sipa changed the topic of #bitcoin-wizards to: This channel 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
BitcoinErrorLog has quit []
melvillian has quit [Remote host closed the connection]
adlai has joined #bitcoin-wizards
frankenmint has joined #bitcoin-wizards
supasonic has joined #bitcoin-wizards
glitch003 has quit [Remote host closed the connection]
rusty has joined #bitcoin-wizards
snthsnth has joined #bitcoin-wizards
glitch003 has joined #bitcoin-wizards
tromp has joined #bitcoin-wizards
dnaleor has quit [Quit: Leaving]
voxelot has quit [Ping timeout: 246 seconds]
dnaleor has joined #bitcoin-wizards
Newyorkadam has joined #bitcoin-wizards
Logicwax has quit [Ping timeout: 250 seconds]
domwoe has joined #bitcoin-wizards
[Derek] has quit [Ping timeout: 260 seconds]
RoboTeddy has joined #bitcoin-wizards
melvillian has joined #bitcoin-wizards
Logicwax has joined #bitcoin-wizards
Guyver2 has quit [Quit: :)]
[Derek] has joined #bitcoin-wizards
[Derek] is now known as Guest95522
tromp has quit [Remote host closed the connection]
RoboTeddy has quit [Ping timeout: 260 seconds]
bit2017 has quit [Ping timeout: 260 seconds]
moa has joined #bitcoin-wizards
Guest95522 is now known as [Derek]
[Derek] has quit [Changing host]
[Derek] has joined #bitcoin-wizards
melvillian has quit [Remote host closed the connection]
RoboTeddy has joined #bitcoin-wizards
RoboTeddy has quit [Read error: Connection reset by peer]
RoboTeddy has joined #bitcoin-wizards
melvillian has joined #bitcoin-wizards
melvillian has quit [Ping timeout: 248 seconds]
snthsnth has quit [Ping timeout: 246 seconds]
tromp has joined #bitcoin-wizards
snthsnth has joined #bitcoin-wizards
pottsy has joined #bitcoin-wizards
dEBRUYNE_ has quit [Quit: Leaving]
dEBRUYNE has joined #bitcoin-wizards
dEBRUYNE has quit [Read error: Connection reset by peer]
Ylbam has quit [Quit: Connection closed for inactivity]
wallet42 has joined #bitcoin-wizards
jtimon has quit [Ping timeout: 248 seconds]
CrazyTruthYakDDS has quit [Quit: Connection closed for inactivity]
aem is now known as aem
dgenr8 has joined #bitcoin-wizards
wiz has joined #bitcoin-wizards
dstadulis has joined #bitcoin-wizards
CrazyTruthYakDDS has joined #bitcoin-wizards
domwoe has quit [Remote host closed the connection]
BCBot_ has quit [Remote host closed the connection]
BCBot has joined #bitcoin-wizards
<JackH>
thats info overload
tromp has quit [Remote host closed the connection]
<BitcoinErrorLog>
i'm gonna go read for 5yrs brb
<kanzure>
BitcoinErrorLog: i'll hold your beer(hat)
mesmer_ is now known as mesmer
<mesmer>
kanzure: thanks!
<kanzure>
JackH: maybe... i used to try to keep up with the wikipedia firehose to disprove people who told me that information overload was a thing. jokes on me becuse now i just know about millions of shitty edits....
fletom has quit [Remote host closed the connection]
<JackH>
we actually just need people to sort this out kanzure
<JackH>
man power can fix it
<JackH>
I tried making my own library of info
<JackH>
I had to give up and just keep it high level, like: www.bitcointalk.org
fletom has joined #bitcoin-wizards
<mesmer>
that's the root of the problem though, it takes time to reach high level understanding if you have been in this space for awhile, but the barrier of entry for eager people and potential devs would have to walk through this same pain every time.
ThomasV has joined #bitcoin-wizards
domwoe has joined #bitcoin-wizards
tr0nk has quit [Ping timeout: 260 seconds]
fletom has left #bitcoin-wizards ["undefined"]
<bsm117532>
I keep a database in Zotero of papers, and have a map in my head of what's important. Everyone builds their own personal knowledge-base a little differently, and that's valuable because everyone gets a different perspective. kanzure likes to aggregate everything, and it's valuable to share, but I personally also find value in "forgetting" things also. Hopefully the stuff I remember is the important stuff. Hopef
helo_ is now known as helo
adlai has joined #bitcoin-wizards
tr0nk has joined #bitcoin-wizards
<harding>
bsm117532: you got cut off at what I'm guessing is the second hopefully: "the important stuff. Hopef"
domwoe has quit [Remote host closed the connection]
<Taek>
JackH, mesmer, bsm117532: wrote the first half of a backend for something that's supposed to help with knowledge curation last Saturday
<instagibbs>
forgetting things are important for personal relationships, not sure about scholarship :P
<Taek>
you upload primary sources (papers, ramblings, etc.) to topic names, then each topic lists all the primary sources that people have added to it
<Taek>
if you click on the primary source, you are brought to a page where annotations/elaborations/criticisms of the primary source can be uploaded
<Taek>
which should make it easier to write criticisms one time and then have an obvious way in which they are linked to the primary source
<JackH>
yeah bsm117532 nothing beats the information you have in your head. it is still the best blockchain for storing and retrieving info
<JackH>
it scales unlimited
<Taek>
I still need to add user accounts, and I'm planning on also adding upvotes + downvotes for primary sources and their annotations
<harding>
Taek: what are left/right/center votes?
<JackH>
the problem is that there is too much
<JackH>
you still need to know what you want to look for, and therein is the problem
<JackH>
you cant know what you dont know
<Taek>
harding: left+right votes help distinguish between beginner topics and advanced topics. There will be two columns for primary sources. Left column would have stuff like 'Mastering Bitcoin' and right column would have stuff like 'Bitcoin Developer's Guide'
<Taek>
oh, finally each topic has a list of other topics at the bottom which are similar
<harding>
Taek: cool.
<Taek>
people can tether topics, and then use voting to indicate how relevant something is
<Taek>
in that case the left/right/center votes indicate whether something is more general or more specific
RoboTeddy has quit [Ping timeout: 276 seconds]
<Taek>
so, 'sha256' would be more general than bitcoin, and '51% attacks' would be more specific than Bitcoin
domwoe has joined #bitcoin-wizards
<Taek>
JackH: the act of linking topics is supposed to help you learn what you don't know
<JackH>
hmm yeah that could work
laurentmt has joined #bitcoin-wizards
<JackH>
well reddit style is great, but the problem is spam there
<JackH>
too many people fucking things up
bit2017 has joined #bitcoin-wizards
bit2017 has quit [Max SendQ exceeded]
melvillian has joined #bitcoin-wizards
melvillian has quit [Ping timeout: 276 seconds]
melvillian has joined #bitcoin-wizards
<bsm117532>
Hopefully. Trying to keep track of everything is exhausting.
* bsm1175321
glazes over word salad. is there a solution in there, or just complaining about doi's?
PaulCapestany has quit [Quit: .]
<harding>
bsm117532: I think it says that if you submit an article to that site, you get a DOI.
PaulCapestany has joined #bitcoin-wizards
jtimon has quit [Ping timeout: 248 seconds]
Guyver2 has joined #bitcoin-wizards
bsm1175321 has quit [Ping timeout: 276 seconds]
voxelot has joined #bitcoin-wizards
priidu has joined #bitcoin-wizards
Erik_dc has joined #bitcoin-wizards
Monthrect is now known as Piper-Off
domwoe has quit [Remote host closed the connection]
Guest50611 has quit [Changing host]
Guest50611 has joined #bitcoin-wizards
Guest50611 is now known as amiller
Eliel_ has joined #bitcoin-wizards
Iriez has joined #bitcoin-wizards
ebfull has quit [Quit: Konversation terminated!]
Piper-Off is now known as Monthrect
ebfull has joined #bitcoin-wizards
_Iriez has quit [*.net *.split]
Eliel has quit [*.net *.split]
tr0nk has quit [Ping timeout: 268 seconds]
domwoe has joined #bitcoin-wizards
melvillian has quit [Remote host closed the connection]
tromp has joined #bitcoin-wizards
tromp has quit [Ping timeout: 248 seconds]
DougieBot5000 has joined #bitcoin-wizards
<amiller>
hi everyone
<kanzure>
sup
ShadeS_ is now known as ShadeS
jaekwon has joined #bitcoin-wizards
<tromp_>
hi, Andrew
paveljanik has joined #bitcoin-wizards
<andytoshi>
hey tromp_
dstadulis has joined #bitcoin-wizards
<andytoshi>
oh amiller
<andytoshi>
sorry, i need coffee.. and to stop highlighting on "andrew"
<amiller>
greetings to all andrews
bsm1175321 has joined #bitcoin-wizards
dstadulis has quit [Client Quit]
<harding>
andytoshi: I also have to figure how to tighten up irssi's default highlight rules as I get highlighted everytime someone here is talking about sharding. :-)
<amiller>
rofl
<bsm1175321>
Hey if you can get more people to talk about sharding, I'm all for it! ;-)
melvillian has joined #bitcoin-wizards
<helo>
harding: i ended up highlighting on "helo " and " helo". seems to work farily well.
<harding>
helo: good idea. Thanks!
<kanzure>
how about "helo: "
<helo>
luckily, the space at the start of the line allows e.g. "< guy> helo, blah" to be matched
Yoghur114 has joined #bitcoin-wizards
melvster has quit [Ping timeout: 248 seconds]
dEBRUYNE has quit [Quit: Leaving]
LeMiner has quit [Ping timeout: 252 seconds]
<Tasoshi_>
with mempool now full, are we to expect nodes to start crashing?
LeMiner has joined #bitcoin-wizards
<Tasoshi_>
or was that just made up by Mike Hearn?
<Tasoshi_>
there is apparently a backlog of 30mb
<Tasoshi_>
and rising
priidu has quit [Ping timeout: 248 seconds]
<arubi>
it's not backlog, it's altruistic node policy.
<arubi>
maybe classic\xt nodes are crashing, I wouldn't know
<Eliel_>
Tasoshi_: #bitcoin would be a more suitable channel for that question. It's off topic here. To answer the question, 0.12 won't crash due to this at all. Older nodes might, if the system runs out of RAM. 30MB is not nearly enough to do much though.
<arubi>
oh wow, I didn't notice we're here
<Tasoshi_>
have there been any tests as to what mempool nodes can handle?
<Tasoshi_>
frankly, I find the situation a bit scary, the backlog keeps increasing.......
bliljerk_ has quit [Quit: Leaving...]
bliljerk101 has joined #bitcoin-wizards
melvillian has quit [Remote host closed the connection]
melvillian has joined #bitcoin-wizards
priidu has joined #bitcoin-wizards
bliljerk101 has quit [Read error: Connection reset by peer]
bliljerk101 has joined #bitcoin-wizards
melvster has joined #bitcoin-wizards
melvillian has quit [Remote host closed the connection]
bliljerk101 has quit [Client Quit]
bliljerk101 has joined #bitcoin-wizards
<kanzure>
wrong channel
dstadulis has joined #bitcoin-wizards
<Tasoshi_>
sorry, I was looking for some sort of technical explanation in regards to what mempools can handle
<Tasoshi_>
and whether we should expect them to start crashing, or otherwise
melvillian has joined #bitcoin-wizards
<amiller>
i think that's on the knife's edge of on-topic, failure modes can be a research question
<amiller>
i think that there's a new policy in .12 that limits the size of mempool, i think that XT has had mempool limits for a while
<amiller>
i think that nodes with old code will just fill up and up, and at some point could use more and more memory until they slow down or crash
dstadulis has quit [Client Quit]
melvillian has quit [Remote host closed the connection]
melvillian has joined #bitcoin-wizards
<Tasoshi_>
hardly any nodes have upgraded to 0.12 though. More specifically, what I am enquiring is whether Hearn was right when he stated that once the mempool starts building up with genuine transactions a systemic node crash will follow or has that been rebutted?
<instagibbs>
this is #bitcoin material
melvillian has quit [Remote host closed the connection]
<amiller>
ok
<instagibbs>
not your fault, Tasoshi_ might be banned from #bitcoin...
<amiller>
i'm interested so i'm reasking the question there
dstadulis has joined #bitcoin-wizards
melvillian has joined #bitcoin-wizards
laurentmt has joined #bitcoin-wizards
Don_John has joined #bitcoin-wizards
dEBRUYNE has joined #bitcoin-wizards
melvillian has quit [Remote host closed the connection]
<gmaxwell>
amiller: It's very hard for lots of mempool memory to be used by 0.11.2 because the base relay fee is high enough that the prior people who were attacking that way were unsucessful. There is no such thing as "starts building up" as different nodes, have different policies: case in point, a long running unmodified 0.12 node here has 286MB in mempool at the moment (0.12 will always have about 300M
<gmaxwell>
B in the mempool since thats its limit). Though sure, if your system can't allocate enough memory/swap eventually the daemon will shut down.
<gmaxwell>
And yes, this is pretty offtopic here.
<amiller>
thanks
domwoe has quit [Remote host closed the connection]
<Tasoshi_>
couldn't a malicious miner easily attack bitcoin right now by mining fully empty blocks?
<Tasoshi_>
since there would be no counterance to that sort of attack, that malicious miner can in effect cripple the entire network?
<Tasoshi_>
is it not the case that, as the system is operating right now, as far as miners are concerned, it is in no way decentralised and the usual incentives no longer apply, for 1 miner can set the policy of the entire network by simply mining empty or 250kb blocks, thus turning the system onto a centralised operation at the mercy of 1 miner and that is possible only because the other miners can no longer operate in competition, because they do
<Tasoshi_>
not have the choice to, say, increase the size of the blocks they mine in return
<kanzure>
so in your view, any non-zero load or backlog is an attack because..?
<Tasoshi_>
my view is simple
<Tasoshi_>
all miners are operating at 1mb, and the system is somewhat functioning, what if one malicious miner decided to give bitcoin a system shoock, and mine 0 blocks
<Tasoshi_>
or 250k blocks, or 500k blocks, whatever
laurentmt has quit [Quit: laurentmt]
<Tasoshi_>
the other miners have no choice or say, there is no competition any longer
<Tasoshi_>
no balance
<kanzure>
why would empty blocks be a shock?
<kanzure>
if anything it is less of a shock, giving low-bandwidth nodes some time to catch up
<Tasoshi_>
let us suppose this miner has 25% hashrate, one can estimate the number of txs, but let us assume they are 50k txs
<Tasoshi_>
having, all the sudden, 50k txs on a backlog breaks everything, including fee estimation etc
<Tasoshi_>
but, more than that, this miner, single handedly can decide network policy, he can decide the fee rate, he can decide all he wishes, 1 miner
<kanzure>
that's already the case- any miner can decide to include whatever fee rate
<Tasoshi_>
there would be no longer any nash equilibrium or decentralisation
<Tasoshi_>
except that other miners, if they have free blocksapce, can compensate, and drive the monopoly miner out of business
<Tasoshi_>
but with blocks limited at 1mb, the other miners have no choice
wallet42 has joined #bitcoin-wizards
<Tasoshi_>
and, since subsidy currently plays the higher part, why shouldn't miners engage in a cartel monopoly behaviour to set fees at, for the long term, unacceptable behaviour, but short term no choice
<Tasoshi_>
so, in effect, well, killing bitcoin, whether they know it or not
wallet42 has quit [Quit: Leaving.]
jtimon has joined #bitcoin-wizards
sparetire_ has joined #bitcoin-wizards
<gmaxwell>
Tasoshi_: your assumptions don't make sense... having a 50k tx backlog doesn't break anything (and, in fact, fee estimation doesn't work without backlog)
<Tasoshi_>
how does fee estimation work?
<gmaxwell>
in any case, today most miners are running with 750k limits. If a miner with less than a majority hashpower stopped mining transactions, at most they could do is effectively halve capacity (or really, decrease it by a quarter relative to 750k).
<gmaxwell>
Tasoshi_: by observing how long it takes for backlogged transactions to get mined.
<kanzure>
here's some ideas on upcoming fee estimation things,
<Tasoshi_>
gmaxwell, how does that observation work when you have a system shock by a malicious miner?
<Tasoshi_>
let us say it takes 0.001 fee to corfirm, everything fine, then a malicious miner decides to set policy, and we suddenly have 50k txs backloged
blkdb has quit [Read error: Connection reset by peer]
<kanzure>
well, for that particular miner you could assume that none of the transaction fees were high enough to entice them (otherwise they would have included the transaction). but this says nothng about the miners that might be mining upcoming blocks.
<Tasoshi_>
is that acceptable gmax?
<Tasoshi_>
what if those txs are urgent?
blkdb has joined #bitcoin-wizards
<Tasoshi_>
but, lets not stop there, the miner wants to play, now he sets the blocks to 250kb
<Tasoshi_>
500kb
<Tasoshi_>
0 again
<gmaxwell>
Tasoshi_: it'll read low for a little while, but catch up. You don't need to hypothesize a malicious miner, the variance of finding blocks is _enormous_ it's not unheared of for there to be long runs of 20 minute apart blocks; and it's completely expected.
<Tasoshi_>
there's a difference, a human is playing
<gmaxwell>
Tasoshi_: you realize this miner only controls his own blocks, right?
<Tasoshi_>
for his gain, and no one can do anything about it
<Tasoshi_>
how would the network be decentralised under that sort of playful miner?
<Tasoshi_>
he, basically, rules
<gmaxwell>
If he does, thats basically a violation of the system security assumptions. But sure. So what? he can decrease capacity by up to 25%-- but swings of that capacity amount are expected just from block finding variance.
<Tasoshi_>
he is setting policy, 50k txs are 1/4th of the network right now
<kanzure>
what does it mean to set policy here ??
<gmaxwell>
huh?
<Tasoshi_>
but it is worse, because he just changed his 1/4th to 0, meaning his effect is half of the network
<gmaxwell>
Sorry, I'm not able to follow your logical leaps here. A miner with 25% hashpower controls a quarter of the blocks on average.
<Tasoshi_>
gmaxwell, let us take f2pool, with say 50k txs or 25%, the network now is operating as it is, then suddenly f2pool creates a system shock and adds 50k to the backlog, while withdrawing their service, that has amplifying effects
<Tasoshi_>
sure, the system adjusts, so now f2pool adds 25k txs, then withdraws them to 0, then adds 50k
<Tasoshi_>
it's chaos
<Tasoshi_>
and everyone else has no power whatever
<Tasoshi_>
no nash equilibrium
<gmaxwell>
what?
<kanzure>
sounds like you are worried about not being able to estimate transaction backlog clearing
<Tasoshi_>
what part do you want me to explain?
<kanzure>
it would be helpful to imagine a constant (or growing) backlog, rather than assuming an intermittent backlog, i think
<gmaxwell>
What you're saying doesn't follow logically. The amount of capacity is inherently unpredictable just by the design of the system. It works fine like that. You've made no argument as to why an additional 25% hashpower pumping the capacity would be "chaos" as you describe.
<gmaxwell>
(relative capacity is also unpredictable because demand is unpredictable)
jtimon has quit [Ping timeout: 248 seconds]
<Tasoshi_>
so instantly reducing transaction capacity right now to say 100k txs would not be chaos?
morcos has joined #bitcoin-wizards
wallet42 has joined #bitcoin-wizards
<gmaxwell>
Tasoshi_: AFAIKCT, you're making a dimensionality error. What does "capacity 100k txs" even mean?
<Tasoshi_>
let us say that right now people can only carry 200k txs a day
<Tasoshi_>
suddenly, with no warning, reducing that to 100k would not cause chaos?
<gmaxwell>
You mean 150k (25% reduction from 25% hashpower not including transactions)?
<Tasoshi_>
well, let's use my numbers
<kanzure>
it might violate the assumptions that some people had, but i'm not sure we should expect to meet unrealistic assumptions
<gmaxwell>
No, it would not cause chaos-- that already happens due to variance, due to existing policy (miners do not have uniform policies to begin with)...
<gmaxwell>
Tasoshi_: your numbers require the attacker be half the hashpower, not 25%
<Tasoshi_>
variance brings tx capacity from 200k to 100k suddenly?
<Tasoshi_>
beyond the system shock though there is a policy setting role
<Tasoshi_>
supply and demand - and their relationship to fees
<Tasoshi_>
now lets get back to the 25% miner
<Tasoshi_>
he can increase supply or reduce it
<Tasoshi_>
if demand remains constant then he is increasing or reducing fees
<Tasoshi_>
in effect, one person, setting policy
<Tasoshi_>
because the other miners have no choice
<gmaxwell>
Tasoshi_: sure, go look over the past week there are swinsgs of 25% even though the whole time 0.12 mempools had near 300mb backlogs. Variance, and policy already produce clearance rate differences on the kind of scale you're talking about with nothing breakin.
<Tasoshi_>
now they can join, but that is a race to the bottom, or a way of making bitcoin fully useless
<Tasoshi_>
gmaxwell, I am not sure what I am failing to explain here
<Tasoshi_>
25% variance and 200k daily txs to 100k all the sudden are very different things
<gmaxwell>
Tasoshi_: again, you seem to be making a logical leap that makes no sense. Yes, if 200k/day is the limit, then 25% hashpower can lower capacity to around 150k/day. Now you suggest that other miners are somehow forced to adopt his policy, but you've suggested no mechenism.
<Tasoshi_>
I know this is all political, but please appreciate facts
<gmaxwell>
Tasoshi_: in terms of how many transactions the network can clear, all causes there are equivilent. If mining variance reduce capacity it's no different for tx getting cleared than an equivilent reduction by some chunk of hashpower.
<Tasoshi_>
don't care how many txs the network can clear
<Tasoshi_>
not the point, at all
<gmaxwell>
There isn't anythin 'political' in this discussion.
<gmaxwell>
Tasoshi_: please-- I asked a direct question. You just insisted that other miners would somehow be compelled to follow the no-transaction miner in their behavior--- please, tell us what mechenism would enable that?
<Tasoshi_>
a super effective attack vector in the short term
<gmaxwell>
Tasoshi_: simply saying there is doesn't make it so.
<Tasoshi_>
let us go to an imaginary worl
<Tasoshi_>
and simplify matters
<Tasoshi_>
let us suppose that tx capacity is 200k txs a day and not 1 more
<Tasoshi_>
let us suppose no miner can increase this tx capacity
<gmaxwell>
several times in this discussion your simplfications have taken things completely out of connection with logic. (much less reality).
<gmaxwell>
Tasoshi_: sure
<Tasoshi_>
let us suppose we have a running system with people creating 200k txs a day, a fee market, and all is running
<Tasoshi_>
let us suppose someone who can decide in regards to say 50k of these txs now says I want higher fees, and thus reduces capacity to 150k per day
<gmaxwell>
yes, 25% hashpower could reduce that capacity by 25%.
<gmaxwell>
if he takes no transtions at all.
<Tasoshi_>
so you agree he can do that?
<Tasoshi_>
and no one else can do anything about it?
<gmaxwell>
Sure, thats how the system works. It's not some specific 50k tx he controls, to the system's transaction processing capacity it is exactly equivilent to him simply turning off. (at least for a couple weeks)
<kanzure>
Tasoshi_: was your expectation earlier today that this was not the case ... ?
<Tasoshi_>
how does he not control those 50k txs?
<Tasoshi_>
he can mine 0 blocks and deny those 50k txs, thus reducing system capacity to 150k no?
<helo>
the assumption interferes with your argument... there will never be "200k/day, exactly", there is the usual variance that users are accustomed to.
<gmaxwell>
Tasoshi_: because there is no mapping of transactions to miners. Let me pose an alternative, how would your situation differ from an alternative sitation where the same miner simply turned off his power and went home?
<arubi>
er, my head hurts because this means a miner acts against their own gain and discards fees, can this be established here?
<Tasoshi_>
you know how it would differ gmaxwell
<Tasoshi_>
that miner would no longer be mining blocks
<Tasoshi_>
there would be no 0 blocks
melvster has quit [Ping timeout: 276 seconds]
<gmaxwell>
Tasoshi_: so? that doesn't have any effect.
<gmaxwell>
(at least not for weeks)
<Tasoshi_>
of course it does
<gmaxwell>
Tasoshi_: What is the effect then?
<Tasoshi_>
a miner mining 0 blocks and a non existent miner are two different things
<gmaxwell>
arubi: sure, but it's hypotetical.
<Tasoshi_>
as you know
<arubi>
gmaxwell, yea, way out there..
dEBRUYNE has quit [Quit: Leaving]
<Tasoshi_>
and agree, I hope?
<gmaxwell>
Tasoshi_: no, actually in the short/medium term .. for transaction processing they are exactly the same.
<Tasoshi_>
not if we go a bit deaper
<tromp_>
no effect on tx/day. some effect on btc rewards/day
<Tasoshi_>
your miner does it once
<Tasoshi_>
my miner continues the game
<Tasoshi_>
he mines 0 blocks, increases the fees
<gmaxwell>
Tasoshi_: if tht 25% hashpower miner turns off one day, the same effect happens, roughly only 150k txn (in our hypothentical 200k constant load world) get processed.. same as the next day.
<Tasoshi_>
then he mines 250k blocks
dEBRUYNE has joined #bitcoin-wizards
<Tasoshi_>
or if he wants chaos he mines full blocks
<Tasoshi_>
now
<Tasoshi_>
do we not have a central bank governor?
gavinandresen has quit [Quit: leaving]
<gmaxwell>
Tasoshi_: go back. You've still not explained where you think the observable difference to users is on that day when the miner turns off.
<gmaxwell>
I think it's essential to clear that up before continuing.
<Tasoshi_>
the miner who turns off does it once
<Tasoshi_>
in a suicide
<Tasoshi_>
my miner is smart
<Tasoshi_>
he is setting policy
<gmaxwell>
Tasoshi_: he could turn off and on as much as he likes, sure'y you've seen a power switch.
<Tasoshi_>
he mines 0 blocks to increase fees
<Tasoshi_>
then 500k
<Tasoshi_>
then full, then whatever he wants
<Tasoshi_>
the crux is no other miner has a say over it
<Tasoshi_>
how is that decentralised?
<gmaxwell>
Tasoshi_: if he's mining 500k blocks than your capacity is back to 175k, fwiw. But again, you're still failing to distinguish turning off with mining nothing for this purpose.
<helo>
Tasoshi_: if everybody decides to do something abusive towards bitcoin, it can break.
<Tasoshi_>
there is no need for a distinguishment
<gmaxwell>
Can you please show the difference between turning off and mining nothing or agree they're equivilent?
<Tasoshi_>
it makes no difference
<Tasoshi_>
he turns it off he mines 0 blocks no one cares
<Tasoshi_>
same thing applies
<Tasoshi_>
and applies for 1 reason
<Tasoshi_>
the other miners have no say
<kanzure>
Tasoshi_: where are your "full blocks" coming from if 25% hashrate is turned off? those blocks will be absent.
<Tasoshi_>
full blocks basically give any miner a policy setting power
<harding>
I think Tasoshi_ is trying to argue that by mining empty blocks (or not mining at all) the fee rate will ratchet up so that when the miner starts mining larger blocks again, the fee rate will be higher. This seems to miss the point that the other miners -- his competition -- will capture all the high-fee transactions.
<gmaxwell>
Tasoshi_: and yes, sure a miner with 25% hashpower can swing capacity around by up to 25%-- e.g. by turning off. So? --- it isn't at all good that any >=25% hashpower miner exists, which I pointed out at the top-- but thats true for many reasons.
<Tasoshi_>
whether they do it by turning on and off or mining 0 blocks is irrelevant
<kanzure>
so you agree they are equivalent
<Tasoshi_>
gmaxwell, you are missing the crucial point of decentralization
<gmaxwell>
harding: yes, it wasn't clear to me above ... sounded like he was thinking there was a mapping from the miner to specific transactions?
<Tasoshi_>
no other miner has a say
<Tasoshi_>
they can't counteract this one miner
<kanzure>
miners always have a say about what they mine
<Tasoshi_>
they can't adapt
<Tasoshi_>
even the network can't adapt
<gmaxwell>
Tasoshi_: they have a say in their own behaivor. Not someone elses, which is itself a critical point of decenteralization.
<Tasoshi_>
it is basically 1 miner setting policy
<Tasoshi_>
no
<kanzure>
what is policy?
<Tasoshi_>
they dont have a say in their own behavior
<gmaxwell>
(and yes, this is a reason why 25% consolidations are themselves absurd... but they're absurd for reasons having nothing to do with capacity)
<Tasoshi_>
and what proves it is simple
<kanzure>
perhaps policy is not happening here?
<Tasoshi_>
what if the other miners want to increase capacity by 50k txs to counteract his decrease?
<Tasoshi_>
well, they can't
<gmaxwell>
Tasoshi_: they do--- that 25% miner, has (limited) control over 25% of the capacity. he deny it, or not, or anywhere in between. He has no control over the capacity of the other miners.
<Tasoshi_>
he does
<kanzure>
backlog size is independent of miner behavior
<Tasoshi_>
let us assume all other miners run at full capacity
<gmaxwell>
Tasoshi_: if they could "increase to offset" that would mean they had control over the 25% of the capacity that he influenced.
<Tasoshi_>
as far as the network is concerned, it is only the 25% miner who decides the overall capacity by increasing it or decreasing it as he pleases
<kanzure>
nope, miners cannot decide the size of the backlog
<Tasoshi_>
we are talking about a network
<gmaxwell>
Tasoshi_: he doesn't decide the overall capacity, he influences the roughly 25% in his pervue.
<Tasoshi_>
if a malicious miner decides to decrease capacity by 25%, the other miners should be able to respond
<Tasoshi_>
under the current system, however, they can not
<gmaxwell>
Tasoshi_: if they 'responded' then they'd have control of his 25% of the capacity then, a failure of decenteralization.
<Tasoshi_>
this opens the space for a clever miner to in effect set network policy, knowing other miners have no choice
<kanzure>
i think you mean that "transaction fees seem to be the only way to figure out prioritization"
<Tasoshi_>
25% of overall capacity
<helo>
(this never stops, by the way...)
<Tasoshi_>
lol
gavinandresen has joined #bitcoin-wizards
<Tasoshi_>
gmaxwell, how did you reach that conclusion?
<gmaxwell>
perhaps, for example, our 25% friend was not malicious, rather, but just required a higher fee to stay in operation. Your hypothetical would leave the others free to kick him out. And then they could kick out the next most efficient and so on, until only one remains (another hypothetical)
<Tasoshi_>
I mean, it is easy to argue whatever
<arubi>
really if all miners follow a "policy", then one miner "turning off" just cuts the bottom 25% of legit, non-spam transactions. smaller than 50k
<gmaxwell>
it's only easy to argue whatever for people who are unattacked to logical reasoning or reality. :)
<Tasoshi_>
but how did you reach the conclusion that miners acting in a decentralized maner suddenly control 25% of a miner's capacity?
<kanzure>
i really think there's some unreasonable assumptions being used here, like transaction reliability or that miners have control of the backlog, which is not true.....
<Tasoshi_>
no
<kanzure>
he said that in a decentralized system the 25% is not controlled by the other 75%
<kanzure>
you seem to have thought he said the opposite
<gmaxwell>
kanzure: yes, lots of weird assumptions-- but so? I think it's okay to talk about the limited case.
asux has quit [Ping timeout: 276 seconds]
<Tasoshi_>
he was arguing the exact opposite of what I said
<Tasoshi_>
while the situation is fully different
<Tasoshi_>
basically, unless I misunderstood, he said that if we turn the coin, then 75% will decide the capacity of the 25% miner
<kanzure>
gmaxwell: i think he is trying to argue that fee prioritization (or other prioritization) is unreasonable, but meanwhile it seems to be the only reasonable option for this system design
<Tasoshi_>
therefore, presumably, it is right that this 25% miner has that power?
<Tasoshi_>
but the situation is fully different
<Tasoshi_>
where the 25% miner decides the situation is limited, with no choice
<kanzure>
in a very vague and abstract way it's actually interesting that bitcoin can be used as a transaction fee to help determine priority, it's nice in the sense that it's a variable internal-to-the-system, rather than an external signal
<Eliel_>
To me it seems like Tasoshi_'s "attack" could be counteracted by a rolling block size limit that allows the next block + N past blocks to total (N+1)*1MB.
<Tasoshi_>
where the whole network of miners has a say, it is dynamic, and it is no longer anyone deciding the capacity of anyone, but everyone, in their complex way, deciding the overall network capacity
<gmaxwell>
Tasoshi_: if someone has 25% influence over the system, then thats what they have. It is a faulure of decenteralization for a single party to have that much; but it is what it is-- the situation is not improved by taking their influence. and handing to to another self-selecting party.
<Tasoshi_>
no
melvster has joined #bitcoin-wizards
<Tasoshi_>
they have what they have because of certain reasons
<kanzure>
Eliel_: there was another "vault" proposal sent to bitcoin-dev a few months ago. did you happen to see that?
<Tasoshi_>
primarely, under full blocks the other miners can not respond
<Tasoshi_>
they have no choice
<helo>
other miners keep doing what they were before the 25% dropped out
<Tasoshi_>
they can not counteract the malicious miner who decides to reduce capacity by increasing their own capacity
<Tasoshi_>
their hands are tied
<gmaxwell>
Tasoshi_: they only influence their own 'share' of the capacity; they can't influence someone elses.
<helo>
or they turn theirs off (for whatever reason, but they have no increased reason to as a result of the 25% drop)
<Tasoshi_>
and you know that, don't you?
<Tasoshi_>
no
<Tasoshi_>
they influence the network share
<Eliel_>
kanzure: I probably missed that. However, that rolling limit sounds like it'd change more than is immediately obvious, so I'm not sure if it's a good idea.
<Tasoshi_>
there is a whole system here
* gmaxwell
is tired of wasting his time.
gmaxwell has left #bitcoin-wizards [#bitcoin-wizards]
<Tasoshi_>
lol
<tromp_>
i'm suprised you laste this long, gmaxwell:)
<helo>
slow clap
<kanzure>
Tasoshi_: it's not an attack because that's how the system is supposed to work
<Tasoshi_>
some told me he was smart
<arubi>
whoever said this is #bitcoin material was really smart
<Tasoshi_>
but sometimes even the smart can't handle facts
<Tasoshi_>
especially the ones who cut halfway through
Tasoshi_ was banned on #bitcoin-wizards by jcorgan [*!*Tasoshi@unaffiliated/tasoshi]
<helo>
Tasoshi_: #bitcoin
belcher has joined #bitcoin-wizards
Tasoshi_ was kicked from #bitcoin-wizards by jcorgan [Tasoshi_]
<kanzure>
what was the evidence that he was aquentus?
<arubi>
kanzure, past sessions, talk of "the test of bitcoin", almost prophecy, cult-like
<JackH>
why is this guy such a pain to listen to? anyone knows who he is or what his problem is?
<arubi>
somewhat similar to that other nickname ^
<Eliel_>
it's difficult to talk to someone who refuses to budge from using their own weird definitions for things.
<kanzure>
Eliel_: i might have rejected that email or something, but i'm confused why i see it on the mailing list
<kanzure>
also for some reason i thought you were eyal or sirer
<bsm1175321>
kanzure: I keep making that mistake too.
<kanzure>
ah then it must be true
Eliel_ is now known as Eliel
<bsm1175321>
Nah, just a semantic name collision.
asux has joined #bitcoin-wizards
<Eliel>
I can get why you'd confuse me with Eyal, but ... sirer?
<kanzure>
well i confuse both eyal and sirer with each other
<kanzure>
*shrug*
<fluffypony>
wow backlog
<fluffypony>
and it's basically gmaxwell trying to help someone
* fluffypony
is never going to read all that
<Eliel>
painful reading, can't recommend.
<JackH>
the guy is insane, and creepy
<Eliel>
All you'll find is gmaxwell unsuccesfully trying to establish a common language with someone who doesn't want to.
<fluffypony>
yeah
<fluffypony>
re: gmaxwell's "It is a faulure of decenteralization for a single party to have that much; but it is what it is"
<fluffypony>
there were some REALLY good conversations I had this weekend with people at the Satoshi Roundtable about them taking on their own chunk of mining hardware, or running nodes spread out where they can host them
<fluffypony>
adam3us also said something similar on one of the panels
<fluffypony>
maybe at some point we forgot to keep educating people on how to retain Bitcoin's decentralisation
melvillian has quit [Remote host closed the connection]
priidu has joined #bitcoin-wizards
<Eliel>
Or perhaps, bitcoin just grew so fast that it all got lost in the noise.
domwoe has quit [Read error: Connection reset by peer]
<JackH>
nah its capitalism in its purest form
* kanzure
weeps for the 56,000 words he typed from the roundtable (bug bruce if you want him to release that)
<fluffypony>
that's the other side of the coin, isn't it - why should CoinBase spend the money to run a chunk of mining hardware when they'll feel like it's a waste of money
<fluffypony>
and you can't convince them otherwise
<fluffypony>
kanzure: I still can't believe the quality of your raw transcribes, it's insane. You must be pushing out well north of 150 WPM?
<amiller>
#bitcoin-wizards-offtopic for typerace tournament chatter
phiche has joined #bitcoin-wizards
Erik_dc has quit [Remote host closed the connection]
glitch003 has quit [Read error: Connection reset by peer]
* midnightmagic
golf-claps for jcorgan
<midnightmagic>
but non-sarcastically
domwoe_ has joined #bitcoin-wizards
domwoe has quit [Read error: Connection reset by peer]
<BitcoinErrorLog>
lol
<BitcoinErrorLog>
any news about transcripts being released? is there anything people would actually find interesting aside from novelty?
<BitcoinErrorLog>
oh you just said bug bruce, nvm
glitch003 has joined #bitcoin-wizards
melvillian has joined #bitcoin-wizards
<fluffypony>
BitcoinErrorLog: maybe it never happens, in which case the live Reddit thread that Shawn and I did is a reasonable overview
<BitcoinErrorLog>
did bruce complain about that too?
Yoghur114 has quit [Remote host closed the connection]
<BitcoinErrorLog>
between the reddit, twitter, and all the first-hand accounts heard over the weekend, it was almost .... like i was there
snthsnth has joined #bitcoin-wizards
tr0nk has joined #bitcoin-wizards
<fluffypony>
no he didn't, but we did tell everyone we were doing it
<fluffypony>
BitcoinErrorLog: you were there in spirit, buddy
zmachine has quit [Remote host closed the connection]
spinza has quit [Ping timeout: 244 seconds]
ThomasV has joined #bitcoin-wizards
spinza has joined #bitcoin-wizards
arowser has quit [Quit: No Ping reply in 180 seconds.]
arowser has joined #bitcoin-wizards
mrkent has joined #bitcoin-wizards
melvillian has quit [Remote host closed the connection]
melvillian has joined #bitcoin-wizards
<JackH>
does anyone know of a place that actually has collected all white papers of pre-era Bitcoin? that speak about systems similar to Bitcoin?
<JackH>
because some past ideas were actually quite good, and interesting to read about, like Szabo's bitgold and similarly
tr0nk has quit [Ping timeout: 248 seconds]
<fluffypony>
I don't know if anyone's made a collection of them - maybe kanzure?
glitch003 has quit [Remote host closed the connection]
<harding>
JackH: I think the collection here is pretty good, although it goes beyond crypto into economics and a few other topics: http://nakamotoinstitute.org/literature/
snthsnth has quit [Quit: leaving]
CrazyTruthYakDDS has joined #bitcoin-wizards
spinza has quit [Ping timeout: 246 seconds]
glitch003 has joined #bitcoin-wizards
maaku has joined #bitcoin-wizards
maaku is now known as Guest98911
domwoe_ has quit [Ping timeout: 248 seconds]
Guest39448 is now known as [Derek]
[Derek] has quit [Changing host]
[Derek] has joined #bitcoin-wizards
<instagibbs>
Adam Back's Hashcash website has a few