Show Idle (>14 d.) Chans
← 2021-06-30 | 2021-07-02 → ↓
00:04 verisimilitude I just remembered I loathe making accounts for anything, so I rescind my request, mats.
00:11 mats didnt see your prev questions, no reason unless you have particular privacy interests
~ 16 minutes ~
00:27 billymg http://logs.nosuchlabs.com/log/asciilifeform/2021-06-30#1041901 << ty
00:27 dulapbot Logged on 2021-06-30 15:30:40 signpost: http://logs.nosuchlabs.com/log/asciilifeform/2021-06-30#1041859 << cl-graph + graphviz
~ 2 hours 27 minutes ~
02:54 * vex has been epicly saddened by the news of mp passing
03:01 mats just remembered D.CBSE after re-visiting mpex homepage
↖
03:01 vex forevermore, magestic pelicans on the wing will remind me. hanbot's eulogy is pretty fucking good
03:01 mats another sad who won't get paid out from the scammer
03:02 vex what was CBSE?
03:02 mats coinbase
03:06 vex what was it? i never had an mpex acct
03:07 mats and reminds me, did BB ever return the money taken from mike_c?
03:08 mats not cool, bud.
03:11 vex its a free market pal
03:12 mats he was rewarded for his ingenuity
03:14 mats it would never have occurred to me to appeal to mp's sadism and use that to coerce a massive personal payment
03:18 mats my only regret is not speaking sooner, calling out the ethically malleable individuals in tmsr, like d_c
03:22 mats staying mum when people you know are being robbed, utter shite behavior
03:22 vex didn't you make money on mpex?
03:23 mats notionally, my withdrawal was ignored
↖
03:24 mats jurov was right, and i was blind to what was going on
03:34 vex you're both right, look around at the castle
03:42 vex dont be greedy matty, it's dangerous
03:52 vex https://www.youtube.com/watch?v=_Kmh4BbJPz8
~ 30 minutes ~
04:22 vex https://www.youtube.com/watch?v=pMYWC0tMZsk
04:22 vex alf. it's onlt a few bytes per week
~ 1 hours 49 minutes ~
06:12 puman "what actually dies and is mourned when actual people die and are mourned isn't the actual people that were there ; but the imagined people that were not, ever, anywhere besides the mourner's own mind"
~ 31 minutes ~
06:44 vex why'd you drop the nk punman?
~ 41 minutes ~
07:25 vex I read it like pooman now
↖
07:26 vex https://www.youtube.com/watch?v=Hy8kmNEo1i8
~ 1 hours 21 minutes ~
08:48 puman http://logs.nosuchlabs.com/log/asciilifeform/2021-07-01#1042058 can't be arsed to auth with freenode. I've been thinking of it as plutoniumman
↖
08:48 snsabot Logged on 2021-07-01 03:25:58 vex: I read it like pooman now
~ 2 hours 31 minutes ~
11:20 puman just went through the saga of zx2c4
11:20 puman http://logs.nosuchlabs.com/log/trilema/2018-04-16#1800034
11:20 snsabot (trilema) 2018-04-16 zx2c4: mircea_popescu: ahh that ignorant and antiquated notion, that "key size implies security size". or do you think there will be some amazing GNFS-like algorithms that come out for ECC, requiring ECC to use absurdly huge keys in the same way as RSA?
11:31 puman figures that if key must be as long as msg for proven "perfect security", it follows that key smaller than message is not "perfect". and that all modern crypto is basically attempt to "show how a moderate amount of random key stream can be used to generate a lifetime supply of keys"
↖
11:37 puman Koblitz has some interesting papers on the subject http://anotherlook.ca/
11:40 puman "newcomers to the field would logically expect that the problems that are used in security proofs come from a small set of extensively studied, natural problems. But they are in for an unpleasant surprise. What they encounter instead is a menagerie of ornate and bizarre mathematical problems whose presumed intractability..."
11:52 puman I send young friends interested in "sekoority" to Koblitz, because they can only find meaning in credentialed authority, they wouldn't "trust" puman or ascii for that matter. Similar to zx2c4 fella, seems he even found a Koblitz paper by himself and came back to share with #t.
11:56 puman If they buy into "modern crypto is shamanism" premise, than interesting conversation is possible. If they don't buy, not much hope.
~ 51 minutes ~
12:47 puman The vermin being offended by Koblitz are also amusing http://www.wisdom.weizmann.ac.il/~oded/X/pmc-jon.txt http://www.wisdom.weizmann.ac.il/~oded/on-pmc.html
12:48 puman "I was shocked and dismayed that the AMS published Neal Koblitz's article without, apparently, any editorial oversight. As one who works in the field of `provable security', I vehemently disagree with most of Koblitz's points"
12:50 puman "Emotional and unfounded attacks against a whole research area and its individuals, as carried by Koblitz, are of no use."
~ 1 hours 26 minutes ~
14:16 thestringpuller http://logs.nosuchlabs.com/log/asciilifeform/2021-06-30#1042035 << iirc D.CBSE was a synthetic, and would convert either into equity of coinbase when it eventually IPO'ed (it did), or shares on MPEx if it ipo'ed there.
14:16 dulapbot Logged on 2021-06-30 23:01:31 mats: just remembered D.CBSE after re-visiting mpex homepage
14:20 thestringpuller http://logs.nosuchlabs.com/log/asciilifeform/2021-06-30#1042049 << rereading the mike_c case and speaking with mod6 years ago, the bank run on MPEx was sorta inevitable. I remember mod6 being really worried, but eventually got paid. starting to think not enough coins to pay out at the time - always suspected smickles did the same to me when he was my broker.
14:20 dulapbot Logged on 2021-06-30 23:23:40 mats: notionally, my withdrawal was ignored
~ 16 minutes ~
14:37 asciilifeform puman: this was in a ~much~ earlier thread, and 0 to do w/ zx2c4 afaik
14:37 snsabot (trilema) 2016-02-05 mircea_popescu: "What really bugs me, however, is how much publicity Koblitz has managed to get out of this. "
14:39 asciilifeform ( and while on subj, see also )
14:39 snsabot (trilema) 2016-06-01 asciilifeform: btw when i went down into the snakepit with several dozen renowned 'cryptographers' earlier this year, i asked a few folks about this.
14:39 asciilifeform asciilifeform actually met, in the flesh, an entire regiment of those loathesome fucks.
14:41 asciilifeform http://logs.nosuchlabs.com/log/asciilifeform/2021-07-01#1042065 << if in fact p!=np, something like current-day popular cryptoism (e.g. rsa) could be 'possible' (i.e. not doomed to certain death)
14:41 snsabot Logged on 2021-07-01 07:31:38 puman: figures that if key must be as long as msg for proven "perfect security", it follows that key smaller than message is not "perfect". and that all modern crypto is basically attempt to "show how a moderate amount of random key stream can be used to generate a lifetime supply of keys"
14:41 asciilifeform but in fact we aint any closer to it today than in 1978.
14:43 asciilifeform puman: koblitz stuck his cock in the wasp hive, there's a 'cryptographic komyooniti' very similar to e.g. the 'climate model' people
14:43 asciilifeform they live on delicious grantola and do not welcome boat-rocking any more than other, similar grant-eaters.
14:45 * asciilifeform has beaten the subj more or less to death in the old logs imho
14:45 snsabot (trilema) 2017-03-02 asciilifeform: 'based on these unproven conjectures, and one or more unproven conjectures on which our logic relies, and one or more circularly-defined conditions, this scheme is secure, for our toy definition of secure that we will not share'
14:45 asciilifeform ( and also, etc.. )
14:45 snsabot (trilema) 2016-08-02 asciilifeform: one of these bargains is that you cannot build a career as a university academic with 'use rsa, kthx, bye'
14:52 asciilifeform http://logs.nosuchlabs.com/log/asciilifeform/2021-07-01#1042060 << you can use any nick you like on dulapnet !
14:52 snsabot Logged on 2021-07-01 04:48:42 puman: http://logs.nosuchlabs.com/log/asciilifeform/2021-07-01#1042058 can't be arsed to auth with freenode. I've been thinking of it as plutoniumman
14:52 puman I'm aware Koblitz peddled ECCisms, even twerked on stage if he was being paid
14:52 asciilifeform of course he did
14:53 asciilifeform or would've been outta job
14:53 asciilifeform welcome to dulapnet , cgra !
~ 1 hours 46 minutes ~
16:40 thestringpuller asciilifeform: are leases still available? or are you booked full?
16:40 thimbronion cgra: indexed db of log links available: http://thimbron.com/2021/06/log-link-index/
16:40 asciilifeform thestringpuller: what are you thinking of getting ?
16:43 thestringpuller something small. going to rewrite blog as just flat files rendered via python, and wanted to migrate it to more trustworthy place
16:44 asciilifeform thestringpuller: i rec the rk then
16:45 asciilifeform thestringpuller: in fact i'm getting moar ips this wk (haggled for a while, finalized yest.) so will be avail. in not too long.
16:47 thestringpuller trying to do math via http://www.loper-os.org/?p=3557 but brain so burned out right now
16:47 thestringpuller Will likely just pay annually to make easy.
16:47 asciilifeform thestringpuller: there's a handy calculator for this.
16:48 asciilifeform anyways please read the instructions cover to cover, in particular the bit re: pgp.
16:48 thestringpuller doing that now. possibly thinking of colocating my FG
16:49 thestringpuller cause no place to plug it up now atm
16:49 asciilifeform thestringpuller: i'm hosting several colo'd FGs atm.
16:49 asciilifeform (incl. mine)
16:50 asciilifeform there is no extra charge for this (assuming you have one)
16:51 thestringpuller Yea it's still sealed and everything. It is the collector in me no doubt.
16:52 * asciilifeform hilariously, still gets spamola from the shop where i had FG baked, asking why not moar, 'reorder?? plz??'
16:53 thestringpuller asciilifeform: qq. if I wanted to do OTP from the colo using FG as my RNG. would i encrypt the rng payload I send to myself to prevent interception? sorry if this is non-sensical question.
16:54 asciilifeform thestringpuller: didja read the link earlier re: what otp is ?
16:56 asciilifeform thestringpuller: when you generate pads, you do it with own hands, at home; then you transport the pads (not the generator..) to the places where they shall be used; then they are used, in whatever increments (historically, pages) which permit burning after use.
16:56 asciilifeform in modern times this takes the form of e.g. 2 hdd's, filled w/ rngolade, and each block is erased after use
16:57 thestringpuller Okay then that answers questions. "The OTP should consist of truely random characters (noise).
16:57 thestringpuller But didn't realize "only generate in physical location"
16:58 thestringpuller perhaps 6th rule should be added to https://www.cryptomuseum.com/crypto/otp/index.htm
16:58 asciilifeform thestringpuller: if you had a magical rng which -- while being trng -- exists in 2 and strictly 2 synchronized copies somehow -- you would not need to transport pads. but no such item exists (or is likely to)
16:58 thestringpuller "Only generate in physically present location."
16:58 asciilifeform thestringpuller: the 2 pads necessarily have to be generated in same location, think about it
16:59 thestringpuller Makes sense now in context of spycraft.
16:59 asciilifeform ( there's a variant devised by asciilifeform , (though almost certainly existed historically, simply 0 mentions in public lit) where indep. generated, but still requires a meeting to trade pieces )
16:59 dulapbot (trilema) 2018-10-23 asciilifeform: Mocky: it's an old idea of asciilifeform's -- otptron gets 2 sd slots. fill switch triggers fill-up of both with identical otp. then you fly to bananistan with ~one~ and trade with the other fella for his. then you both have identical xor of pad-a and pad-b, in the respective slots.
17:00 thestringpuller Makes more sense now: "The major problem with OTPs, is their distribution. A unique set of OTP booklets needs to be issued and distributed to each individual spy or agent abroad. As the OTP was destroyed immediately after use, sufficient and timely supply of new OTPs had to be guaranteed.
17:02 asciilifeform thestringpuller: owning a working trng solves ~one~ traditional difficulty w/ otp; but certainly not all.
17:10 thestringpuller Makes more sense now: "The major problem with OTPs, is their distribution. A unique set of OTP booklets needs to be issued and distributed to each individual spy or agent abroad. As the OTP was destroyed immediately after use, sufficient and timely supply of new OTPs had to be guaranteed.
17:10 thestringpuller ugh sorry for spam
17:10 asciilifeform lol
17:16 asciilifeform !w poll
17:16 watchglass Polling 16 nodes...
↖
17:16 watchglass 185.85.38.54:8333 : Could not connect!
17:16 watchglass 84.16.46.130:8333 : Could not connect!
17:16 watchglass 185.163.46.29:8333 : Could not connect!
17:16 watchglass 205.134.172.26:8333 : Alive: (0.082s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Blocks=689319
17:16 watchglass 108.31.170.100:8333 : (pool-108-31-170-100.washdc.fios.verizon.net) Alive: (0.105s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Blocks=689319 (Operator: asciilifeform)
17:16 watchglass 205.134.172.4:8333 : (172-4.core.ai.net) Alive: (0.082s) V=70001 (/therealbitcoin.org:0.7.0.1/) Jumpers=0x1 (TRB-Compat.) Blocks=689319
17:16 watchglass 54.39.156.171:8333 : (ns562940.ip-54-39-156.net) Alive: (0.173s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Blocks=689319
17:16 watchglass 205.134.172.28:8333 : Alive: (0.084s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Return Addr=0.0.0.0:8333 Blocks=689319 (Operator: whaack)
17:16 watchglass 192.151.158.26:8333 : Alive: (0.205s) V=70001 (/therealbitcoin.org:0.7.0.1/) Jumpers=0x1 (TRB-Compat.) Blocks=689319
17:16 watchglass 208.94.240.42:8333 : Alive: (0.160s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Blocks=689319
17:16 watchglass 143.202.160.10:8333 : Alive: (0.270s) V=70001 (/therealbitcoin.org:0.7.0.1/) Jumpers=0x1 (TRB-Compat.) Blocks=689319
17:16 watchglass 176.9.59.199:8333 : (static.199.59.9.176.clients.your-server.de) Alive: (0.248s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Blocks=412062 (Operator: jurov)
17:16 watchglass 54.38.94.63:8333 : (ns3140226.ip-54-38-94.eu) Alive: (0.262s) V=88888 (/therealbitcoin.org:0.8.88.88/) Jumpers=0x1 (TRB-Compat.) Blocks=689319
17:16 watchglass 213.109.238.156:8333 : Alive: (0.328s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Blocks=689319
17:16 watchglass 103.36.92.112:8333 : (terebe.ns01.net) Alive: (0.665s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Blocks=689319
17:16 watchglass 205.134.172.6:8333 : (172-6.core.ai.net) Alive: (0.377s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Return Addr=0.0.0.0:8333 Blocks=689186
17:26 cgra asciilifeform: fwiw, finally bothered with a bouncer, too
17:27 cgra thimbronion: thanks! will check it out sometime soon(ish)
17:28 asciilifeform a++ cgra . keep in mind that this is a temporary item, to be replaced w/ arsenet. (but not w/out warning)
17:29 asciilifeform ( thread )
17:29 dulapbot Logged on 2021-06-18 18:35:02 asciilifeform: the troo p2p topology i propose removes all kindsa fundamentally palace-flavoured concepts -- 'joining', 'kicking', 'banning' -- and replaces simply w/ freedom of association, i.e. peering & unpeering.
17:35 cgra asciilifeform: yeah. i tried to briefly study (from rfc's + some) how do irc servers intercommunicate user messages. then you said this :)
17:35 dulapbot Logged on 2021-06-28 10:16:54 asciilifeform: has been reading the src of 'ratbox', thinking 'so how does it deduplicate messages..?', answr is, 'doesn't'
17:38 asciilifeform cgra: apparently irc has this subtle braindamage baked in at algo level. and is, evidently, why all the great irc nets eventually not only schismed, but acrimoniously schismed and turned into palace-reichs and eventually burn down
↖ ↖ ↖
17:39 asciilifeform see e.g. this history re subj.
17:43 asciilifeform cgra: orig. thrd where asciilifeform summarizes proposal for 'return to 1980s irc'.
17:43 dulapbot Logged on 2021-06-18 15:31:54 asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2021-06-18#1039833 << i may've let this slip, oughta answr directly -- what i'm proposing is to depart from last 25yrs of internet decay, where errything came to consist of 'herd of cattle' and 'farmers' to corral, milk, and occasionally devour the former; and instead experiment where people stand up on own legs.
17:44 asciilifeform ( spoiler : probably not trivially sewable outta existing irctrons after all )
~ 25 minutes ~
18:09 shinohai $vwap
18:09 busybot The 24-Hour VWAP for BTC is $ 33425.35 USD
18:10 asciilifeform reminds me, gotta add busybot to the botlist
18:11 asciilifeform 1s
18:11 asciilifeform ..done
~ 1 hours 6 minutes ~
19:18 shinohai tyvm asciilifeform
19:19 asciilifeform shinohai: currently all it does is to set'em to be coloured correctly
~ 1 hours 25 minutes ~
20:44 gregorynyssa /quit
20:48 * gregory4 had missed the original message about `dulapnet`, finally here.
~ 20 minutes ~
21:08 gregory4 is anyone here excited for the upcoming Dune adaptation by Villenueve?
21:15 mats https://cdn.b12.io/client_media/2nMpL8WQ/2b43d97c-c061-11ea-a23f-0242ac110002-Cryptodollars_Castle_Island_Ventures_2020-07-07.pdf
21:15 mats https://f.hubspotusercontent00.net/hubfs/6024551/Genesis%20-%20Quarterly%20Reports/Genesis21Q1QuarterlyReport-H%5B2%5D.pdf
21:27 asciilifeform welcome to dulapnet, gregory4
21:28 asciilifeform mats: loox like umpteenth permissioned-altcoin-paypalism crapola
21:28 asciilifeform mats: or is there sumthing unusual about this one?
21:35 mats its a systemic review of the space
21:36 asciilifeform mats: how wouldja describe -- which -- space ?
21:36 mats crypto-eurodollars
21:38 asciilifeform mats: you do understand that nobody would simultaneously bake a decentralized altcoin ~and~ honestly offer to 'back' it with whatever xyz, right?
21:38 asciilifeform typically 1, the other, or both, is a straight out lie
21:38 mats you understand i've been around almost as long as you, right?
21:39 asciilifeform mats: aha, so i'd expect you would know this. so whence the interest in subj ?
21:39 asciilifeform simply entomological?
21:39 asciilifeform mats: pretty sure you were 'tuned in' longer than me, iirc
21:40 mats some parts entomological, some parts exploratory, this is the 'kyctronic-less' fiatola you've wondered about b4
21:41 mats i'm curious about the risks that stablecoins might pose to fiat, and where the levers are
21:41 asciilifeform mats: in 19th c usa, recall, there were various non-usg-issued currencies; possibly will play out similarly
21:41 mats yup.
21:42 mats stablecoin issuers are now buying as much short dated commercial paper as the largest money market funds
21:42 mats i'd like to know where i can pour the petrol
21:44 asciilifeform i suppose that for folx w/ what to pour, it is an interesting puzzle
21:44 * asciilifeform not spent time contemplating, rather like does not spend much on naval strategy, has no navy
21:45 asciilifeform welcome to dulapnet, bingoboingo !
21:45 bingoboingo Made it to this side, will fuck with the bouncer later
21:46 asciilifeform i think this leaves only puman
21:46 asciilifeform ( and jurov, supposing he's alive )
21:46 * bingoboingo sent an email to jurov
21:47 asciilifeform currently this is the most barbaric possible arrangement, there is no autofallback. if asciilifeform and/or his box get flattened by asteroid -- plox to manually walk to signpost(trinque)'s box
21:47 asciilifeform ( see his www re where )
21:55 gregory4 asciilifeform: any interest in bridging "dulapnet" with "xrsenet?" that way, you can use "unreal" and avoid the 15 second issue with "ratbox."
21:55 asciilifeform gregory4: afaik 'unreal' won't bridge w/ anyffin other than self
21:56 asciilifeform gregory4: it sucks mightily, tries to force sslism, and is massive ball o'shit , even heavier than 'ratbox'
21:56 asciilifeform selected it simply because it had a working dulap-gentoo port
21:57 gregory4 asciilifeform: you were able to turn off SSL though, right?
21:57 gregory4 I don't think we are on SSL right now.
21:57 asciilifeform gregory4: i haven't tried server-server bridging with it at all
21:57 asciilifeform my understanding is that it mandates sslism for it, and would have to be patched to even try
21:57 gregory4 asciilifeform: oh, it only requires SSL for peering.
21:57 gregory4 what a strange decision.
21:58 asciilifeform the more pressing issue imho is this
21:58 dulapbot Logged on 2021-06-28 10:15:42 asciilifeform: 'If a SERVER message is parsed and attempts to introduce a server which is already known to the receiving server, the connection from which that message must be closed (following the correct procedures), since a duplicate route to a server has formed and the acyclic nature of the IRC tree broken.
21:58 asciilifeform i.e. that there simply aint a usable ircd
21:59 shinohai wb bingoboingo
21:59 gregory4 asciilifeform: the "ircd" which I am currently using on "xrsenet" is my own implementation.
21:59 gregory4 I might try adding a stateful deduplicator.
21:59 asciilifeform gregory4: is this item posted somewhere?
22:00 gregory4 asciilifeform: not yet, but I am open to the idea.
22:03 gregory4 assuming that you are able to turn off mandatory SSL for peering of "unreal," I could serve as bridge between "dulapnet" and "xrsenet" for a while.
22:03 gregory4 obviously, this is not a long-term solution.
22:04 asciilifeform gregory4: i'm looking for shortest path to '1 talker -- 1 relay'.
22:04 asciilifeform just about anything else is essentially same thing as what we have now, i.e. single point of failure
22:04 gregory4 asciilifeform: that path, IMO, is IRC with a stateful deduplicator bolted on.
22:05 asciilifeform gregory4: can you elaborate re 'stateful deduplicator' ?
22:06 gregory4 server saves last N messages received from peers. refuses to pass on identical messages.
22:08 asciilifeform i.e. 1 nick can't utter e.g. 'yes' or '$vwap' etc within <N min ?
↖
22:09 gregory4 asciilifeform: that is a side-effect, yes. same person cannot say same thing within N message-cycles.
22:09 asciilifeform not only this, but different nodes may have, at a given time, different states
22:09 asciilifeform and thereby disagree re what's a dupe
22:10 gregory4 doesn't matter IMO, because the goal is to prevent cycles.
22:11 asciilifeform tbh i'd rather implement this as a udpism w/ irc frontend for operator , and dedupe based on a random nonce field in ea. msg
22:11 dulapbot Logged on 2021-06-23 09:12:36 asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2021-06-23#1040513 << for one thing, if yer 'something else' doesn't have at least irc-compat. front-end, you're then also signing up to write errybody a new client. for all known os and ui paradigms. goodluck w/ that
22:11 asciilifeform (with this as bonus)
22:11 dulapbot Logged on 2021-06-30 15:36:08 asciilifeform: wb signpost . asciilifeform had a minor thought, 'ddos-proof udp can be baked with btc-like bruteforce 'must have n zeros in hash' nonce' even w/out actual crypto
22:11 asciilifeform because it aint clear to me that this'd be necessarily more complex than fixing an ircd as described earlier
22:12 gregory4 asciilifeform: your product would basically listen on two ports: one for peers, and one for clients.
22:12 gregory4 whereas IRC uses the same port for both.
22:12 asciilifeform aha
22:13 asciilifeform irc, there, simply to plug into existing client
22:14 gregory4 the BTC-like brute-force requirement is basically Adam Back's old idea for email-filtration, which inspired Nakamoto Satoshi.
22:14 * asciilifeform was laying this out on chalkboard, and currently stuck on the puzzler of 'how do peers know 'the current difficulty' ?'
↖ ↖
22:14 asciilifeform gregory4: aha, 'hashcash', see thrd
22:17 gregory4 the way to know the current difficulty is to have a chronically updated chained ledger, where shorter chains are discarded when a longer one appears.
22:17 gregory4 the other way is simply for each peer to set his own difficulty in his configuration-file, and risk ostracism of other peers, if set too high.
22:17 asciilifeform gregory4: this doesn't map to this problem, sadly. you gotta be able to drop enemy packets (incl. replays of your own, valid ones) into /dev/null ~immediately~
22:18 asciilifeform without storing, necessarily, anything
22:18 asciilifeform gregory4: peer can't 'set own difficulty' because this would require answering packets which come in w/ insufficient pow
↖
22:18 asciilifeform they MUST NOT BE answered
↖
22:18 asciilifeform in any way.
22:18 asciilifeform or yer carrying your own cross
22:18 asciilifeform i.e. helping the ddoser
22:18 gregory4 you are building a mountain-fortress.
22:19 asciilifeform hm?
22:21 gregory4 how about this: just don't reply with "insufficient PoW" if difficulty is not met.
22:22 asciilifeform gregory4: and if met? then must reply?
22:22 asciilifeform 1 of the appeals of udp is that you can operate 'simplex'
22:22 gregory4 let every peer manually troubleshoot his connection to every other peer. it is a one-time effort (per peer).
22:23 asciilifeform until somebody gets flooded and then you have nfi what happened
22:23 gregory4 set the difficulty to X. then type "hello can you hear me?" if no reply, then set difficulty to X + 1.
22:23 asciilifeform say he tripped over his mains cable. and now ~your~ cpu to its knees ?
↖
22:24 * asciilifeform realizes that the problem, as posed, may not necessarily have a solution
↖
22:24 gregory4 I have barely worked with UDP.
22:25 gregory4 assuming that maximum message-length is < MTU, how do you plan to confirm that message was received anyways?
22:25 asciilifeform gregory4: there is absolutely nothing particular about udp, it simply is 'postcard' w/ N bytes payload.
22:25 gregory4 AFAIK, where UDP is used in industry, such as video-games, a subset of TCP is reimplemented.
22:25 asciilifeform gregory4: you don't need to confirm.
22:26 gregory4 asciilifeform: then how do you live with the uncertainty that every time you send a chat-message, there is a 5% chance of it not going through?
22:26 asciilifeform gregory4: there's a thread in #t, where asciilifeform tested udptron across oceans w/ various folx, packet reception was ~100%
22:27 gregory4 asciilifeform: I didn't know that. nonetheless, you better be sure. even 99% is unacceptable IMO.
↖
22:27 asciilifeform gregory4: given this, plus the fact that w/ 'flood routing' one can expect to receive just about every message N times -- imho no reason to send ack's or anything of the kind
22:28 asciilifeform (exercise for reader: by what amount does sending an ack change the probability of total loss of a message?)
22:28 gregory4 better be between two and three sigma.
22:29 asciilifeform gregory4: imho suffices to ask to get the packet you sent, echoed back.
22:30 asciilifeform (and this way the receiver doesn't need to solve another pow simply to reply to you)
22:30 asciilifeform however this would mean that dest.ip cannot be included in the hashed-over payload
22:30 asciilifeform (potentially making replays easier)
22:31 gregory4 asciilifeform: right, I see. echoing doesn't require memory-allocation.
22:31 asciilifeform aha
22:32 * asciilifeform wishes to encourage all tuned-in folx to contemplate this puzzler and write in.
22:39 gregory4 another way is to use network-history to seed a PRNG, and use that to give permission to a particular peer to determine the difficulty of the next epoch.
22:39 gregory4 however, he must make his choice within a window of N seconds, or the PRNG is re-seeded.
22:40 gregory4 this is basically how PoS blockchains work.
22:40 asciilifeform gregory4: you haven't any guarantee of a shared history or of particular propagation speeds, though
22:40 asciilifeform there is not necessarily 'a network'
22:40 asciilifeform simply a bunch of nodes, coming and going
22:40 gregory4 asciilifeform: it doesn't need to be our own network-history. we could use hashes of recent Bitcoin blocks, for instance.
22:41 asciilifeform ugh now bringing in one of the least reliable things on satan's green earth as external dep ?!
↖
22:41 asciilifeform as in what, i can't talk unless i'm near a trb box that's reliably eating blox?
22:41 gregory4 asciilifeform: yep.
22:42 asciilifeform fuck that.
22:42 asciilifeform observe that trb nodes don't ever all agree re what is latest block.
22:42 dulapbot Logged on 2021-07-01 13:16:40 watchglass: Polling 16 nodes...
22:43 asciilifeform no external liquishit plz.
22:43 asciilifeform even having to sync clocks is, imho, lethal.
22:44 asciilifeform gregory4: centralization is like the vacuum of space, if we were building a space vehicle : it will always 'want' to penetrate. through ANY pore you allow it.
22:44 dulapbot Logged on 2021-07-01 13:38:53 asciilifeform: cgra: apparently irc has this subtle braindamage baked in at algo level. and is, evidently, why all the great irc nets eventually not only schismed, but acrimoniously schismed and turned into palace-reichs and eventually burn down
22:45 asciilifeform or seawater, in submarine, etc
22:45 asciilifeform it is always, always looking for a hole through which to go.
22:45 asciilifeform and then it goes in, and EVERYTHING you did was for nought, turns to shit.
22:46 asciilifeform and you end up, at best, w/ mp's little reich
22:46 asciilifeform or with the original irc, its fleanodes, etc.
22:46 asciilifeform fuck that. all of it.
22:47 asciilifeform decentralization o muerte, venceremos.
22:47 gregory4 venceremos = we will conquer?
22:48 asciilifeform gregory4: aha (cribbed, of course, from the cubans)
22:49 asciilifeform it is (was?) on their coins, 'patria o muerte..'
22:49 gregory4 I haven't studied Spanish, but a few days ago, I learned that ahogado means drowned.
22:49 asciilifeform point being, there are things on which there cannot be compromise, or better to give up straight off.
22:50 asciilifeform gregory4: lolyes
~ 17 minutes ~
23:07 shinohai $uptime
23:07 busybot The bot has been up for: 2 days 1 hours 2 minutes and 46 seconds
23:07 shinohai For some reason bot is 1000x faster @ responses here than EVER on fleanode.
← 2021-06-30 | 2021-07-02 → ↑

Random(asciilifeform) | Download hourly DB snapshot | Get Source Code

Theme: Light | Dark | System Default