Hide Idle (>14 d.) Chans
← 2021-11-26 | 2021-11-28 → ↓
vex: I've never parken of the peruvian marchin powder, but around when bc was @ $20, I was msoking a gig out the back of a maquarie st watering hole.. heard about bitcoin?
vex: da what?
vex: bitcoint cunt
vex: davidson? yes please
vex: what are those walls strett ciggies called BingoBoingo?
vex actually can't brain
vex: I smoke jet's in vn. folks know i'm aweful
vex: all the indo chop for minimum dong
vex: as soon as they let me out, I'll need to smoke for about 10 year to get the dollar coast average down to something sensible
vex: china won
vex: wait, erryone needs taiwan
vex: sup tuttie?
atuttle: asciilifeform, did you ever get around to making that defritzing module? you know, the one that goes in place of the cr50. if not, it looks pretty easy; most of the pins can just be passed straight through. at most, maybe a CPLD to generate the 3v3 enable and the keyboard strobe
atuttle: vex, hi
vex: you're new here?
atuttle: on this IRC server? yes, I am. Archibald Tuttle, heating engineer, at your service.
atuttle extends hand
atuttle: we're all in this together.
vex shakes atuttle's hand.
atuttle: I have been reading loper-os.org for a while.
vex: you know you're approaching in a spooky manner right?
atuttle: in what sense?
vex: maybe it's just me. nevermind
vex: you're a seppo?
atuttle: the whole movie is great, you should watch it sometime.
vex: Who is your daddy and what does he do?
atuttle: (I don't know what a "seppo" is...)
vex: us citizen
atuttle: hrm, is you next question going to be a request for a social security number? ;)
vex: on second recolation; the name sounds famililar, go right through
atuttle: I also have a few questions about Pest, but it might be a while before I get around to trying it in a serious way.
atuttle: the main question being, isn't the "try to decrypt every packet using every key" a scalability concern? this was/is bitmessage's achilles heel.
vex: there's knob for that
atuttle: hrm, quick search of spec 0xFC for "knob" yields a few hits but none seem relevant... please explain?
atuttle: FWIW I read 0xFF completely but have only skimmed 0xFE and 0xFC
vex: skimming is likely to be your downfall
atuttle: it would be nice if the versions included a "changes" section to help people know what parts they need to re-read carefully
vex: don't be rude
atuttle: I don't consider it a substitute for thorough reading
atuttle: rude?
vex: don't ttell people what to do
atuttle: did I?
jonsykkel: it only has to do verify hmac for each peer for each incoming packet
jonsykkel: botleneck will most likely be deduplication of replayed packets
atuttle: jonsykkel, yes, O(N*M) for N peers and M packets...
jonsykkel: easy peasy
jonsykkel: unless plan on having 10k peers
atuttle: jonsykkel yes, that's the concern. although it could be considered a "feature not a bug" in that it makes it painful for people to create large "star topology" networks. I was wondering if the scalability problem was actually a deliberate design feature.
atuttle: I mean if freenode/libera.chat/whatever decided tomorrow that Pest was the future and they were just going to switch everything over to it immediately, well, I think they'd hit a brick wall with this N*M thing.
jonsykkel: i dont think this system is intended for unwashed masses
atuttle: well, I think it's pretty awesome, and the unwashed masses have a habit of adopting awesome things more suddenly than anybody expects.
jonsykkel: could very well be
atuttle: it's a bit disappointing to see all the energy being devoted to moving people from IRC to matrix.
atuttle: matrix is decentralized in theory, yes, but like web browsers it is so outrageously complicated and fast-changing that there will never be more than one or two complete implementations.
atuttle: pest doesn't have that problem.
jonsykkel: i think itis great, cleans things up
jonsykkel: indeed, xmpp syndrome
atuttle: also, pest doesn't bootstrap itself from DNS. that is a huge, huge advantage that is very hard to find in protocols these days.
atuttle: so, I would kind of like the unwashed masses to adopt pest. maybe nobody here wants that. if so, that's okay.
atuttle: i also like the "kelvin" versioning scheme. it implies that the protocol will someday (soon?) reach a state of being "final" and stop accreting features.
atuttle: just like TeX.
jonsykkel: yeah, protocol doesnt care where packets come from either so even ip only used to know where to send shit
jonsykkel: "cooling down to final crystalline form"
jonsykkel: was described like that somewhere
atuttle: anyways, if the n_peers*m_packets scalability obstacle was *not* deliberate, there are a few ways to fix it.
atuttle: the easy one is to include in the black packet the hash of the most recent red packet sent from that sending station to that receiving station.
atuttle: then the receiving station, by keeping a table of hash-of-most-recent-packet-received, can know immediately which key it should use to try to decrypt the next incoming packet
atuttle: this is an opportunistic optimization, so you still need to fall back to "try every key" when no match is found in the table. this can be exploited by DDoS attackers.
atuttle: if that is a problem, wireguard has a solution that can be adapted fairly directly.
vex: dafuq?
atuttle: on the other hand, if n_peers*m_packets scalability obstacle *was* deliberate, that's okay, but acknowledging the deliberate obstacle would be helpful to people trying to understand the spec. Also there might be better ways (I cannot think of any right now, but might in the future) to achieve the same goal.
atuttle: sorry to dump so much text.
atuttle: I may have to step away soon, and wanted to put that into the logs in case asciilifeform comes back later.
atuttle: sorry to run my mouth so much. I'll shut up now.
asciilifeform gotta afk shortly but will try to anwsrc coupla of atuttle's qs
jonsykkel: atuttle: wouldnt that only optimize for valid packets which ~never arrive anyway, not for ddos spam
asciilifeform: atuttle: it's deliberate. pest is designed for decentralization ruat caelum. and for total resistance to ddos. hence verifications are to take place in worst-case time entirely by design.
dulapbot: Logged on 2021-11-21 13:37:34 asciilifeform: the pill against the former is simply to be able to operate over N fallback links, w/out enemy necessarily having any idea of their existence. the pill against the latter is to actually have a protocol where a spamola packet is rejected no later than 1/t, where t is the interval the nic takes to receive the packet.
atuttle: jonsykkel, yes, it would speed up the processing only for valid packets. it would not reduce the CPU load caused by a ddos. I'm not sure what you mean by valid packets "~never arrive anyway" though.
asciilifeform: atuttle: it is very easy to speed up processing for valid packets -- simply use sender ip as a hint to associate with key. but asciilifeform ~specifically~ recommended agains this, for the above reason.
atuttle: asciilifeform, thank you for clarifying that it is deliberate.
atuttle: I humbly suggest including this clarification in the specification, but "it's your baby". Thanks for creating Pest.
asciilifeform: verifying in not-worstcase time makes it tricky to calculate exact station capacity.
asciilifeform: atuttle: imho is quite pedantically described in the spec.
asciilifeform: this (aside from being entirely unnecessary) would violate '1.2.8. nothing to the snoop.'
asciilifeform: i.e. folx w/out key have no biz learning ANYTHING at all from a black packet. other than the fact of its existence.
dulapbot: Logged on 2021-11-27 01:29:50 atuttle: i also like the "kelvin" versioning scheme. it implies that the protocol will someday (soon?) reach a state of being "final" and stop accreting features.
atuttle: asciilifeform, what would the snoop learn from the hash of the most-recently-sent red packet?
dulapbot: Logged on 2021-11-15 21:29:41 asciilifeform: a properly-adaized replacement-trb (i.e. with ZERO traces of derpcoad and demonstrably bug-free) would be what one could call 'permanent software'
asciilifeform: atuttle: potentially helpful for cryptoanalysis. i'ma not describe how, 'the space in the margin of this pg is too small'(tm)(r)
asciilifeform: atuttle: point being, asciilifeform deals in absolutes. when he says 'NO information to nosy parker w/out key' -- means it.
asciilifeform: no means no.
atuttle: (can anybody explain to me why "dulapbot" repeats randomly-selected comments of mine?)
asciilifeform: atuttle: they aint random -- it triggers on log links
dulapbot: Logged on 2021-11-27 01:45:53 asciilifeform: atuttle: they aint random -- it triggers on log links
atuttle: Oh I see; you posted a link to a log of something I said, so dulapbot spoke the comment you were linking to. Thanks, now I get it.
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2021-11-27#1067799 << making would-be star topology peddlers choke on own vomit is simply a pleasant side-effect.
dulapbot: Logged on 2021-11-27 01:25:05 atuttle: jonsykkel yes, that's the concern. although it could be considered a "feature not a bug" in that it makes it painful for people to create large "star topology" networks. I was wondering if the scalability problem was actually a deliberate design feature.
atuttle: yes, I knew the channel was logged, I just didn't make the connection between the bot speaking and log links being posted. Now I get it.
atuttle: asciilifeform, by the way, if you never got around to making the fritzchip-replacement PCB I might try it. I have a pair of gru-bobs and a scarlet which are fritzed, unlike my kevin. But if you already designed one and are willing to share it there is no sense duplicating the effort.
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2021-11-27#1067809 << no dns, no sslism, no wwwism, misc. liquishit. no central anything.
dulapbot: Logged on 2021-11-27 01:28:12 atuttle: also, pest doesn't bootstrap itself from DNS. that is a huge, huge advantage that is very hard to find in protocols these days.
asciilifeform: atuttle: re cr50 -- it's certainly physically possible, i've simply no interest in a cure that requires per-unit bga rework, it's uneconomical
atuttle: woah, the fritz chip is BGA? I thought it was QFN. are you serious?
asciilifeform: atuttle: fwiw i've long ago lost interest in the box (which as i understand is outta print for a while)
asciilifeform: atuttle: iirc bga
atuttle: crap
atuttle: wish I had boardviews+schematics for the gru series as well :(
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2021-11-27#1067810 << for all i know unknown folx already using. beauty is, they don't need to tell asciilifeform et al about it.
dulapbot: Logged on 2021-11-27 01:28:35 atuttle: so, I would kind of like the unwashed masses to adopt pest. maybe nobody here wants that. if so, that's okay.
dulapbot: Logged on 2021-11-17 19:27:29 asciilifeform: fwiw asciilifeform suspects that there were folx 'using pest' long before i described it on www, and before thimbronion wrote his pre-pest thing, or even before any of us were wanking re 'gossipd', etc.
vex: mebbe not. not your daddy's internet protocol
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2021-11-27#1067800 << approx. as likely as u.s. fed switching from printed toiletpaper money to btc, imho
dulapbot: Logged on 2021-11-27 01:25:43 atuttle: I mean if freenode/libera.chat/whatever decided tomorrow that Pest was the future and they were just going to switch everything over to it immediately, well, I think they'd hit a brick wall with this N*M thing.
asciilifeform: think, no place for censors, cia-funded 'swedish' 'foundations', 'consciousness raising' misc. liquishit.
atuttle: asciilifeform, if you asked me, in 2010, the odds that El Salvador would make BTC legal tender in 2021, I would tell you "0% odds"
asciilifeform: atuttle: 'no actual btc was harmed' in the making of the salvador circus.
atuttle: alright I should have picked a different example.
asciilifeform: atuttle: the ~total stagnation in p2pism in last 20y, interrupted only by single event (the publication of orig. bitcoin) , imho not accidental.
atuttle: I still maintain that cryptocurrency has exceeded our wildest expepctations from ten years ago.
atuttle: asciilifeform, well, yes, I agree with you there: outside of the "extended family" of cryptocurrency-related work, internet decentralization has been in decline wfor a while now.
atuttle: on the other hand maybe it was such a remarkable breakthrough that it simply attracted all the excellent people. like a powerful magnet. there aren't any fewer excellent people than before; they just all wound up working on the same thing in the early 2010s.
asciilifeform: anyways for all i know 1e6 pest users already. and there's absolutely no reason they'd 'stand up to be counted' here or elsewhere.
asciilifeform: there's simply no reason to.
asciilifeform: and that's the way it oughta be.
atuttle: agreed.
vex: people usually talk tho
asciilifeform: pest is not built to 'make rich' anyone, or 'famous', or for any such social butterfly scamcircuit garbage. built simply to ride a proper internet parasitically and unstoppably over the reich-coopted one.
asciilifeform must bbl but encourages atuttle to read log and comment ( link to logline being commented, and make use of the link syntax , makes for over9000x easier reading )
asciilifeform: [link][annotation].
vex embarrassed; to bed
atuttle: asciilifeform yes, that's a wonderful thing. I would just hate to see it become willdly successful only to run into a brick wall of "32-bit IP addresses are plenty" or "640kb of ram is enough for everybody."
atuttle: good night everyone.
asciilifeform: further re the O(peers * packets) thing -- factually it's O(packets), exactly like anyffin whatsoever to do with packets (yer looking at each one, or whatever % cpu capacity allows) given as in practice you'll want at least as many cpu cores as you have peers, to achieve line-rate (Gb/s or whatever yours is) liquishit-rejection
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2021-11-27#1067893 << asciilifeform is of the school of thought that there aint, nor will ever be on planet3, in fact even anything near 2^32 people who belong using the net.
dulapbot: Logged on 2021-11-27 02:01:56 atuttle: asciilifeform yes, that's a wonderful thing. I would just hate to see it become willdly successful only to run into a brick wall of "32-bit IP addresses are plenty" or "640kb of ram is enough for everybody."
asciilifeform: hence 0 support for ipv6 or similar heresies. anyffin with ipv6ism purporting to be a pest implementation is heretical and asciilifeform will not support. (can't prevent bullshit artists from perpetrating whatever atrocities -- but will not help)
dulapbot: Logged on 2021-11-27 01:57:47 atuttle: on the other hand maybe it was such a remarkable breakthrough that it simply attracted all the excellent people. like a powerful magnet. there aren't any fewer excellent people than before; they just all wound up working on the same thing in the early 2010s.
asciilifeform: virtually all of the work in the space has been -- and remains -- in perpetrating scams.
dulapbot: Logged on 2021-11-14 13:49:26 asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2021-11-14#1065677 << erry half-serious shitcoin has a 'lure'. for ethertards, the 'smart'; for tortards -- 'anonymism'.
dulapbot: Logged on 2021-11-14 11:56:58 asciilifeform: and given as the entire purpose of a shitcoin is to 'replay history' with perp at the center.
asciilifeform: ... while, in contrast, barely dozen people to date actually cared to try to bring the real thing up to adult standard.
asciilifeform: atuttle: if yer interested in pest, i rec to join the wot -- then possibly can peer with someone on the experimental pestnet.
signpost: imagine thinking you were escaping communism by defecting to the US.
signpost: poor thing
signpost starting a rewrite of the xor-a-tron today in common lisp, wishing to publish in a respectable language.
signpost: anyone started a pest in CL btw?
mats: pest is probably not suitable for malware c2 huh
mats: weird encrypted udp traffic is suspicious
signpost: seems like making one protocol appear to be another is an orthogonal problem
signpost: pest-a-tronics oughta be liftable in principle onto any other medium
signpost considers the "there is a place you are not" aspect of highly-visible pest traffic a feature, not bug, for the first use-case.
signpost: but there are certainly others.
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2021-11-27#1067910 << asciilifeform did. but don't let this discourage from baking own !
dulapbot: Logged on 2021-11-27 16:09:10 signpost: anyone started a pest in CL btw?
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2021-11-27#1067911 << you don't want symmetric crypto for these ( idjits do it anyway, then e.g. asciilifeform picks it apart and gets key and muhaha )
dulapbot: Logged on 2021-11-27 16:20:07 mats: pest is probably not suitable for malware c2 huh
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2021-11-27#1067912 << loox exactly like e.g. vpn traffic. but in principle one could masquerade as dns, say, or whatever you like.
dulapbot: Logged on 2021-11-27 16:20:37 mats: weird encrypted udp traffic is suspicious
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2021-11-27#1067913 << indeed is. stego is a++ but outside scope of spec imho.
dulapbot: Logged on 2021-11-27 16:32:21 signpost: seems like making one protocol appear to be another is an orthogonal problem
dulapbot: Logged on 2021-11-27 16:33:03 signpost: pest-a-tronics oughta be liftable in principle onto any other medium
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2021-11-27#1067788 << the spec is distributed as a vtree, imho is 'list of changes' quite enuff
dulapbot: Logged on 2021-11-27 01:19:23 atuttle: it would be nice if the versions included a "changes" section to help people know what parts they need to re-read carefully
← 2021-11-26 | 2021-11-28 → ↑
Random(asciilifeform) | Download hourly DB snapshot | Get Source Code