parent, so he can reissue it to another child. this would make for a less unforgivably-tightly-coupled system ☝︎
mircea_popescu: it may well be argued that user is actually worst-fucked, as per 6.a or .b examples above.
mircea_popescu: "no, you're a bunch of orcs with delusions of agency in others."
mircea_popescu: "british empire decided women should be fucked -- we're a different empire, decide otherwise"
mircea_popescu: whatever attempt to "turn the cock against the man" will fizzle just as soon as man feels like it. the cockbearer is a matter of fact not a matter of convention.
asciilifeform: classical bitcoin is a massive cock, with the sharp end pointed at the node.
asciilifeform: second, while sleeping asciilifeform did realize that he had nailed down a few parts that ought to have been hinged, and hinged some parts that oughta have been nailed.
mircea_popescu: amusingly enough, the whole segwit thing could be best described as "a rather braindamaged attempt to implement half of stan's casks"
a111: Logged on 2017-02-14 17:51 mircea_popescu: trinque understand the problem with "I will also be piling up the signed material for my records" : if during your lifespan you manage to lose eg a 30yo hdd, because you've built a reviewable system those txn will be reviewable. are you on the hook now for refunds ?
mircea_popescu: a seventh and final problem : you now constructed a chain of casks, on top of the blockchain. consider what happens if node fails to keep his ENTIRE history of casks EVER issued : http://btcbase.org/log/2017-02-14#1613884 ☝︎☟︎
mircea_popescu: fenses a la "he didn't really do it, someone else did, and he's unloved!!1"
mircea_popescu: and if i try to meet some... they don't roll anything down... a day later finally someone rolls something down and so i put my txn in and it... doesn't get included. wtf is this shit ?
mircea_popescu: a sixth, also major problem is that the system offers no serious guarantees to the user. think of some situations : a) i mined some bitcoin back in 2011, today i clean out the closed and boot up old laptop, to check out the tits and bits of ex gf who was back then hot and heavy into me. i jack off, and then snoop around and find a bitcoin wallet with ~500 btc in it. i go to spend it... and discover... that i know no nodes... ☟︎
mircea_popescu: on the internet"), the closest thing is google-ish "advertising bidding" whereby google presents you with some details about a click and you have less than 300 ms to make a bid, which it matches and delivers ; b) is VERY likely to cause major delays. your system would, on the basis of pissguessing, require hourlong blocks or somesuch.
mircea_popescu: a fifth, major point is that the current ~ten minute blocks are predicated on loose coupling of "casks" (current situation can be deemed to be that all miners roll infinity of casks at all times to all comers, and all pubkeys are nil) and no negotiation whatsoever occurring either downward or backward. the dual-negotiation system you propose instead a) has never been tried in practice (which practically reduces to "can't work
mircea_popescu: a fourth, minor point, is that you have your market primitives ass-backwards. no price formation is ever envisaged to have "asks may depend on bids" as an edulcoration of "asks are fixed ; bids are fixed" because it is always possible to produce a pricing function for the product (based on cost) whereas it's never possible to produce a pricing function for the demand (demand is a psychological, not a physical phenomenon). the ☟︎
mircea_popescu: a third, and rather major, problem is that you will have serious trouble creating fixed width transactions in general. the reason is that the amount of information itself varies ; you can pretend to push this inconvenience all the way to the user ("hey, always use two inputs and two outputs and fu!") but it's somewhat unlikely to ever stick. ☟︎☟︎
mircea_popescu: moreover you even want it to invalidate blocks, which is certainly too much. how is a node to even know.
mircea_popescu: if however i offer half the "current price" for a share, i am not so engaging. moreover the case exists in history, for instance at the root of the rotschild fortune
mircea_popescu: there's a different, much less notable problem wrt to what constitutes "a rotten fill". if i go to the stock market (i mean the old, gentlemany, pit of hand gestures thing) and bid "sentiments" for a certain share, i am in fact engaging in retarded behaviour, on the level of sjwing, and will, and should be kicked out. ☟︎
mircea_popescu: to demand miner makes a certified statement to the effect of "offset 0xa to 0xb in block y is reserved for txn 0xc" is putting a constraint on this p2p system the likes of which it can't conceivably bear. ☟︎
mircea_popescu: there's a reason a man will fuck ~any woman he encounters naked or in any even vague kind of peri-nude state. without this property, reproduction'd have long ago failed. there's similarily a reason miners currently accept absolutely any txn into a pool of liquid shit.
mircea_popescu: there's a major problem with the very tight coupling your write-up may be read to envisage wrt cask filling. there are, as far as reality goes, two lists : a list of available txn slots in blocks, which is fixed and aforeknown, and a list of desired txn, which is neither fixed nor truly aforeknown. the workers moving the latter into the former (miners + nodes) are necessarily going to have serious trouble doing a very fixed a
mircea_popescu: asciilifeform re your piece (you know if you had intelligently published it ON YOUR OWN BLOG i could have left a comment there instead, and in a year or whenever this were to be reviewed we wouldn't have to go grepping the logs like idiots) :
mircea_popescu: wtf "nothing sexual happens". in a fucking theatre ? that's insanity. ☟︎
ben_vulpes: trinque: horny teen girls need a culturally sanctioned place to be horny in public, 'twas the best most of america had post y2k
ben_vulpes: i don't much know about paths out, but hatred is definitely the fuel for the torch i use to burn a moat around my tiny castle.
trinque: to anyone born in a cultural sewer, hatred is along every path out, even/especially the laziest one.
trinque: while harlan ellison is still on the mind, to a young trinque, his story was one of the first (yes) indications that total rejection of others was a thing.
trinque: I'll be up early with a fresh mind to read.
phf: trinque: it's double entendre in a musical form. she's singing about a girl who likes lollipops, but of course everyone know what the author (gainsbourg) ment. later she grew up a bit and had a falling out with gainsbourg once she learned the trick
mircea_popescu: 40s. 40s is fucking big band / andrews sisters and nobody giving a shit about getting the clap.
mircea_popescu: in the background, there's this steve martin - kathleen turner movie in which she's a black widow and he's a brain surgeon. woman has nice legs and a fine disposition ; what a fucking waste of a reel.
trinque: what a place
mircea_popescu: it's a wonder they're not in turtlenecks.
danielpbarron: trinque, it's supposed to take a url to the ascii armored key right?
mircea_popescu: that's about 2% of the federal budget, ie barely a ripple.
trinque: I'll be fixing !!key for a sec
trinque: do we know a full fingerprint for that nick?
trinque: there are a handful of those
asciilifeform: could take a day or 2.
asciilifeform: (and if a miner issues moar tickets than he has plane seats -- it becomes known)
mircea_popescu: anyway. taking away the miner choice wrt to what transactions are included would be a major step forward, goes on the list.
asciilifeform: (if there is a shortage of casks coming to me from upstream -- i raise my price for downstream. etc)
asciilifeform: with this algo, you can in fact market access to a miner. and n levels out.
asciilifeform: the terminal sig can belong to a node, rather than a miner, if miners prove resistant to signing anything
mircea_popescu: mno, miner ifnormally tells his nodes he wants txn ; much like lord of manner tells women of the house he wants a girl added.
asciilifeform: to rewrite formally, a cask is a C = sig_h1(sig_h2(sig_h3(.....sig_m('i'll mine a tx for q btc/kb...'))); tx author signs, author_sig(my_tx + C)
asciilifeform: at the bottom of the pyramid, a tx author requests a cask from a node, and fills it (assigns his node-and-miner fee to the caskchain specified by the hash he was given.)
asciilifeform: here is how it'd work, roughly. a miner generates rsa-signed messages , occasionally, let's call'em 'empty casks'. a cask consists of a declaration, 'i'll include a tx within N blocks, i promise, for Q btc/kByte; and oh, here's a nonce.' the casks are distributed to the next level of nodes away from the miner ;
a111: Logged on 2017-02-27 21:41 asciilifeform: actually i know an algo that does this. will post it later, if it isn't obvious to mircea_popescu et al after a few minutes' thought.
a111: Logged on 2017-02-27 22:07 mircea_popescu: now, the historical solutionb to the problem, as well as perhaps a workable solution here, is the intrinsic oracle. if user relays txn to a node WHO MAKES A PROMISE (such as for instance "the txn will be included before block n" ?) then the nodes can be scored by their oracle value ("what he said turned out true!) and suddenly you have a more meaningful node market.
asciilifeform: mircea_popescu: seems more of a http://btcbase.org/log/2017-02-25#1618577 neh ☝︎