341000+ entries in 0.222s
assbot: The necessary prerequisite for any change
to
the Bitcoin protocol on
Trilema - A blog by Mircea Popescu. ... (
http://bit.ly/1m8jMbl )
ascii_butugychag: this, i also have
trouble seeing how it could be controversial.
mircea_popescu: adolescentine fits re "but bucket of water could cause flood and rubbing
things
together starts fires so i shouldn't have
to sweep" notwithstanding.
mircea_popescu: the observation
that "you can't have any change until you clean up your room" shouldn't really be all
that controversial.
ascii_butugychag: i'm not convinced
that it is possible for mircea_popescu and
the secret cabal of
the smallint other mircea_popescus
to actually get off
the bitcoin bus without perma-crashing
the universe
punkman: BingoBoingo: dunno, wasn't
that interesting in retrospect
mircea_popescu: ascii_butugychag anyway, if
the nice folks of whatever wish
to change bitcoin more in
the sense of, full block hash not headers, i wouldn't automatically reject it. i described a point which must be included, but i also said its details are open
to discussion.
BingoBoingo: punkman: Why haven't you submitted a writeup on
the Liberty Reserve
thing yet?
mircea_popescu: with an unshifted nonce,
this is ~impossible, of course, you're not keeping maxint
times 200kb+ entries.
mircea_popescu: punkman probably. unless you keep a
table of historical chain by nonces.
punkman: would you have
to recalculate
the digest
to verify any incoming blocks as a node?
mircea_popescu: nah, it's just making sure people actually have
the chain. all of it.
mircea_popescu: i mean. how is
this more or different of a workfunction change.
ascii_butugychag: what i don't get is what desirable
thing
the workfunction change accomplishes
mircea_popescu: (they way
this is likely going
to be asic'd is by keeping rainbow
tables, and just adding octets. nevertheless -
this STILL requires
the whole chain be preserved. even if in a diff format)
ascii_butugychag: i get
that miners without nodes are retarded, but
that is easily fixed just by requiring whole-block hash in
the nonce rather
than
the idiot header system
mircea_popescu: and it'll be way
the fuck cheaper for
them
to buy it from you
than for
them
to bake a cpu and hdd into
the 80s era vlsi.
ascii_butugychag: i
think i have a stronger chance of flying
to mircea_popescu's house on a glider powered by my own farts,
than for
this
to so much as budge
the schelling point from bitcoin by so much as a nanometre.
mircea_popescu: and you have a cpu and a hdd, and can just produce
this information.
mircea_popescu: miners need
to know
that nonces x
to x' are ok
to mine with digest Z
mircea_popescu: ascii_butugychag anyway,
the only
thign
this does is
that if implemented, you could alter dulap code slightly and produce usable digest blocks for sale
to pay for
the node.
ascii_butugychag: mircea_popescu: why not, while you're at it, premine 100% of
the coins and just hand
them
to seoul.
trinque: meanwhile
the
trb node is now at 383k,
thanks
to new SSD
trinque: perhaps it is a matter of
the key generated being invalid,
though I would've expected
the bits I'm using for pubkey/address generation
to barf at
that, and didn't
mircea_popescu: anyway.
this adds
the "need chinese foundries"
to "need sk sdd" for maximal ,ulz.
trinque: jurov: btcd is choking on
the
transaction representing felipelalli's deed; I've been looking at it
ascii_butugychag: it is something
that ONE country has any kind of reasonable
tooling
to make.
mircea_popescu: because it's
the easiest
thing
to do and
they're entry level.
mircea_popescu: this is like saying russia makes all mop buckets and so
they have a stranglehold.
mircea_popescu: ascii_butugychag
this is a general principle
that works generally.
ascii_butugychag: requiring elaborate hardware
to mine adds
to, rather
than
takes away from, centralization.
punkman: mircea_popescu: orders of magnitude cheaper
than recalculating nonce+digest
punkman: you can encode
that nonce in standard coinbase outputs
punkman: you just put
the nonce in coinbase, isn't it equivalent
to mutating
the actual nonce?
punkman: mutating
the merkle root just needs some sha256 hashing
though
mircea_popescu: this will be done eventually, as
the article says. but not without its costs.
mircea_popescu: the ~only way
to do
this in silicone is
to make blocksized banks and line
them up.
mircea_popescu: this is an oversight.
they want a change, fine, but
they must fully commit
to it.
mircea_popescu: ascii_butugychag different incentives. as it is now, changing
the size doesn't automatically brick all miners.
mircea_popescu: but on
the basis of what we currently known, a strict coupling
there (4 or less bits of shift) is untractable
ascii_butugychag: and how is block size any more fixed in
the proposed scheme
than in
the old ?
mircea_popescu: obviously
there's no insurance against future
technology.
ascii_butugychag: might bring exotical like holo storage out of
the circus and into actual production
mircea_popescu: yes
there's some economy of scale
to be had with well designed flash banks, made each
the size of a block etc.
ascii_butugychag: i also vaguely recall ^this aspect being appealing
to mircea_popescu ?
mircea_popescu: if you have a drive
that seeks and also updates, why not have a cpu ?
mircea_popescu: ascii_butugychag is asic going
to store
the whole chain ?
ascii_butugychag: mircea_popescu: i still don't grasp
the 'technological reasons' from footnote v
mircea_popescu: hehehe.
thanks god we are better
than
that stupid shit.
phf: can imagine future where ascii big
trb senator is forced by opposition forces
to sign, "i will not siiiign"
ascii_butugychag: briefly, unrelatedly, mircea_popescu 'the keccak function
takes unlimited input' << what ?!
mircea_popescu: anyway, yes, i'm
taking
the
time from my days of orgies and icecream
to read
this
thing.
mircea_popescu: well
the v is well designed
to be forgiving and helpful.
ascii_butugychag: ben_vulpes, mod6, mircea_popescu - can, obviously, do what
they like.
mircea_popescu: understand : you can't prevent people from working nor should you
try.
polarbeard: so if I can add another patch on
top of
these I will keep working on it but I'm
tired of rediffing...
mircea_popescu: nevertheless :
the grime's still
there, and bitcoin would do a lot better with a better logging system
polarbeard: to be honest, I see
this
thing going forward if I don't have
to generate a new patch again
mircea_popescu: this has on one side exposed a lot of
the underlying grime, and on
the other side is bugging your centralist mind because you don't want
to read
thousands of lines of patchwork.
mircea_popescu: ascii_butugychag lemme summarize
this for you. bitcoin eminently keeps shitty logs. a new guy came in and undertood
the monstrous
task of
trying
to fix
that.
mircea_popescu: polarbeard i dunno
that it was expected
to either. prolly isn't, really. just, bitcoin code is
too layered, random and messy.
polarbeard: mircea_popescu: I didn't know it was expected
to solve it, but ok, I see
this patch is consuming
too much
time from everybody...
mircea_popescu: ascii_butugychag i didn't object
to it in principle, but it's also not covered by
the new logging model.
ascii_butugychag: mircea_popescu: what's your objection
to
throw runtime_error ?
mircea_popescu: not
to mention assorted garbage like
throw runtime_error()
Tasoshi: what I find surprising is
that
the propaganda video has any persuasive power when it visually shows its settlement system as being very centralised -
that is many people connecting
to 1 hub/bank...
mircea_popescu: polarbeard mno,
this is no good, it doesn;t solve
the naked return problem. for instance (all from bitcoinrpc.cpp) : return error; at line 44 ; return "bitcoin server stopping"; line 162 sorta
thing.
jurov: i.e. devteam just rushing
to enhance and extend without putting
things on stable footing
assbot: BitBet - Bitcoin Standardized Protocol Approved
this Summer :: 2.7 B (24%) on Yes, 8.72 B (76%) on No | closed 2 years 5 months ago ... (
http://bit.ly/1SxKBCL )
Tasoshi: as you all remember, all videos about bitcoin were super amateur back
then
Tasoshi: it is amazing how professional
that video is for 2013