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
Chris_Stewart_5 has quit [Ping timeout: 272 seconds]
Chris_Stewart_5 has joined #bitcoin-wizards
<sipa>
midnightmagic: heh, never saw that
frankenmint has joined #bitcoin-wizards
jiggalator has joined #bitcoin-wizards
frankenmint has quit [Remote host closed the connection]
Chris_Stewart_5 has quit [Ping timeout: 250 seconds]
andytoshi has joined #bitcoin-wizards
<midnightmagic>
sipa: :)
bsm117532- has joined #bitcoin-wizards
<bsm117532->
So it's occurred to me several times that with Schnorr signatures, it's possible to combine all P2PKH transactions in a block. Does any existing currency do this? Monero? Dash?
<bsm117532->
e.g. a single signature of witness data per block...
<othe>
dash is a plain bitcoin copy, theres no different crypto stuff used at all
<othe>
i dont think thats done by any of the cryptos at all
<othe>
we use some aggregated schnorr signatures in ringct but not in the context u mentioned tho
<bsm117532->
Interesting... it reduces the P2PKH data to a single signature, making it extremely compact.
<maaku>
bsm116532 that would require an interactive step
<bsm117532->
I was looking at Ethereum's transaction graphs today...95% of addresses are PKH, not contracts. Even there, most transactions are moving the underlying coin, not contracts...
<bsm117532->
@maaku can you elaborate?
Chris_Stewart_5 has joined #bitcoin-wizards
<maaku>
With pairing crypto you can do it non interactively, and that is what OWAS is about
<maaku>
Bsm, how do you combine signatures of different messages?
SyncYourDogmas has quit [Quit: Leaving]
<bsm117532->
What's OWAS?
MaxSan_ has quit [Quit: Leaving.]
MaxSan_1 has joined #bitcoin-wizards
<bsm117532->
Correct me if I'm wrong, but Schnorr signatures can be summed.
<gmaxwell>
as maaku mentioned, schnorr aggregation requires interaction, it's a two phase protocol, first agree on the nonce, then form the signature.
<bsm117532->
I wish I had been at this meeting...
<gmaxwell>
BLS signatures can do the same thing without interaction.
<gmaxwell>
all of that has been discussed in this channel and on bitcointalk before.
<bsm117532->
@gmaxwell is it possible to take a well-known nonce (like last block hash)
<gmaxwell>
The _funny_ thing is that we measured the aggregation gain for existing transactions in the network, and found it not that much smaller than the maximum gain from all being aggregated.
<bsm117532->
Yeah, I'm (relatively) new here. And won't set foot in bitcointalk... :-/
<gmaxwell>
bsm117532-: no, the nonce has to be secret or you reveal private keys.
<gmaxwell>
you can read bitcoin talk without posting there.
<gmaxwell>
and generally the technical subforum is pretty good
<bsm117532->
Is revealing an aggregate privkey a vulnerability? Wouldn't it require all-but-1 colluding participants to exploit?
CubicEarth has quit [Remote host closed the connection]
<gmaxwell>
I think the figures we got were something like 40% gain assuming everything was aggregated vs 30% on the existing transation load.
<gmaxwell>
bsm117532-: this is unrelated to aggregation, if I give you a DSA signature where you know the nonce (or even where you don't know it, but I produced another signature with the same nonce before) then you can immediately recover my private key.
<gmaxwell>
(or perhaps I should have said more general instead of unrelated.)
<bsm117532->
@gmaxwell what additional assumptions go into that? The figure I saw the other day was ~87% P2PKH...
<gmaxwell>
bsm117532-: 87% what?
<gmaxwell>
are you saying 87% could be aggreated? the figure I gave is byte savings.
<bsm117532->
OK then those numbers jive. I was extrapolating from 13% P2SH usage.
<gmaxwell>
aggregation only avoids signatures, not the other bits of data in a transaction, so the byte savings is less than the total amount.
<maaku>
You'd still have to be signing the same thing, which different transactions or not
Ylbam has quit [Quit: Connection closed for inactivity]
<gmaxwell>
P2SH doesn't diminish the potential of aggregation gains, since most p2sh is multisig it increases it.
<gmaxwell>
s/most/virtually all/
<bsm117532->
Don't worry gmaxwell, I have a cool new use for the hash preimage script... ;-)
<bsm117532->
Ok, off to read transcripts, thanks @kanzure, @gmaxwell
jee56 has joined #bitcoin-wizards
rusty2 has joined #bitcoin-wizards
CrazyTruthYakDDS has quit [Quit: Connection closed for inactivity]
jiggalator has quit []
AndChat-7049 has joined #bitcoin-wizards
bsm117532- has quit [Ping timeout: 250 seconds]
AndChat-7049 has quit [Ping timeout: 240 seconds]
bsm117532- has joined #bitcoin-wizards
belcher has quit [Quit: Leaving]
roman__ has quit [Read error: Connection reset by peer]
CubicEarth has joined #bitcoin-wizards
c0rw1n_ has quit [Ping timeout: 240 seconds]
Aranjedeath has joined #bitcoin-wizards
CubicEarth has quit [Remote host closed the connection]
mdavid613 has quit [Quit: Leaving.]
Chris_Stewart_5 has quit [Ping timeout: 240 seconds]
Chris_Stewart_5 has joined #bitcoin-wizards
Chris_Stewart_5 has quit [Ping timeout: 246 seconds]
CubicEarth has joined #bitcoin-wizards
Noldorin has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
chjj_ has quit [Ping timeout: 244 seconds]
jiggalator has joined #bitcoin-wizards
hashtag has quit [Ping timeout: 244 seconds]
mdavid613 has joined #bitcoin-wizards
CrazyTruthYakDDS has joined #bitcoin-wizards
HostFat has quit [Quit: Leaving]
Tiraspol has quit [Remote host closed the connection]
Tiraspol has joined #bitcoin-wizards
Burrito has quit [Quit: Leaving]
CubicEarth has quit [Remote host closed the connection]
CubicEarth has joined #bitcoin-wizards
chjj_ has joined #bitcoin-wizards
pro has quit [Quit: Leaving]
jee56 has quit [Read error: Connection reset by peer]
jee56 has joined #bitcoin-wizards
Giszmo has quit [Quit: Leaving.]
frankenmint has joined #bitcoin-wizards
veleiro has joined #bitcoin-wizards
ThomasV has joined #bitcoin-wizards
CubicEarth has quit [Remote host closed the connection]
go1111111 has quit [Ping timeout: 276 seconds]
jiggalator has quit [Remote host closed the connection]
jiggalator has joined #bitcoin-wizards
jiggalator has quit [Client Quit]
TheSeven has quit [Disconnected by services]
[7] has joined #bitcoin-wizards
dingus has quit [Ping timeout: 240 seconds]
Chaterz has quit [Ping timeout: 250 seconds]
copumpkin has quit [Ping timeout: 260 seconds]
othe has quit [Excess Flood]
copumpkin has joined #bitcoin-wizards
othe has joined #bitcoin-wizards
dingus has joined #bitcoin-wizards
CubicEarth has joined #bitcoin-wizards
chjj has quit [Ping timeout: 244 seconds]
CubicEarth has quit [Ping timeout: 246 seconds]
CubicEarth has joined #bitcoin-wizards
chjj has joined #bitcoin-wizards
mdavid613 has quit [Quit: Leaving.]
sneak has joined #bitcoin-wizards
sneak is now known as Guest59696
ThomasV has quit [Ping timeout: 240 seconds]
CubicEarth has quit []
<gmaxwell>
bsm117532-: can't quite figure out the falcon-net page? is that a commercial service?
jee56 has quit [Quit: Leaving]
<gmaxwell>
(more tools in that space are good, in any case)
<gmaxwell>
but the claims about relay network being unmaintained and what not are untrue. (though matt has tried pretty hard to get people to setup alternatives, including by threatening to stop maintaining it :))
<sipa>
having another relay network, curated by a different group, is not bad
<gmaxwell>
not not bad, it's fantastic.
<gmaxwell>
The rest I dunno about, and don't care. :)
<sipa>
so, what is the problem? were they expecting their not-yet-announced application-level-cut-through-routing to be integrated into bitcoin directly?
<gmaxwell>
I emailed Gun, and had a back and forth discussion. His position is perplexing to me. It appears to me that he is basically referring to the whole bitcoin industry (not just Bitcoin Core) as blockstream. I wrote him a long list of people who have no asscoiation with blockstream and came up with items in the core capacity roadmap, and he went through it and said that Christian Decker was the on
<gmaxwell>
ly person who wasn't a blockstream contractor.
<gmaxwell>
He then gave me a list of things we supposidly ignored, which was mostly things first published in feburary, months after that capacity roadmap, most not related to capacity/scaling.
<gmaxwell>
He also complained that bitcoin-ng was not part of that roadmap. In my response I pointed out that the preprint was first published a few weeks before that roadmap, and that it's a pretty radical change to the system.
Transisto2 has joined #bitcoin-wizards
<gmaxwell>
(I continue to think we can get the same benefits without giving miners (ephemeral) identities as bitcoin-ng does)
<gmaxwell>
so I don't get it, I think he may have unrealistic expectations at the pace that academic research can enter production.
<rusty2>
gmaxwell: he may also underestimate the other low-hanging fruit before we consider such a radical change.
<qpm>
tx:<Jeremy_Rand> gmaxwell: In my experience, people who complain about "problems being ignored" usually either are confused or have an agenda. To the point that I usually don't engage such people anymore. Props to you for engaging him.
<gmaxwell>
I have a lot of respect for him and think basically everything he works on is interesting. It's really demoralizing to get that kind of comment, I think it's predicated on misunderstandings and I don't want to write it off.
<gmaxwell>
There is also the issue that his attitude will potentially be impressed on his students.
<gmaxwell>
rusty2: and perhaps overestimates how many people who are working on these things in earnest on the bitcoin side.
<rusty2>
gmaxwell: very good point.
rusty2 is now known as rusty
ThomasV has joined #bitcoin-wizards
<sipa>
qpm: what is that <tx> ?
<gmaxwell>
and just the general cadance, e.g. he complained that no action has been taking relative to http://fc16.ifca.ai/bitcoin/papers/MES16.pdf ... but a big part of the SW motivation is to make further script extensions deployable without as much work as CLTV/CSV took.
<qpm>
tx:<Jeremy_Rand> sipa: Tx is a private Tor-only IRC network; qpm is a relay between Tx and Freenode.
<sipa>
ah
<rusty>
gmaxwell: That sounds like him being overly optimistic about how fast bitcoin moves, yeah.
<sipa>
was bitcoin-ng ever implememted?
<gmaxwell>
I don't know. They presented simulation results, they have not published code. We could ask Ittay.
<gmaxwell>
man, I didn't notice this before but their freeking bitcoin-ng announcment has this kind of junk in it.
<gmaxwell>
---
<gmaxwell>
In Context
<gmaxwell>
Since the Bitcoin world is rife with conflicts of interest, there is value in reproducing part of the announcement that accompanied this note on the bitcoin developers list:
<gmaxwell>
NG is compatible with both Bitcoin as is, as well as Blockstream-like sidechains, and we currently are not planning to compete commercially with either technology -- we see NG as being complementary to both efforts. This is pure science, published and shared with the community to advance the state of blockchains and to help them reach throughputs and latencies required of cutting edge fintech app
<gmaxwell>
lications. Perhaps it can be adopted, or perhaps it can provide the spark of inspiration for someone else to come up with even better solutions.
<gmaxwell>
---
gsdgdfs has joined #bitcoin-wizards
Transisto2 has quit [Ping timeout: 246 seconds]
paveljanik has quit [Quit: Leaving]
rusty has quit [Ping timeout: 272 seconds]
MaxSan_1 has quit [Ping timeout: 264 seconds]
MaxSan_ has joined #bitcoin-wizards
raedah2 has joined #bitcoin-wizards
raedah has quit [Ping timeout: 246 seconds]
<Taek>
It was a breath of fresh air to go to the decentralized web summit. The attitude was a lot more collaborative, to the point I hadn't realized how difficult things were at Bitcoin conferences.
<Taek>
It's helps a lot that most of the tech there didn't rely on consensus, different implementations could coexist a lot easier than they can within the bitcoin ecosystem
<qpm>
tx:<Jeremy_Rand> Taek: it was my first time at a conference representing Namecoin. I also enjoyed it a lot. I'm a bit sad to hear you say that most Bitcoin conferences aren't as collaborative; that's unfortunate.
<Taek>
My first conference was after the blocksize rift happened, I can't speak to the environments before that
<gmaxwell>
before the scaling conferences there was very little in the way of bitcoin conferences with significant technical content.
<qpm>
tx:<Jeremy_Rand> Taek: ah, I see. Yeah, I have to admit I'm enjoying the fact that Namecoin doesn't have any equivalent controversy going on. I don't know how the Bitcoin people can handle the craziness. Hats off to everyone who's still getting productive work done.
<Taek>
I think most of the tension, even in the technical half of the community, is rooted around investments that people have made into Bitcoin. When half of your savings are invested into an asset, it puts you on edge to see things you feel might threaten the project
<Taek>
gmaxwell: were there moments at Wikipedia that parallel the blocksize debates?
<qpm>
tx:<Jeremy_Rand> Taek: that is a good point. And might also explain why Namecoin seems to be unaffected so far -- none of the Namecoin developers have a large investment in NMC.
<qpm>
tx:<Jeremy_Rand> (well, and also the fact that NMC's market cap is too small for anyone to care)
<gmaxwell>
Taek: there has been drama in many project's I've seen (also none in others)-- each case is pretty different.
<qpm>
tx:<Jeremy_Rand> My sample size is really small, but cryptocurrency projects have the most drama of any projects I've been even loosely involved in.
<Taek>
I seem to remember that some pretty terrible things happened with ffmpeg/libav
<gmaxwell>
yes, well... thats to say the least.
<gmaxwell>
Bitcoin drama is exacerbated by the Yahoo Stock Forums factor. If in 2001 you asked me where the most toxic place on the internet was, I'd have easily answered the yahoo stock trading forums.
<gmaxwell>
Seems like a unique mixture of the kind of personalities who are not just investing, but so opinionated about it that they want to argue with each other, with their positions backed up by large bets, and perhaps even the hope that their trolling will influence the price...
<gmaxwell>
Bitcoin picks up some of that.
<gmaxwell>
Another factor is that bitcoin is an awful big tent-- drawing in people from all kinds of perspective, a necessary thing for a money to do-- means that there is a less of a common basis to discuss things from.
<Taek>
2001 was after the collapse of the dot-com bubble. Did that play into it? A lot of leftover tension from failed investments.
<fluffypony>
libertards are gonna libertard.
<gmaxwell>
It was also toxic before, actually I think the upswings mostly made it worse.
<fluffypony>
and capitalists are gonna capitalist, I guess
<gmaxwell>
There were some pretty big disputes between wikipedians early on, but they usually shared a lot more common basis than random bitcoin people will.
<gmaxwell>
Similar to disputes in standards orgs.
<fluffypony>
I think that the situation is exacerbated by a lack of technical understanding among most of the participants
<gmaxwell>
In WP but unlike my expirence in standards orgs, a lot of the "dispute" incidents were driven by clear outsiders. Like when scientology started a major campaign against wikipedia. This ultimately increased community cohesion, defending against a common threat.
<Taek>
Makes sense that diverse backgrounds would increase strife
<fluffypony>
so you end up with people feeling like something makes "more" sense to them, without an understanding of the technical complexities / perverse incentives / attack vectors involved
<qpm>
tx:<Jeremy_Rand> As a privacy-interested dev who basically ended up in cryptocurrency dev by accident, I have to admit that it was kind of jarring seeing the bad behavior that shows up in cryptocurrency circles, whether it's caused by money being involved or just a wider tent than I'm used to
<fluffypony>
Jeremy_Rand: it's the money thing
<gmaxwell>
In SDOs there are sometimes the concern that people are trying to undermine efforts of competing technology, but in those places they have anti-sockpuppetry measures that mean that the allegiences are clear. Trust in Bitcoin gets unmined somewhat for lack of that. (like, is this dude arguing with me in good faith, or is he just trying to promote some altcoin he's invested in?)
<Taek>
I wonder what sort of external threat to Bitcoin would be big enough to draw everyone together but small enough to leave the project intact.
<fluffypony>
never seen anything like this in FOSS projects, even ones that were poisonous and had an idiot savant leader
[7] has quit [Ping timeout: 250 seconds]
TheSeven has joined #bitcoin-wizards
<gmaxwell>
Most of the time people who don't like the technical stuff ignore the technical stuff.. and all are happy. But I think the Yahoo Stock Factor makes everyone opinioned about everything. :)
<gmaxwell>
Falkvinge's comments on reddit the last couple days are an interesting example. Super opinionated that segwit is totally unneeded and that malleaiblity will somehow fix itself by the wisdom of crowds.
<fluffypony>
oh the other thing is that technical criticisms are easily handwaved with accusations of "FUD"
<Taek>
You necessarily think about it more, which gives you more confidence in your opinion, making you less likely to realize that the guy across from you has 5 years of expert experience on you.
<Taek>
*Yahoo Stock Factor
<gmaxwell>
This is the beer-cup-hat factor that I lamented in that post, which a lot of people didn't like, and found to be unduly insulting; but I think it captures one element of things that go on.
<gmaxwell>
we also have pressure to express no doubt, actually--- this is perhaps one of the good things about the blocksize-saga.
<gmaxwell>
Some of the most vicious people in the community that would tolerate no technical person discussin a risk in bitcoin technology ended up on the must-hardfork-now side and themselves have been blah blah blah about risks from not hardforking for months. (Cypherdoc, being a good example of that). So perhaps in the future when someone talks about some risk that crew will be less likely to come out o
<gmaxwell>
f the woodwork with pointed sticks.
<qpm>
tx:<Jeremy_Rand> Taek: btw, I don't recognize your username -- sorry if this is a stupid question, but did we meet in person at the Decentralized Web Summit? (Trying to jog my memory.)
<Taek>
David Vorick, Sia, we may have conversed briefly, there were a lot of new faces and I didn't get to everyone
<qpm>
tx:<Jeremy_Rand> Taek: ok cool -- I'm pretty sure that if we talked, it was very brief. Same here, it was impossible to talk to everyone, despite my attempts to do so. :)
AaronvanW has quit [Remote host closed the connection]
MoALTz has quit [Ping timeout: 276 seconds]
AaronvanW has joined #bitcoin-wizards
Inaltoas1nistra has joined #bitcoin-wizards
CrazyTruthYakDDS has quit [Quit: Connection closed for inactivity]
Inaltoasinistra has quit [Ping timeout: 272 seconds]
MoALTz has joined #bitcoin-wizards
Ylbam has joined #bitcoin-wizards
dEBRUYNE has joined #bitcoin-wizards
go1111111 has joined #bitcoin-wizards
<midnightmagic>
sipa: sorry. I forgot to ask permission in here while I was in #freenet asking.
<midnightmagic>
qpm is a relay bot between a Tor-only IRC network I built, and here. It's in a number of channels; given that I am 100% pro the insta kickban of a misbehaving remote nework (and given only a small handful of people are even using it right now) does anyone have any objections to qpm sitting here doing two-way relay?
<midnightmagic>
(or just +q I suppose)
<midnightmagic>
I am also oper over there: so problem users can be dealt with. Currently I don't expect issues because it's invite-only. But they're pretty regular users.
<midnightmagic>
I run qpm as well.
<gmaxwell>
thanks for creating that, I'm really disappointed with freenode's poor handling of tor.
<midnightmagic>
You're welcome to use it, aside from the inevitable netsplits it's been fairly reliable. Plus there's me, and I have not much else to do aside from respond to outages
<midnightmagic>
I'd say.. what, Jaremy? Maybe 98% at this point?
<midnightmagic>
oh, except for those days.
<midnightmagic>
Anyway if anyone can think of any objections, please let me know. I don't want to cause any problems.
frankenmint has quit [Remote host closed the connection]
akeyla420 has quit [Quit: Leaving]
<fluffypony>
midnightmagic: same reason we have MRL-Relay routing to a private network, as gmaxwell indicates there's not much of an alternative right now
<midnightmagic>
true enough.
murch has joined #bitcoin-wizards
<gmaxwell>
unfortunately it seems that bitcoin-otc was a major driver in freenode policy--- they stopped allowing tor because of nickserv brute forcing.
<midnightmagic>
we were, apparently hated and tolerated in about 50% portions.
Guyver2 has joined #bitcoin-wizards
c0rw1n_ has joined #bitcoin-wizards
dEBRUYNE has quit [Quit: Leaving]
<fluffypony>
midnightmagic: so like Bitcoin, then :-P
Guyver2 has quit [Ping timeout: 260 seconds]
alferz has quit [Ping timeout: 240 seconds]
alferz has joined #bitcoin-wizards
ThomasV has quit [Ping timeout: 260 seconds]
mrpocoyo has joined #bitcoin-wizards
Guyver2 has joined #bitcoin-wizards
Logicwax has quit [Ping timeout: 240 seconds]
Logicwax has joined #bitcoin-wizards
NewLiberty_ has quit [Ping timeout: 258 seconds]
ThomasV has joined #bitcoin-wizards
mrpocoyo has quit [Ping timeout: 276 seconds]
atgreen has quit [Ping timeout: 252 seconds]
ThomasV has quit [Ping timeout: 244 seconds]
nanasho has joined #bitcoin-wizards
totemizer has joined #bitcoin-wizards
<totemizer>
obviously in the wizards channel there are more people than in the marketing channel...
<totemizer>
can anyone from here give me some links to blockchain related benchmarks? I depleted google and several other sources but there are only a really few I could find, maybe I am missing something
spinza has quit [Ping timeout: 260 seconds]
<murch>
totemizer: What properties are you trying to find statistics on? It might be helpful to ask on http://bitcoin.stackexchange.com
hashtag has joined #bitcoin-wizards
<totemizer>
thanks for the suggestion
<totemizer>
i am mostly interested in technical stuff, performance related things, anything to put a number on it really
<totemizer>
I came here instead of stackexchange because I haven't had a concrete question to put it there
<sipa>
perfoance of what?
<sipa>
blockchains are easy, maintaining a ledger is not
<totemizer>
performance of a blockchain, but it seems our terminology doesn't match up. i think of blockchain as a distributed database
<sipa>
it's not
<sipa>
a blockchain is a transaction log
<sipa>
it defines a database by replaying it
<totemizer>
right, so as I said our terminology doesn't match up, I hope this doesn't get into the way of communicating effectively
<sipa>
well it would help to clarify what you want to know about
<sipa>
"distributed database" is pretty vague
skang404 has joined #bitcoin-wizards
<fluffypony>
totemizer: also "blockchain" is not generic
<totemizer>
sipa that is only because the words mean different things to you. I to me the core concept I care about is that the nodes are equals and there is no predetermined center to rely on, and that it stores data.
<totemizer>
fluffypony: what do you mean by generic?
<fluffypony>
totemizer: the performance benchmarks would differ across varying implementations of the same cryptocurrency, much less across different blockchain applications
<fluffypony>
eg. compare Bitcoin's performance on bdb vs. leveldb
<sipa>
yes, actual implementations use an actual local database
<totemizer>
fluffypony: I realize that the performance benchmarks would differ a lot, that is why I expected to find a lot of them, not a few...
<totemizer>
yeah, i do not meant "node performance" obviously
<sipa>
totemizer: what kind of benchmark do you mean? for what kind of operation?
<sipa>
"database updates per second" ?
<totemizer>
i meant "network performance". I was expecting stuff like "using this kind of consensus algorithm, this kind of implementation, etc etc etc technical details, these are the numbers..."
<totemizer>
it seems that there is no one doing this
<totemizer>
is this not something meaningful to do?
<sipa>
it is not without specifying what you mean by blockchain
<sipa>
if you're referring to a proof-of-work consensus algorithm, that pretty much only makes sense in case there is a financial incentive to cooperate
mrpocoyo has joined #bitcoin-wizards
<sipa>
and its performance depends on what the participants' nodes can tolerate
<sipa>
plus the inequality you're willing to live between between worse and better connected nodes
<fluffypony>
yeah so Bitcoin could "scale" to thousands of transactions per second
<fluffypony>
as long as each node was in a datacenter and had tons of highly performant hardware + lots of low-latency bandwidth between them
<sipa>
right, it depends on what the nodes and the connections between them can do
<sipa>
not on the kind of "blockchain" used
<totemizer>
sipa: I am not sure how else to specify it differently than what I already wrote: core concept I care about is that the nodes are equals and there is no predetermined center to rely on, and that it stores data.
<sipa>
well you can't use proof of work for that, as there is no incentive to cooperate
<sipa>
nodes don't lose anything by working adverserially
<sipa>
you can use a traditional consensus algorithm like paxos or pbft for that
<sipa>
if the nodes have known identities
<sipa>
in which case "blockchain" is just a new hyped name for something 20 year old
<sipa>
(for example, google uses paxos based distributed databases internally, with very high performance, and they don't have any fixed center)
<totemizer>
sipa: I see. I think that when asking about technical performance benchmarks I can ignore the theoretical part about incentives and malicious nodes and just ask what is the case in the optimal situation, as most benchmarks do that anyway.
<fluffypony>
totemizer: the problem is in defining "optimal"
mrpocoyo has quit [Ping timeout: 272 seconds]
<sipa>
totemizer: if you only care about the optimal situation, you don't need a blockchain, use paxos
<totemizer>
optimal in this case is that no nodes try to act adverserially and every node does the same thing
<sipa>
use paxos
<totemizer>
sipa: paxos?
<sipa>
google it
<totemizer>
oh, i did, i am learning, thanks
MaxSan_ has quit [Ping timeout: 240 seconds]
<sipa>
if nobody will act adverserially, you don't need a blockchain at all
<fluffypony>
blockchains are expensive and horribly inefficient, but for a reason
<totemizer>
sipa: ok, so for you blockchain has no meaning unless there are malicious nodes?
<sipa>
totemizer: to me, it is a data structure
<totemizer>
fluffypony: i know that, but thanks for pointing it out
<totemizer>
sipa: same with me, but i can only hear about the worst case scenarios
<sipa>
totemizer: but using it to define a distributed database only makes sense to establish consensus between distrusting partoes
<murch>
I've been trying to find that blog from a bitcoin startup that detailed various cases when companies tried to hire them. There was a nice overview of what else you should be using and why each of these cases didn't fit the blockchain use-case. Can't find it though. Does that ring a bell for one of you?
<sipa>
where you have the distinction between the case where the parties are known (use pbft) or are unknown and you'll use economic properties instead (proof of work)
<totemizer>
sipa: i am still with you, distrusting doesn't mean malicious though
<sipa>
totemizer: well in bitcoin parties assume that every single party can be malicious
<murch>
totemizer: Well, if the other won't act maliciously, why would you distrust them? o.0
<sipa>
totemizer: if yiu assume none will be malicious, you have no reason to distrist
<totemizer>
so the system does consider it, but i want to know what is happening when there are no "active" nodes like that :)
<sipa>
totemizer: sure, you can talk about a system where the system can deal with malicious parties, but you're talkong about benchmarks where none (or only a small number) are
<murch>
totemizer: Perhaps you have a look at that and you'll probably better understand why you're getting the answer you're getting here.
<totemizer>
sipa, murch it is a false dichotomy. it is completely possible to imagine a situation where you have a blockchain built with no trust in mind between the nodes but between time A and time B no active nodes act maliciously. theoretically this is possible i think
<sipa>
totemizer: no, i'm absolutely with you
<sipa>
totemizer: in fact, that is common
<totemizer>
i read that blogpost last week, but thanks :)
<sipa>
if the system is designed to deal well with malicious parties, it is likely that none will actually act maliciously, as there is nothing to gain
<totemizer>
sipa: ok, so what am I missing still?
<totemizer>
sipa: yeah, that was my thoughts too
<sipa>
totemizer: bottom line: if you want a blockchain to define a database where parties don't trust each other, performance is limited by what the weakest participant in the system can validate
<totemizer>
yay, thanks
<sipa>
because blockchains solve the trust problem in the absolutely most naive way possible: everyone validates everything
<totemizer>
yeah, you see, now you telling it, i find it logical, but i am just learning still so it was not obvious before if there is no other factor even more restraining
<totemizer>
i can imagine for example exactly PoW to be limiting of course :)
<fluffypony>
totemizer: also to add, the block size limit artificially limits the performance so that the weakest participant isn't cut off
<totemizer>
yeah, that is a fun thing with bitcoin
<sipa>
if you don't need that, and are willing to trust that all faults will be only due to race conditions and lost/slow communication, but not actively malicious parties, use paxos
<totemizer>
what do you think segregated witness will be adopted?
Chris_Stewart_5 has joined #bitcoin-wizards
<sipa>
i will quit if it doesn't :)
<totemizer>
i am reading up on paxos for sure, really good that you pointed it out, it was in my blindspot
<totemizer>
read that too, but i like asking other people :D
<sipa>
well, i'm biased
<fluffypony>
I think that's more a #bitcoin-dev question
<murch>
fluffypony: Not updated since 6th of May. ;)
<murch>
I hope it's not up-to-date because then none of those listed finished integrating support since then.
<fluffypony>
oh lol, didn't even notice that
<murch>
fluffypony: Well, at least nobody has posted on the corresponding github issue since 6th of May.
spinza has joined #bitcoin-wizards
<totemizer>
yeah, there are two pull requests and there is a comment on some change being needed but it is so ezoteric that I have no idea what it is about ÉD
<sipa>
the ml discussion about extending the size of witness programs?
<instagibbs>
murch, yes, it needs to get updated, but volunteer effort :P
<totemizer>
sipa: I don't think so
ThomasV has joined #bitcoin-wizards
<murch>
instagibbs: I don't mean to diss anyone's effort, I'm just trying to point out that the site is likely not representing the current stage of SegWit progress accurately. :)
pro has joined #bitcoin-wizards
NewLiberty has joined #bitcoin-wizards
dnaleor has quit [Ping timeout: 244 seconds]
atgreen has joined #bitcoin-wizards
<bsm117532>
gmaxwell: Re: bitcoin-ng I like the idea in principle, but it's a merging of two disparate technologies. I'd prefer to see Bitcoin stick with Nakamoto-style consensus rather than make a hybrid. That's what I'm trying to do with my braids project.
<fluffypony>
bsm117532: last sentence is a bit ambiguous - do you mean you're trying to stick with Nakamoto-style consensus with braids, or make a hybrid with braids?
<bsm117532>
I've been concerned for some time that academia in general doesn't have the patience to shepherd such a change through the process of getting it into core. I've seen zero efforts (post announcement and paper) to get shepherd it (or several other academic projects).
<bsm117532>
It's time consuming to build a testnet, shepherd patches/PR's, and the incentive structure in academia is to get the paper, not do the subsequent shepherding.
<bsm117532>
With OSS projects there's always this nasty expectation that someone else will do the work for you... :-/
<bsm117532>
fluffypony: I'm sticking to Nakamoto-style consensus. That is, leaderless, and not identifying nodes (which pulls in BFT considerations).
<bsm117532>
Personally I'm seriously wondering how I'm going have the time to shepherd my own ideas...but you won't find me bitching that Blockstream isn't doing it for me... ;-)
<totemizer>
bsm117532: but where are the papers then though?
<totemizer>
i read the lightning network paper, or at least part of it. I don't have an academic training so maybe I shouldn't say this, but I find it really convoluted compared to the actual content
<fluffypony>
totemizer: read Rusty's write-up on it, it's a lot easier to digest
<fluffypony>
bbiab, have to go buy a new stepladder
<instagibbs>
bsm117532, what you said about shepherding is spot-on. Write the code, make an NG testnet, get feedback, etc.
<bsm117532>
totemizer: I'm giving a talk about braids next week. Paper will follow shortly thereafter. The ideas are basically done, it's time to build a braided testnet.
<bsm117532>
I know I know, I've been talking about this for a long time. ;-)
<bsm117532>
It's really almost done this time, I promise! ;-)
<sipa>
the ideas for segwit were done in november :)
ThomasV has quit [Ping timeout: 240 seconds]
<totemizer>
ideas are always done :D
<totemizer>
i can give zillion ideas that are done
<bsm117532>
I had to do quite some numeric analysis for this one. ;-)
<bsm117532>
Grahps can be complex.
frankenmint has joined #bitcoin-wizards
<totemizer>
bsm117532: well, i hope your paper will not be behind a paywall then and I am curious of those graphs :D
<bramc>
Basically if you access a distant part of memory you haven't hit in a while it will have to be swapped in, and everything will have to sit around waiting for the shipment to come in.
<bsm1175321>
I mean, once the set is computed, we're talking about adding/removing a few thousand elements, on average, no?
rusty2 has quit [Ping timeout: 250 seconds]
<bsm1175321>
I'm well aware, but even including that, I'm surprised.
<bramc>
Yes that's correct, it requires enough recomputation of the while utxo set to be dicey though. Just roots of those few thousand transactions are no big deal.
<bramc>
Also you want delays hitting block propagation to be really, really low. And naive merkle tree implementations are really, really bad.
<bramc>
My argument is roughly along the lines of what you're saying though - my thesis is that the performance problems can be gotten under control using a good implementation and don't need any fancy MMR stuff
<bsm1175321>
So, in some of the braids work I'm doing, I moved from tree-based sets to hash-based sets, which generally gives orders of magnitude improvement in speed. Is there a reason a hash-and-bucket set implementation wouldn't work here?
<bramc>
Yes, you need to compute a merkle root.
<bsm1175321>
Ooooh yeah.
<bsm1175321>
fart.
frankenmint has joined #bitcoin-wizards
frankenmint has quit [Remote host closed the connection]
<bsm1175321>
Couldn't the hash-and-bucket structure itself be encoded, instead of a merkle tree root?
frankenmint has joined #bitcoin-wizards
<bramc>
You generally want to avoid using a merkle set even when you have one and it's highly performant. A parallel regular set will always kick its ass in performance.
<bsm1175321>
I mean, to first order the txids are already hashes, so one could use the first few bytes of that as your first try.
<bramc>
My implementation uses balanced prearranged things with that sort of technique. It's orthogonal to whether the underlying merkle set is binary. There's no benefit to going non-binary or unbalanced.
<bsm1175321>
Then you just need to encode the buckets with multiple elements...
<bramc>
My documentation at the top of the file which explains the data format is fully fleshed out by the way, I can explain the details.
<bsm1175321>
So then am I correct that a c++ unordered_set<> would be superior, all we need is a way to compute a commitment of one?
<bramc>
Yes a regular hash set is clearly better. I think I can get performance within an order of magnitude though.
adnn has joined #bitcoin-wizards
<bsm1175321>
A concrete comparison would definitely be worthwhile, before accepting delayed commitments, for example.
<bsm1175321>
(FWIW, I have a use for these commitments -- and I'd like them to be timely!)
<bramc>
Yes everybody wants a performant merkle set implementation. It's a useful toy for all kinds of stuff.
<bramc>
I need to get my implementation finished. Nobody else is volunteering to do one, and I don't blame them.
<bsm1175321>
I've found unordered_set to be sufficient for my purposes... :-/
<bramc>
The really meaningful comparison is how quickly you can do a simple lookup in a regular set versus a lookup with a proof in a merkle set
<bramc>
Doing boolean membership lookups in a merkle set is just dumb.
<bsm1175321>
bramc: given the entire set in memory, stored as a regular set (with insertions/deletions from a new block), how much time is it to compute the Merkle root from scratch? You wouldn't have a cache miss problem there.
Davasny has quit [Remote host closed the connection]
<bsm1175321>
The Merkle roots of sub-trees could be cached and invalidated when their contents modified accessed, so you know which sub-trees need to be recomputed, where the size of a sub-tree is set to the CPU cache size...
<bsm1175321>
s/accessed//
nonaTure has quit [Ping timeout: 276 seconds]
nonaTure has joined #bitcoin-wizards
bramc has quit [Ping timeout: 250 seconds]
bramc has joined #bitcoin-wizards
adnn has quit [Remote host closed the connection]
<bramc>
In my implementation the merkle root is calculated lazily. If you add everything before asking for the root and then request the root it will mostly do a linear pass
<bramc>
It also clusters sub-trees together. That's what cache coherence is :-P
ThomasV has quit [Ping timeout: 244 seconds]
<bsm1175321>
bramc: You've got a few syntax errors in your master HEAD
<bramc>
bsm1175321: You mean like all the lines which are written in english and need to be translated into code?
<bramc>
When I said it isn't done yet, that was meant literally
<bramc>
That said, the explanation of the data format at the top should be final, as are all API methods (the ones which don't start with an underscore)
<bramc>
The _ref and _deref methods are ridiculous and should be eliminated completely on a port to C
murch has quit [Quit: Leaving.]
mkarrer has quit [Ping timeout: 276 seconds]
<bsm1175321>
No I mean indentation and unmatched parens
mkarrer has joined #bitcoin-wizards
frankenmint has quit [Remote host closed the connection]
<bramc>
Eh probably, it isn't at all finished. I'm focusing on getting it all written before doing any debugging.
<midnightmagic>
Oh, hey bramc nice to see you again.
Church_ has joined #bitcoin-wizards
priidu has quit [Ping timeout: 272 seconds]
optimator_ has joined #bitcoin-wizards
metric_ has joined #bitcoin-wizards
bramc has quit [Ping timeout: 250 seconds]
JHistone has quit [Ping timeout: 264 seconds]
nephyrin` has joined #bitcoin-wizards
qawap_ has joined #bitcoin-wizards
qawap_ has quit [Changing host]
qawap_ has joined #bitcoin-wizards
yorick__ has joined #bitcoin-wizards
Jaamg_ has joined #bitcoin-wizards
comboy_ has joined #bitcoin-wizards
Alanius_ has joined #bitcoin-wizards
gwollon has joined #bitcoin-wizards
espes___ has joined #bitcoin-wizards
Anduck_ has joined #bitcoin-wizards
buZz_ has joined #bitcoin-wizards
katu_ has joined #bitcoin-wizards
bertani has joined #bitcoin-wizards
ryan`c has joined #bitcoin-wizards
AEM- has joined #bitcoin-wizards
warren_ has joined #bitcoin-wizards
[Derek]_ has joined #bitcoin-wizards
liviud_ has joined #bitcoin-wizards
henrytill_ has joined #bitcoin-wizards
phantomcircuit_ has joined #bitcoin-wizards
K1773R_ has joined #bitcoin-wizards
frankenmint has joined #bitcoin-wizards
Guyver2 has quit [*.net *.split]
johntramp has quit [*.net *.split]
nephyrin has quit [*.net *.split]
henrytill has quit [*.net *.split]
liviud has quit [*.net *.split]
d9b4bef9 has quit [*.net *.split]
ryan-c has quit [*.net *.split]
espes__ has quit [*.net *.split]
Alanius has quit [*.net *.split]
qawap has quit [*.net *.split]
katu has quit [*.net *.split]
AaronvanW has quit [*.net *.split]
yorick has quit [*.net *.split]
mountaingoat has quit [*.net *.split]
face has quit [*.net *.split]
K1773R has quit [*.net *.split]
arubi has quit [*.net *.split]
Guest54896 has quit [*.net *.split]
optimator has quit [*.net *.split]
comboy has quit [*.net *.split]
epscy has quit [*.net *.split]
buZz has quit [*.net *.split]
catern has quit [*.net *.split]
jaromil has quit [*.net *.split]
phantomcircuit has quit [*.net *.split]
[Derek] has quit [*.net *.split]
aem has quit [*.net *.split]
metric has quit [*.net *.split]
bertani_ has quit [*.net *.split]
Anduck has quit [*.net *.split]
gwillen has quit [*.net *.split]
warren has quit [*.net *.split]
Jaamg has quit [*.net *.split]
so has quit [*.net *.split]
K1773R_ is now known as K1773R
ryan`c is now known as ryan-c
maaku has quit [Remote host closed the connection]
Guest1996 has joined #bitcoin-wizards
maaku has joined #bitcoin-wizards
catern has joined #bitcoin-wizards
bramc has joined #bitcoin-wizards
qpm has quit [Ping timeout: 260 seconds]
epscy has joined #bitcoin-wizards
face has joined #bitcoin-wizards
mountaingoat has joined #bitcoin-wizards
qpm has joined #bitcoin-wizards
johntramp has joined #bitcoin-wizards
buZz_ is now known as buZz
<bramc>
Hey midnightmagic
wumpus has quit [Ping timeout: 264 seconds]
AaronvanW has joined #bitcoin-wizards
AaronvanW has quit [Changing host]
AaronvanW has joined #bitcoin-wizards
Church_ has left #bitcoin-wizards ["WeeChat 1.5"]
go1111111 has quit [Ping timeout: 244 seconds]
arubi has joined #bitcoin-wizards
wumpus has joined #bitcoin-wizards
frankenmint has quit [Remote host closed the connection]
heretolearn has quit [Quit: Connection closed for inactivity]