log☇︎
800+ entries in 0.249s
phf: neither current wot nor gossipd spec wot have the history component, because if you go by "there's no rating outside of rater" past ratings make absolutely no difference.
trinque: whereas I see them as distinct; wot is the history of my past and present gossipd connections, and indications of what I thought of them
trinque: I could see an argument that the WoT evolves into the gossipd graph.
trinque: it becomes hard to distinguish from the gossipd-graph, eh?
trinque: asciilifeform: release gossipd already so I can write a DHT for it !1!!1 solves WoT, solves "DNS", solves ...
danielpbarron: this is counter the the gossipd model of "i heard so-and-so said whatever"
trinque: under the present scheme I could do the same, but have to trust the wire to be honest, or find a better wire, which proceeds towards dragging gossipd in as a dependency
mircea_popescu: my mind's eye foresees a time when all comms will consist of this. gossipd over irc!
Framedragger: re tmsr dns, in principle don't see why not - it could just as well just update quite often. but not sure if gossipd is taken into account (but then not sure about ip stack either, so..)
gabriel_laddel: Framedragger: if you want to do something useful, build the CLIM web I've described inbetween here and gossipd.
gabriel_laddel: mircea_popescu: 1. he can fly 2. he can do something useful for tsmr~ == get the CLIM-web thing up and running inbetween here and gossipd
gabriel_laddel: Rather than debating the fine points of gossipd, just implemented it and move on.
Framedragger: (ideally, i would guess, it'd munch datagrams, too, because gossipd is to have no stateful connections of the tcp kind.)
gabriel_laddel: ^ exactly. then we just wait for someone to get annoyed enough with the backend that they swap it out with gossipd.
Framedragger: "A0 can be implemented by regarding an authority's PGP public key as being its public routing address!" << nice kademlia and/or gossipd vibes
trinque: mircea_popescu: no. once I know a key I can interact with deedbot if I can make it there over gossipd peerings
Framedragger: asciilifeform: how about: make a proof of concept name system, use it instead of current dns root server set for now, later enable every gossipd user to run their own instance of name system if they prefer the fully-decentralized-dictionary path; the initial PoC will still have been useful.
mircea_popescu: feel free to query it over gossipd, or over whatever else.
a111: Logged on 2016-11-15 18:35 trinque: one central table of all things seems entirely anti-gossipd
Framedragger: so, yeah. and i'm no longer convinced it would redundant, in the sense that when gossipd cometh, one must throweth the 'general name system' away
Framedragger: trinque: right, this is cf. gossipd's "everyone has their own a la hosts file, and does with it what they like". is that what you meant?
trinque: one central table of all things seems entirely anti-gossipd ☟︎
trinque: describe "wrapped in gossipd" ?
Framedragger: re. plaintext for all NSAs to read, sure, there's that. goes against gossipd's "no free bits for the unauthenticated" i guess. but again, this can be wrapped in gossipd later.
Framedragger: asciilifeform: do note that mircea_popescu's idea of keeping dns is more akin to a general WoT-enforced hashtable, update-able via (in principle) gossipd-compatible pgprams, and (for the time being) transportable over dns/udp. the latter so that dns clients can make use of it.
mircea_popescu: (note that they quite exactly identify the exact reason gossipd makes empire impossible : "In the case of a public communications system this condition of a common symbol set with a common semantic interpretation must be further strengthened to that of a unique symbol set with a unique semantic interpretation. This condition of uniqueness allows any party to initiate a communication that can be received and understood by any
Framedragger: well you cant fault them for not having heard of gossipd. (cue everyone: YES YOU CAN). re. i2p, well its an interesting project. but - doomed.
ben_vulpes: "gossipd? what dat. have you ever heard of i2p? what's an ethernet frame? shortwave?"
Framedragger: what is nice is that the dns transport itself is quite elegant - questions/answers - one packet for query, another for answer. so transport is (in principle) compatible with session-less gossipd model, i think.
Framedragger: yes indeed; and now i see that maybe there wouldn't be too much of a redundancy there - that is, when gossipd arrives, dns server could still accept pgprams, they'd simply come via gossipd - and the latter may even sit behind an abstracted network interface, etc.
mircea_popescu: Framedragger yes, but communicatin by sending pgpgrams is already a simplified gossipd is the point.
mircea_popescu: at first, gossipd replaces the current "talk to servers to admin your strings" which happens, essentially, over substitute-email and http-sessions.
Framedragger: and maybe it's worth to focus on gossipd first. but i see what you're saying regarding these two things being parallel. i guess.
Framedragger: the whole "patch a DNS server and run alternative root" effort sounds interesting and useful, but, as you said, eventually the underlying layer would need to swapped for gossipd anyway. in gossipd, UDP/TCP as currently used by DNS may not even work. hence there may be a redundancy of effort;
mircea_popescu: there is EXACTLY ONE solution to "zooko's triangle" - and it is called gossipd. it works by crushing the socialist mind and its foul expectation of "any person is any person". not fucking so.
Framedragger: to an extent, i would say, no? in gossipd, user would have their own "hosts" file, mapping mircea_popescu to $fingerprint or w/e
mircea_popescu: this thing and gossipd are parallel anyway.
Framedragger: there is a nuance: if lotsa customization is needed, perhaps time/energy is better spent on gossipd. a logistical question i guess.
asciilifeform: the correct way to 'dns' is to 1) gossipd , which gives you 2) http://mypubkey/ahostnameofmine/...
asciilifeform: well, an old trabant-engine powered empire, cannot win against spherical frictionless gossipd, no.
mircea_popescu: and in these terms, a very simple explanation of why the republic is indefeatable is very simply forthcoming : given two users a and b who, through division of labour, have expertise in fields f1 and f2, the empire path for a to access f2 or for b to access f1 is to take some money all the way from -3 to 3 TWICE. whereas the republican model simply has a talk to b, on the mediation of a gossipd/wot
mircea_popescu: certainly the discussion of gossipd needs some updating.
mircea_popescu: emacs frozen as part of gossipd, other implementations allowed at user's discretion.
asciilifeform: mircea_popescu: imho a correct gossipd simply emulates irc server on the user end.
mircea_popescu: asciilifeform well with any luck tmsr fork of emacs will be part of gossipd anyways.
mircea_popescu: i jus' did, and best i can see it is sufficient, and how we'll do stuff like rsa encryption/decryption over uci once gossipd is here.
mircea_popescu: trinque he has a serious issue with things that i don't yet fully grok, but i can not presently distinguish from this irrational meltdown and the irrational meltdown on the gossipd spec.
mircea_popescu: in other words : paillier in UCI over gossipd seems capable of bignum.
thestringpuller: power rangers version of gossipd like stuff? https://github.com/bitcoin/bips/blob/master/bip-0150.mediawiki
mircea_popescu: anyway. jus wait for gossipd, it'll solve this.
mircea_popescu: anyway. should be fun to see how this plays out once gossipd.
mircea_popescu: no, evidently it works fine in gossipd
trinque: atop gossipd I think the *entire* "this website" thing goes, replaced by "I would like the thing identified by hash ..., do my friends have it?"
trinque: whole thing is broken until gossipd-year-0
mircea_popescu: which kind-of requires working gossipd.
asciilifeform: gossipd, in all proposed variants, relies on numbertheoretical crypto
mircea_popescu: the correct pill to all of this being, of course, gossipd.
shinohai switches to gossipd
a111: Logged on 2016-09-14 12:34 mircea_popescu: incidentally, and re http://trilema.com/2016/gossipd-design-document/#comment-119099 : if anyone likes to fuck around with markov chains etc, a prototype implementation of that would be perfect at this juncture.
mircea_popescu: http://log.mkj.lt/trilema/20160922/#311 << word. meanwhile this merged in my head with the eliza-for-gossipd thread (re http://btcbase.org/log/2016-09-14#1541643 ) ; there should in principle be a way to have the same tool do both things, and therefore probably should do it that way. ☝︎
PeterL: could relay messages from one gossipd node to another
asciilifeform: remember that the hypothetical scheme is not gossipd.
Framedragger: asciilifeform: ah, so 'private key' is purely 'station key' - i guess like in latest gossipd discussion
asciilifeform: this is not, note, a 'gossipd' net per se, but potentially component in one.
Framedragger: need to minimize energy expenditure at $job. "go home tired" sort of case, it's sad. i'd like to, e.g., among other matters (incl mkj log etc of course), write a very initial very to-be-criticized gossipd for udp/ip, using (kill me now) python twisted. i guess many live like this, savouring free time, with personal projects falling behind. [i rant today, because exhaustion + coffee infusion => wee bit psychedelic]
mircea_popescu: ah also : re "encryption" : we CAN use the eliza involved in gossipd spec (anyone actually did this or just some quick napkin work and then forgot about ?) to eliza the actual emissions.
Framedragger: asciilifeform: so that's the thing, maybe prototyping gossipd designs over usual packet-switched internet is not even worth it?
asciilifeform: at any rate, pretty much 100% of asciilifeform's work re gossipd, incl. prior to mircea_popescu's publication, was so that eventually radio.
mircea_popescu: but anyway, to remind teh esteemed lordship that http://trilema.com/2016/gossipd-design-document/ is still open for comments etc.
asciilifeform: https://github.com/trevp/double_ratchet/wiki << and a usg gossipd.
Framedragger: one side-effect from the above: it may therefore be that as one approaches the gossipd model / state of affairs, cooperation > minmax-style competition
mircea_popescu: you notice gossipd as specified is not related to implementation yes ?
gabriel_laddel: why even bother with gossipd?
asciilifeform: (it can, e.g, readily serve www, or even gossipd etc)
trinque: the trouble of not choosing your gossipd-node peerings wisely
mircea_popescu: you realise that the gossipd model with markov chains is probably the one source of the singularity
mircea_popescu: yeah, by now it's becoming evident that gossipd is to be delivered with a config file and a VERY LONG helpfile.
mircea_popescu: incidentally, and re http://trilema.com/2016/gossipd-design-document/#comment-119099 : if anyone likes to fuck around with markov chains etc, a prototype implementation of that would be perfect at this juncture. ☟︎
trinque: commands-over-gossipd *entirely* replaces the www, and there will eventually be UI for that.
scriba: Logged on 2016-09-13: [22:22:28] <pete_dushenski> frankly, it'd be swell as hell if mp and alf continued this debate in the comments section there - a la gossipd thread model - and then we'd have something to look back on weeks, months, or years from now.
pete_dushenski: frankly, it'd be swell as hell if mp and alf continued this debate in the comments section there - a la gossipd thread model - and then we'd have something to look back on weeks, months, or years from now.
asciilifeform: at any rate, i stand by my assertion that a synchronous (if you don't know what this means, look it up) carrier is unsuitable for gossipd.
asciilifeform: Framedragger: i won't touch a tcp 'gossipd'
Framedragger: imho one should not start prototyping gossipd on tcp due to the whole stateful nature of tcp. it would make things *appear* to be so much easier - implicit packet ordering and stream control, etc. - and when things eventually need changed it's gonna suck00r bigtime
Framedragger: asciilifeform: right. which is not a super strange thing to require anyway imho; this may be required when IP is dissected for gossipd purposes, anyway. (sure, ideally the whole thing should be (eventually) ditched.
trinque: freenode adding its 2 bitcents to the gossipd discussion.
Framedragger: http://trilema.com/2016/gossipd-design-document/#comment-119079 s/DZ/SD/ i don't know why i wrote DZ, very unclear freudian slip
scriba: Logged on 2016-09-13: [15:28:28] <asciilifeform> mircea_popescu et al : http://trilema.com/2016/gossipd-design-document/#comment-119069 << proposed 'lighthoused' spec
asciilifeform: http://trilema.com/2016/gossipd-design-document/#comment-119078 << seems like we're back to start, eh.
scriba: Logged on 2016-09-13: [15:28:28] <asciilifeform> mircea_popescu et al : http://trilema.com/2016/gossipd-design-document/#comment-119069 << proposed 'lighthoused' spec
asciilifeform: mircea_popescu et al : http://trilema.com/2016/gossipd-design-document/#comment-119069 << proposed 'lighthoused' spec
a111: Logged on 2016-09-13 08:11 mircea_popescu: trinque asciilifeform & everyone : without detracting from the more important gossipd spec work, im thinking of making a bot spec. it'd require bots to answer in pm to help, help json and help sexpr. comments/ideas ?
mircea_popescu: trinque asciilifeform & everyone : without detracting from the more important gossipd spec work, im thinking of making a bot spec. it'd require bots to answer in pm to help, help json and help sexpr. comments/ideas ? ☟︎
asciilifeform: sooooo it'd seem like we're pretty close to gossipd schematic
asciilifeform: was pete_dushenski here for the gossipd mega-thread ?
trinque: then later we're on gossipd and "online" doesn't mean anything
mircea_popescu: aaand gossipd updated.
PeterL: asciilifeform http://trilema.com/2016/gossipd-design-document/#comment-119021 response
asciilifeform: http://trilema.com/2016/gossipd-design-document/#comment-119016 << addendum.
asciilifeform: http://trilema.com/2016/gossipd-design-document/#comment-119015 << and yes i know mircea_popescu has auto-ping, but linking here for other folks.