log☇︎
67100+ entries in 0.021s
asciilifeform: if tomorrow i sign a message with my key ' asciilifeform fleanode nick was compromised 2ks ago, ignore what the idjit said' -- also stands. names are clothing.
asciilifeform: trinque: i apparently don't
asciilifeform: ( if tomorrow asciilifeform signs a message 'call me pile_of_cement nao plox' everyone will laugh, but his v-sigs will still stand )
asciilifeform: http://btcbase.org/log/2017-11-08#1734603 << names are cosmetic. q is re 'what is a complete pubkey' ☝︎
asciilifeform: that's what bothers me, that i can't yet prove.
asciilifeform: mircea_popescu may well be right incidentally, re the impossibility of somebody lying about the e in a key, doing any damage . but i cannot yet prove.
asciilifeform: what happens if you change e to e', e' != e ?
asciilifeform: *tuple
asciilifeform: anyone see any typo ?
asciilifeform: this omits all mention of paddings etc.
asciilifeform: ( same exact op, but different colourings depending on whether encrypting or decrypting )
asciilifeform: m = c^d mod n
asciilifeform: c = m^e mod n
asciilifeform: for completeness, let's also note the rsa op :
asciilifeform: output is pub/priv pair; pub is tuble (n,e) , priv is (d, p, q) .
asciilifeform: ( can also write , d == e^-1 mod phi )
asciilifeform: 4) compute seekrit d, 1 < d < phi, where e*d === 1 mod phi
asciilifeform: 3) choose e, 1 < e < phi , where gcd(e, phi) == 1
asciilifeform: 2) phi = (p-1)(q-1)
asciilifeform: 1) we plug in the FG. p, q are primes. n = pq.
asciilifeform: let's take interlude to review ?
asciilifeform finds that the l0gz do not currently contain a full description of rsa
asciilifeform: does mircea_popescu know a proof that i cannot choose an alternate e that will result in a pubkey that 1) verifies some or all of the existing signatures made with his genuine e + 2) verifies another, which he did not produce ?
asciilifeform: the use of a hash ( i.e. item guaranteed not to be 1:1 mapping ) to 'reference unambiguously' to something, is a fractionalreserveism.
asciilifeform: if e is free variable ( vs. '65537 4evah 4all' ) then having n is having half a pubkey.
asciilifeform: because e is a free variable ?
asciilifeform: above however was written prior to discarding 'let's all 4evah use 65537'
asciilifeform: it's about asciilifeform reviewing for davout why http://btcbase.org/log/2017-04-09#1640824 was on point ☝︎
asciilifeform: what then, instead, about fram, ursul polar ?
asciilifeform: because it is about pubkey ?
asciilifeform: the pubkey is the man. per mircea_popescu's own theory and practice.
asciilifeform: you can't sign with either.
asciilifeform: like knife exists solely to cut.
asciilifeform: crypto sigs are to be unambiguous. is the entire point of'em.
asciilifeform: mircea_popescu balked at 'unfragging udp is 512byte' but for some reason doesn't barf at irc
asciilifeform: as i see it it's an arbitrary turd.
asciilifeform: however they don't use hashes, so no collisions
asciilifeform: ( they die faster than born , so no shortages )
asciilifeform: the ham radio people still have iirc what, 5 letter names
asciilifeform: mircea_popescu: why not take to logical conclusion : 1 bit : 1 == mircea_popescu , 0 == everybdy else, lol
asciilifeform: discussion is re permanent design for sane crypto.
asciilifeform: dun matter for how long it was state of the art in my chukchistan.
asciilifeform: after bath (ffa) is built, i have 0 intention of continuing to rub against tree trunk to clean.
asciilifeform: the historic 'it wasn't visible broken in my face' can justify any kind of shitcrypto, the use of winblowz, whatever customary syphilis . i ain't buying.
asciilifeform: you build bridge for 500 tanks, not for the wheelbarrows crossing river to date.
asciilifeform: that nobody saw it fit yet to do this in front of asciilifeform's face, is neither here nor there
asciilifeform: 20-50k usd gets you another 17215D118B7239507FAFED98B98228A001ABFFC7-sha1 but diff modulus.
asciilifeform: are we building a thousand year reich or not, lol
asciilifeform: has yet to materialize to naked eye.
asciilifeform: 'can always find right key'
asciilifeform: with this argument you could justify arbitrarily breakable fp, say.
asciilifeform: mircea_popescu's key with altered pubexp is a valid key, and not mechanically distinguishable from the original's in the absence of sigs.
asciilifeform: it does no such thing .
asciilifeform: it is relevant to the screaming wrongness of http://btcbase.org/log/2017-11-08#1734499 statement. ☝︎
asciilifeform: not knowing the e has exactly same effect as not knowing half of the n. ☟︎
asciilifeform: re fps yes. re 'e is part of a pubkey' - it is.
asciilifeform: names are not involved.
asciilifeform: enemy can offer a diddled key with your n but his e, that he can decrypt encrypts to.
asciilifeform: the historic cheat was to standardize the e. which imho was retarded.
asciilifeform: so they are both rightfully part of a pubkey.
asciilifeform: just same as if you did not know the n.
asciilifeform: not deliberately, lol. but if you don't know my e , you can't encrypt to me.
asciilifeform: review the algo.
asciilifeform: nope
asciilifeform: change pubexp ? you get a different pubkey
asciilifeform: pub exponent isn't a function of modulus, it's an independent variable
asciilifeform: http://btcbase.org/log/2017-11-08#1734480 << how are they to find out ? by same argument you could as easily say 'post first half of modulus, let'em look for other half' ☝︎
asciilifeform: davout: 1) in person 2) via secure channel from people who you already know, who know $person.
asciilifeform: observe the mendacious idiocy of koch's signature code, where if sha1 hash collision is found , can forge sigs ~regardless of what sig algo hashing was set to~
asciilifeform: but outside , my key is my key. not a hash.
asciilifeform: how you keep straight the contents of own household, is none of my business
asciilifeform: and you're welcome to hash on own end, 'in harem'. just don't refer to my key as its hash, it is unseemly
asciilifeform: what spotting.
asciilifeform: diff exists.
asciilifeform: davout: http://btcbase.org/log/2016-12-27#1590896 ☝︎
asciilifeform: hashes are voodoo, there is not a hash of demonstrable strength for the same reason there is not block cipher of same ( they are different sides of same medal, we had thread with elementary proof )
asciilifeform: and it doesn't matter what you make it out of.
asciilifeform: davout: it has to be a fixed in advance hash function.
asciilifeform: yes indeed the Actual Thing costs moar than the plastic doll.
asciilifeform: this echoes the ffa 'omfg slow' discussions.
asciilifeform: get used to doing ~the actual thing~ instead of 'convenient' simulacra.
asciilifeform: entirely practical.
asciilifeform: 1024 byte all.
asciilifeform: moar if people get to pick own exponent
asciilifeform: http://btcbase.org/log/2016-12-27#1590895 + http://btcbase.org/log/2017-04-09#1640824 for completeness. ☝︎☝︎
asciilifeform: ( has links to earliers )
asciilifeform: http://btcbase.org/log/2017-08-15#1698400 << most recent one ☝︎
asciilifeform: year+
asciilifeform: lemme dig out the thread
asciilifeform: i disagree
asciilifeform: there's no provision for changing the fp algo.
asciilifeform: all stuck with sha1 4evah
asciilifeform: davout: speaking in general of all even vaguely pgp-like compatibles
asciilifeform: it's approx on par with the old chinese leather cannon, even chance of exploding
asciilifeform: rfc specified sha1.
asciilifeform: davout: YOU CANNOT USE SANE HASH WITH GPG FP
asciilifeform: this is The Right Thing
asciilifeform: pull by getting key... from people
asciilifeform: and cannot be changed per the rfc.
asciilifeform: davout: it's a sha1