269900+ entries in 0.187s
mircea_popescu: and importantly,
to continue
this discussion asciilifeform , enemy can not reconstruct fp on
the basis of intercepting your comms
to me.
mircea_popescu: and you can check
that it is, but only BECAUSE you are you and have
them both.
phf: asciilifeform: wouldn't
there just be a process
to work with primes primes directly?
mircea_popescu: asciilifeform not manually ; but you will get a notification from me saying "please use 54kh6kl54j3l45 instead of ,mn54,35lkjh4l3 from now on. it's
the same modulus".
phf: mircea_popescu: comment needs
to be separately signed?
mircea_popescu: phf
the only problem is
that i want
the comment
to also be in
there.
mircea_popescu: as
they say, "save
this sort of idea for when your father gives you your part of heavenly kingdom"
phf: well, primes stand on
their own,
then
there's clientside mapping mp -> sexp of primes. anything else can be distributed by any other channel, including "with key", signed by
the primes.
mircea_popescu: it doesn't come from "this item assures me it is
this item"
mircea_popescu: asciilifeform
the "selfsig" comes in
the form of "i obtained
this key of mp from verified channel
to mp".that's
the sig.
mircea_popescu: phf
they COULD just have made
that "Comment" in some serialization and did
the fp over it like i propose and been scott free.
phf: mircea_popescu: oh bunch of random shit, first name last name comment date generated random DOS era flags. all
the stuff
that's been brought up before
mircea_popescu: asciilifeform if you agree
to package it separately, add
to
that "no automated signing of anything" and you've got my position.
mircea_popescu: this is
the problem with socialist solutions :
they cap
the benefit certainly ; and offer some solution
that half works. so you've
traded
the long
tail of
the future for
the guarantee of current failure.
mircea_popescu: the fact
that
the youth is imbecile crates such designs
mircea_popescu: the fact
that it appears
to breeds idiocy in
the youth.
mircea_popescu: 3) as it's
theoretically nonsensical, it feeds a bunch of idiotic yet unconscious expectations in
the mind of
the user. i am well convinced
that
the general "i'll pull myself by own breeches" attitude of "creative" people is both due
to and resulting in
the gpg selfsig
mircea_popescu: as it's not acctually correctly designed it 1) creates false sense of security ; 2) creates unnecessary byzantinism and "can't pop
the hood on
this"
☟︎ mircea_popescu: sure. some classes : as it's unnecessary in
the discussed scheme, it 1) adds weaknes
through requiring unneeded computation ; 2) gives crevice for shitgnomery for no reason (see anvin bs) ; 3) possibly creates weakness
through unnecessary computation on key generation
mircea_popescu: though material encrypted
to it will still only be readable by me.
mircea_popescu: i want it
to work
this way
though. if you
take away my comment from my key and add something about how "miss piggy is
the hawtest!!1" it's not, in my view,
the same item.
mircea_popescu: the first
time you encounter
them you're in
the presence of
their owner.
mircea_popescu: anyway, fp is "checksum" de facto - if your modulus rotted it will no longer match
the fingerprint
mircea_popescu: it's not "the sks server"
that is retarded. is
the concept of machine-spread rsa key
that's retarded ; much in
the way of "machine-generated
trust", be it embodied in "dao" or "colored coins" or "safe bitbet"
☟︎☟︎ mircea_popescu: machine processes ; people spread/administer.
that's rsa key lifecycle.
mircea_popescu: if owner asks "which
this" you are not sufficiently connected so as for you
to has his key.
mircea_popescu: a) it becomes no longer obvious for enemy whether key is or is not = key he wants
to look for.
mircea_popescu: i am in favour of includingthe comment. for
the following
two reasons :
mircea_popescu: the one remaining cowlick is, should fingerprint cover
the whole e, N, comment ?
mircea_popescu: and for
the record, base-tmsr=` !"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_abcdefghijklmnopqrstuVWXYZ{|}~`
mircea_popescu: because hex
takes 2 characters per byte ; whereas
this would
take ~1.232 characters/byte
mircea_popescu: asciilifeform ideally
tmsr rsa, such as for eulora etc, uses a fp and armored base90 as above. so
there!
mircea_popescu: ascii 36
through 126 inclusive (91 elements). much better.
mircea_popescu: actually, why
the fuck are we even using "base 56", which is braindamaged (no 1 AND no l ? really ?) when we should really be using base-120 or somesuch, ascii-high-byte half.
mircea_popescu: yeah
this entire "id of rsa key"
thing needs a fix. i imagine sha512 of
tmsr standard is
the only way
to go.
phf: i suspect
that phuctor work has increased attention
to gpg both on wrecker and on attacker side. no way
that short-id collision would've been a
thing without phuctor