wumpus changed the topic of #bitcoin-wizards to: This channel is is for discussing theoretical ideas with regard to cryptocurrencies, not about short-term Bitcoin development | http://bitcoin.ninja/ | This channel is logged. | For logs and more information, visit http://bitcoin.ninja
<kanzure>
but really- my request for -wizards logs was genuine. early 2014 stuff is missing. and i don't know when the channel was actually started. i have some 2013 logs but not before.
<kanzure>
whoops i mean mid 2014 stuff is missing
<zooko>
tromp: so
<gmaxwell>
andytoshi: ^
<zooko>
tromp: was the benchmark which rated your GTX 980 as about as good as an i7, was that running 8 parallel instances of Cuckoo on the GTX 980?
<kanzure>
ah correction: i do have all 2014 logs, but not pre-2013 logs.
<kanzure>
gmaxwell: it was andytoshi's logs from 2013 that i have.
<gmaxwell>
oh well then you have the start
<gmaxwell>
before then -wizards was #bitcoin-dev and private messages between myself and petertodd.
<kanzure>
this is like scraping the bottom of a nice cream bowl
<kanzure>
hm okay..
shesek has quit [Read error: Connection reset by peer]
shesek has joined #bitcoin-wizards
mengine has quit [Ping timeout: 265 seconds]
kmels has quit [Ping timeout: 244 seconds]
<andytoshi>
early 2014 i think i was not involved with bitcoin heavily, sorry
<andytoshi>
but petertodd gave me some logs that i thought covered that era..
<kanzure>
yeah i was mistaken, sorry about that
jaekwon has joined #bitcoin-wizards
<andytoshi>
apparently not, just january
<andytoshi>
oh, sorry, never mind, i was not reading all the scrollback
<kanzure>
i have your logs 2013 to 2014-03-12, then i have my own logs from 2014-02-23 to now, and i had a logbot starting around 2014-10-19 and that's why i thought i was missing 2014-03-12 to 2014-10-19
<andytoshi>
excellent
<kanzure>
log reading is going as planned
<kanzure>
i definitely don't have old #bitcoin-dev logs... mine only go back to 2013-03-11 (surprise surprise).
Newyorkadam has joined #bitcoin-wizards
<kanzure>
i might have to limit the irc log reading to "interesting sections" defined by bursts of messages in a time period, because reading a 105 MB irc log file is not so easy.
Burrito has quit [Quit: Leaving]
<Luke-Jr>
mine probably go back to 2011 Jan
<kanzure>
i'll figure out a way to read ~50 million words of logs if you hand them over?
<Luke-Jr>
do I have to? :P
<gmaxwell>
careful that your IRC client doesn't log PMs in random channels.
<Luke-Jr>
gmaxwell: it does :<
<Luke-Jr>
only ones I send though
<Luke-Jr>
(eg, via /msg)
<gmaxwell>
right.
<gmaxwell>
irssi does
<gmaxwell>
sipa: that failure is because benchsign does not re-increase the buffer size in the loop.
zooko has quit [Ping timeout: 246 seconds]
<gmaxwell>
this is something of an API footgun-- a downside of an inout argument-- that maybe we should call out...
<gmaxwell>
an adequate fix is
<gmaxwell>
- int siglen = 74; for (i = 0; i < 20000; i++) {
<gmaxwell>
+ int siglen = 74;
<gmaxwell>
oh well assuming my irc client hadn't butchered that. :)
<gmaxwell>
damnit wrong channel
AaronvanW has quit [Ping timeout: 246 seconds]
mengine has joined #bitcoin-wizards
shesek has quit [Read error: Connection reset by peer]
shesek has joined #bitcoin-wizards
shesek has quit [Read error: Connection reset by peer]
<smooth>
I dont think this is correct really: "So to verify that a given transaction is valid, you'd have to have access to all the block chains"
<smooth>
with a tree structure you could restrict transactions to only span one level up or down
shesek has quit [Read error: Connection reset by peer]
rusty has joined #bitcoin-wizards
shesek has joined #bitcoin-wizards
<gmaxwell>
that post doesn't seem very useful to me? it looks like its saying they can never merge again if there is more than 1MB of data off to the side?!
btcdrak has quit [Quit: Connection closed for inactivity]
roxtrong_ has joined #bitcoin-wizards
starsoccer has quit [Ping timeout: 244 seconds]
starsoccer has joined #bitcoin-wizards
starsoccer is now known as Guest99747
Guest99747 has quit [Changing host]
Guest99747 has joined #bitcoin-wizards
roxtrong_ has quit [Ping timeout: 264 seconds]
snthsnth has quit [Ping timeout: 260 seconds]
shesek has quit [Read error: Connection reset by peer]
shesek has joined #bitcoin-wizards
moa has quit [Ping timeout: 244 seconds]
<smooth>
gmaxwell: i think by merge it means that one new block would have two parents, allowed if the sum the sizes of both parent chains was <1 MB for 1000 blocks
<smooth>
maybe that means average or something, hard to say, and probably never thought out
<gmaxwell>
not sure how to grade proposals like that on our list. Like.. I'm not sure what it means well enough to say anything about it.
<ryan-c>
if anyone wants to see it, my defcon talk about breaking brainwallets is up on youtube https://youtu.be/foil0hzl4Pg
<gmaxwell>
ryan-c you're about 5 hours behind me. :)
c-cex-yuriy has quit [Quit: Connection closed for inactivity]
shesek has quit [Read error: Connection reset by peer]
shesek has joined #bitcoin-wizards
GGuyZ has joined #bitcoin-wizards
Populus has quit [Remote host closed the connection]
shesek has quit [Read error: Connection reset by peer]
PRab has quit [Read error: Connection reset by peer]
shesek has joined #bitcoin-wizards
<kanzure>
is there any chance of probabilistic payments + payment channels working?
PRab has joined #bitcoin-wizards
<kanzure>
also:
<kanzure>
20:32 <@gmaxwell> I argue a specific commitment structure where miners, armed with a succinct ZKP for NP statements, create blocks which provide only an update to the UTXO set, and a constant size proof that the new utxo set was an authorized modification according to some unspecified number of undisclosed transactions.
<kanzure>
20:33 <@gmaxwell> It's lovely, except for the current infeasability of running ECDSA verification in the prover unless we don't mind 12 hour blocks. :)
<kanzure>
with payment channels already setup i don't think i'd mind 12 hour blocks
<kanzure>
plus someone convinced rusty that channel setup can be (trustlessly) "subsidized" by hubs for users waiting to get utxos
<kanzure>
also i think there was a proposal made somewhere "recently" about large multisig pools (using fancypants signature schemes) being onboarded on to lightning network, without causing utxo bloat? i don't know where i saw that.
<rusty>
kanzure: hmm, I didn't see that...
<kanzure>
well that's troubling, i was about to blame you for that one too
<kanzure>
maybe aj
snthsnth has quit [Ping timeout: 256 seconds]
CodeShark has quit [Ping timeout: 250 seconds]
<aj>
it was not me, perhaps the one armed man?
dignork has quit [Ping timeout: 244 seconds]
larraboj has quit [Ping timeout: 244 seconds]
shesek has quit [Read error: Connection reset by peer]
dignork has joined #bitcoin-wizards
larraboj has joined #bitcoin-wizards
shesek has joined #bitcoin-wizards
shesek has quit [Read error: Connection reset by peer]
<ryan-c>
After having done a few security evaluations, I assume https libs fuck up cert validation until they pass a test suite of shady certs
<ryan-c>
and non-http protocols using tls are even worse...
<ryan-c>
I found that $VENDOR who sells popular software for intranet wikis, ticketing, source control, etc had LDAPS support (which would be used for active directory in an enterprise environment) that didn't validate the server certificates.
<ryan-c>
another vendor had a server management system (that ran as root) that would connect to a master config server and pull stuff down and execute it as root
<ryan-c>
not only did it not validated https certificates, but if you blocked port 443 in the firewall it would fall back to using plain http with no additional integrity protection
<ryan-c>
also it was written in perl and was obfuscated by encoding the sourcecode as whitespace
dEBRUYNE has joined #bitcoin-wizards
<gmaxwell>
there have been a couple bitcoin webwallets that used websockets for all their communications and either didn't use the https version or would silently fall back if the https failed. :) not limited to custom non-browser apps.
<ryan-c>
gmaxwell: wut? was there some time period where unencrypted websockets were allowed from https pages?
<gmaxwell>
oh yea, I think only FF23 started to actually block that.
ThomasV has quit [Ping timeout: 246 seconds]
dEBRUYNE has quit [Ping timeout: 244 seconds]
<ryan-c>
in retrospect, http javascript from https pages should never have been allowed
<gmaxwell>
but advertisements!
<gmaxwell>
would just be another reason to not deploy https at all. :(
<gmaxwell>
now, ... https that looked entirely like HTTP and didn't display the lock icon, on any kind of mixed content would have been a boon.
<ryan-c>
ad industry will probably be 99% https within a year
<ryan-c>
yeah, it's a pity unauthenticated opportunistic crypto never really happened
<gmaxwell>
I tried unsucessfully for a while to get firefox to artifically delay http loading so that it was never faster than https. Unsurprisingly this was not a popular idea. :)
kmels has quit [Ping timeout: 246 seconds]
<ryan-c>
heh
<ryan-c>
chrome will, in some cases which i have not yet bothered to find the constraints of, predictively establish https sessions to servers
<ryan-c>
it'll do the tcp connection and handshake, then sit there waiting to make the http request
<gmaxwell>
As much as I distrust Yudkowsky writing about sociology (esp evolutionary motivated) I'm sure there is an element of that to some of these communications challenges.
<gmaxwell>
I think we've seen a lot of issues there with bitcoin's rapid growth. For most of 2014/2013 if you were new to bitcoin you were _likely_ learning about it from someone who was only a month before you, because the userbase was growing very ast.
<gmaxwell>
er fast.
<Eliel>
Well, he might not have the whole picture, but the world started making more sense to me after reading his texts and applying the concepts in them.
<gmaxwell>
And we ended up with pretty big cultural gaps where those of us around from earlier don't even speak the same language... on top of the normal background gaps.
<Eliel>
but yes, you're probably correct that there are language problems too
waxwing has quit [Read error: Connection reset by peer]
<gmaxwell>
Eliel: yes, well the thing to keep in mind is that all ideas are tools. Some ideas are better tools than others, an idea doesn't even have to be _correct_ to be a useful tool, so long as you use it prudently. Yudkowsky's work has a high level of opinion to researched substance though, and often it makes domain experts cringe. Sometimes severely. It also happens to resonate well with people, but
<gmaxwell>
thats not correlated with it being actually true (something you should know from his writing. :)) so a grain of salt or six is aways advisable (as with any author too). :)
<gmaxwell>
Well 'language' also in this case can mean the inferrential gaps. "How could the block size have anything to do with security or decenteralization" to quote a prominent CEO, who moments before would have described himself as being well read on the subject.
<Eliel>
ah, yeah, that one is definitely an inferential gap problem right there.
roxtrong_ has joined #bitcoin-wizards
<stonecoldpat>
the problem in oct 2013 (when I started) was also a lack of learning material apart from a handful of papers and asking questions in IRC/reading code, so I can't imagine many others going through the trouble to explore it that deeply
chmod755 has quit [Quit: Ex-Chat]
<stonecoldpat>
I made jokes to others that the SoK papers were basically what took me a year to learn
c-cex-yuriy has joined #bitcoin-wizards
<stonecoldpat>
(and would have been invaluable back then!)
waxwing has joined #bitcoin-wizards
<gmaxwell>
Yea, we still don't know how to train people on this stuff. SoK helps, but a lot of the best learning has been interactive.
<gmaxwell>
Also people's tolerances for teaching have diminished over time. :(
ThomasV has quit [Quit: Quitte]
roxtrong_ has quit [Ping timeout: 244 seconds]
rustyn has quit [Read error: No route to host]
rustyn has joined #bitcoin-wizards
Guyver2 has joined #bitcoin-wizards
Guyver2 has left #bitcoin-wizards [#bitcoin-wizards]
<amiller>
i'm pretty sure i wouldn't have gotten into crypto research at all if it weren't for cool ideas and problems i heard about in irc chat
<amiller>
it's also been really useful to scour mailing list posts from decades ago too
<amiller>
i'm really keen to help preserve all the stuff from this era
AaronvanW has quit [Ping timeout: 246 seconds]
spinza has quit [Excess Flood]
<gmaxwell>
I'll break my normal trend of finding the subject irritating, and post the idea that hard limits of any kind rule system are inherently politically unstable. Even when the limit is protecting an important tradeoff or major design consideration, limits themselves are prone to be seen as obstacles independant of their goals; and only by sustained effort does any limit of political important survive
<gmaxwell>
without being set to 0 or infinity.
<gmaxwell>
By this analysis perhaps Bitcoin's coin supply cap is inherently more stable because inflation becomes 0, one of the two stable states I suggest for any 'limit'.
<gmaxwell>
And any of these scale proposals that suggest other fixed preprogramed limits are inherently unstable.
ThomasV has joined #bitcoin-wizards
<gmaxwell>
I wonder if anyone has an examples from orginizational or public policy that contradict that intution?
spinza has joined #bitcoin-wizards
<gmaxwell>
Another reason that non-extremal limits are unstable is that they split their support. No matter what the value is, almost no one loves it. People who agree in principle can find plenty of room to disagree about any particular boundary.
<gmaxwell>
This kind of activity is obvious, I think, in things like US immigration policy, and things like enviroment emissions policy whenever they escape the bounds of a boring administrative agency and intersect a political process.
<Eliel>
I suspect a consensus about limits would be easier to reach if it was somehow possible to bill the costs for having higher limits to those who support them.
<jcorgan>
indeed, talk is cheap; opinions tend to be worth the cost of expressing them :-)
adam3us has quit [Quit: Leaving.]
bedeho has quit [Ping timeout: 246 seconds]
dc17523be3 has quit [Ping timeout: 250 seconds]
<gmaxwell>
Eliel: well thats why I suggested the difficulty based dynamic sizing. I think it's almost impossible to make progress though while people are conflating scale with scalability, and I think some of that is the shadow of a hard limit. Hard to say.
<Eliel>
gmaxwell: I think flexcap should be combined with one of the time based solutions so that the real limit is the smaller of the two.
<gmaxwell>
Eliel: yes, well ideally what you can do is make it so that the time thing is hopefully never the actual limit; so that it doesn't get attacked.
sparetire_ has joined #bitcoin-wizards
<Eliel>
as in, combining BIP101 with flexcap?
* gmaxwell
refuses to go into the weeds!
<gmaxwell>
I'd started to write a long respose there, as its easy to get into minutia design. :) harder to think about the more general contours.
<Eliel>
:)
<Eliel>
although, it'd probably be better to not have an absolute upper limit, even a growing one but rather a cost that declines as time passes to increase the limit.
dc17523be3 has joined #bitcoin-wizards
c0rw|zZz has quit [Ping timeout: 240 seconds]
<gmaxwell>
Yes, well a resistance that goes asymtopic.
<gmaxwell>
May be a lot more poltically stable at least. Also, safer in some sense-- but bad engineering, as you can't reasonably test to the 'limit' of what the software would be demanded to do.
dEBRUYNE has joined #bitcoin-wizards
<Eliel>
true, but at least it will result in a system where you can pretty accurately predict the actual limits.
<gmaxwell>
yea, but safty is not about the average case, its about the 0.0001% of edge cases. :)
<Eliel>
would it not be good enough?
<gmaxwell>
well that depends, if someone can profitably push a system into an edge case, a one in a billion 'chance' is not very secure!
<gmaxwell>
it's happen with probablity 1.
<Eliel>
hmmh... looking at the growth of average block size for the last 3 years, it looks like even BIP101 would halve the speed of growth.
<gmaxwell>
Eliel: thats meaningless-- in the sense that block size growth has been almost exclosively controlled by the soft limiting settings since late 2012, and has perfectly filled whatever was available, also I think you're actually not measuring the average.
<Ylbam>
We now have three different versions of the TXO set: ∑a, ∑a + ∑b, and
<Ylbam>
∑a+∑c. Each of these versions is consistent in that for a given txoutid
<Ylbam>
prefix we can achieve consensus over the contents of the TXO set.
<Ylbam>
could someone explain please?
bedeho has joined #bitcoin-wizards
heretolearn has joined #bitcoin-wizards
eudoxia has joined #bitcoin-wizards
b_lumenkraft has joined #bitcoin-wizards
StephenM347 has joined #bitcoin-wizards
realcr has quit [Quit: WeeChat 0.4.2]
ASTP001 has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
<Ylbam>
are tree-chains still being worked on by petertodd? haven't found any recent article about it
<Ylbam>
only cited in some recent block size papers
roxtron__ has quit [Remote host closed the connection]
roxtrong_ has joined #bitcoin-wizards
nwilcox has quit [Quit: leaving]
ThomasV has joined #bitcoin-wizards
<kanzure>
can the trustless properties of lightning network-style payment channels be applied across merged-mined chain boundaries? or does that lose some of the trustlessness.
<kanzure>
also i think that if it can be shown that those trustlessness properties hold for that sort of scenario, that a bip70-style payment protocol could be developed where wallets do lightning network-style negotiation and also chain selection where your hubs might have cross-chain presence. then you don't entirely need your merchant (or whoever) to be expecting your original funds to be originating from the same blockchain as they wish to ...
<smellymoo>
fluffypony, I think I solved the problem. Want your critic.
<kanzure>
these pseudonyms are dope
<smellymoo>
kanzure, heh.
Dizzle has quit [Quit: Leaving...]
<smellymoo>
fluffypony, I'll pm you.
zooko has left #bitcoin-wizards ["#tahoe-lafs"]
<kanzure>
benefit of bip70-like payment protocol + payment channels + cross-chain hubs is that you could have committed transaction volume shifting around between different chains. and then you don't have to bottleneck a single blockchain with a billion transactions/second.
<kanzure>
although you would have to come up with a good way to convince people to receive outputs on the other chains....
<kanzure>
or er, to shift their preferences. probably with lower fees.
<kanzure>
also i should note that a bip70-style payment protocol that is interactive between sender and receiver was likely to happen anyway for payment channel reasons and lightning network reasons. offloading even more work into that wallet-to-wallet communication seems like an easy win for everyone.
user7779078 has quit [Remote host closed the connection]
chmod755 has joined #bitcoin-wizards
user7779078 has joined #bitcoin-wizards
jaekwon has joined #bitcoin-wizards
c-cex-yuriy has quit [Quit: Connection closed for inactivity]
<kanzure>
(((also when i say merged-mined chains above i am also talking about compact spv proof magic for bitcoin teleporation or bitcoin token migration; so convincing someone to accept outputs on another chain is not as hard as otherwise.)))
CodeShark has quit []
kmels has joined #bitcoin-wizards
user7779078 has quit [Read error: Connection reset by peer]
<smellymoo>
kanzure, are you talking to yourself?
<ryan-c>
braindumping
<smellymoo>
ryan-c, yeah, I got a totally bitching idea, but I'm going to hold onto it until I figure out the details and write my white paper. Not so good at keeping things in my brain though.
<kanzure>
ryan-c: correct
<kanzure>
irc is mostly asynchronous communication anyway
b_lumenkraft_ has joined #bitcoin-wizards
b_lumenkraft has quit [Ping timeout: 250 seconds]
b_lumenkraft_ is now known as b_lumenkraft
kang_ has joined #bitcoin-wizards
Dizzle has joined #bitcoin-wizards
trippysalmon has joined #bitcoin-wizards
user7779078 has joined #bitcoin-wizards
CodeShark has joined #bitcoin-wizards
ASTP001 has joined #bitcoin-wizards
priidu has quit [Ping timeout: 244 seconds]
ThomasV has quit [Ping timeout: 272 seconds]
nwilcox has joined #bitcoin-wizards
dEBRUYNE_ has joined #bitcoin-wizards
triazo has left #bitcoin-wizards ["end of my coinyDays"]
priidu has joined #bitcoin-wizards
GGuyZ has quit [Quit: GGuyZ]
GGuyZ has joined #bitcoin-wizards
GGuyZ has quit [Client Quit]
eudoxia has quit [Quit: Leaving]
paveljanik has joined #bitcoin-wizards
nullbyte has joined #bitcoin-wizards
dc17523be3 has quit [Ping timeout: 260 seconds]
dc17523be3 has joined #bitcoin-wizards
chmod755 has quit [Quit: Ex-Chat]
davec has quit [Read error: Connection reset by peer]
rubensayshi has quit [Remote host closed the connection]
davec has joined #bitcoin-wizards
GGuyZ has joined #bitcoin-wizards
chmod755 has joined #bitcoin-wizards
GGuyZ has quit [Quit: GGuyZ]
rodarmor has joined #bitcoin-wizards
rustyn has quit [Ping timeout: 240 seconds]
GGuyZ has joined #bitcoin-wizards
p15x_ has quit [Max SendQ exceeded]
c0rw1n has quit [Ping timeout: 260 seconds]
c0rw1n has joined #bitcoin-wizards
moa has joined #bitcoin-wizards
Guyver2 has joined #bitcoin-wizards
Guyver2 has left #bitcoin-wizards [#bitcoin-wizards]
p15x has joined #bitcoin-wizards
b_lumenkraft has quit [Quit: b_lumenkraft]
trippysalmon has quit [Read error: Connection timed out]
trippysalmon has joined #bitcoin-wizards
frankenmint has quit [Remote host closed the connection]
nwilcox has quit [Ping timeout: 272 seconds]
LeMiner2 has quit [Read error: Connection reset by peer]
c-cex-yuriy has joined #bitcoin-wizards
LeMiner has joined #bitcoin-wizards
afk11 has quit [Ping timeout: 246 seconds]
chmod755 has quit [Quit: Ex-Chat]
b-itcoinssg has joined #bitcoin-wizards
frankenmint has joined #bitcoin-wizards
gill3s has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
ghtdak has left #bitcoin-wizards [#bitcoin-wizards]
mjerr has quit [Ping timeout: 246 seconds]
rustyn has joined #bitcoin-wizards
GGuyZ has joined #bitcoin-wizards
GGuyZ has quit [Ping timeout: 250 seconds]
GGuyZ has joined #bitcoin-wizards
copumpkin has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
GGuyZ has quit [Ping timeout: 268 seconds]
jaekwon has quit [Remote host closed the connection]
roxtrong_ has quit [Ping timeout: 250 seconds]
roxtrongo has joined #bitcoin-wizards
ThomasV has quit [Ping timeout: 268 seconds]
copumpkin has joined #bitcoin-wizards
roxtrongo has quit [Ping timeout: 252 seconds]
roxtrongo has joined #bitcoin-wizards
adam3us has joined #bitcoin-wizards
<kanzure>
giant n-of-m pools based on web-of-trust reputation could allow for much lower fees for various transaction types, but entry into the pool will be much more scrutinous because nobody wants a troublemaker that causes delays or for fraud-correction or revocation transactions to have to hit the blockchain (which might even squader prior fee profitability). but anyway, because the fee savings are so great, this might exert a pressure on ...
<kanzure>
... scrutiny of members that join an m-of-n signing ring, although personally i hope to see 10k-member signing rings that work reliably for multiple years.
jaekwon has joined #bitcoin-wizards
wilbns has left #bitcoin-wizards [#bitcoin-wizards]