All Core Devs - Execution #54

Transcript

[Music] [Music] [Music]

[Music] [Music] [Music]

[Music] [Music] [Music] you

hey everybody let's get started on all this we're gonna go ahead and get to the first item on the agenda which is some of the roadmap discussion the first one is the robson fork for Constantinople before we started Martin said it was about 3,000 blocks away which should be 11 and a half hours I think there was a release yesterday forget or a little over 24 hours ago forget and then parity had their release how long ago offeree a few days ago now pretty much commuted but they released

just a couple of hours up yeah okay cool so the Robson Fork should be going on later let's see I think I'm getting feedback somewhere oh I think Greg could Jamie thank you okay well I'll go ahead offeree if you had anything sorry did you release yeah I asked about basically how long ago parity released and Martin said like yesterday yes right now

wonderful so do we need to do any kind of prep for tonight as far as or today for some of you for the Robson Fork I'll be online I already told the Qataris that we have a fork very soon so they are aware so yeah they we also have spreading the word in case we need to - to upgrade and okay sounds good and I think yeah most of the miners I mean if there is a if there's another incident like last time with the rob stand fork we'll

just learn more lessons from that see what happens when we have a two forks at the same time oh no actually for the Robson Fork we're just doing one fork aren't we yeah that's right well I would assume that there will be a split and there will be some fallout it doesn't bother me that much but it might be good to be prepared that it might happen because I would guess that a lot of people of mine maybe they're not that active I'm not sure they about dated do we know was mining on Robson do we know

anyone that's kind of part of the problem we I don't know anyone who's mining on Rob's tannot the mining pools who are mining on main net yeah well I think Jamie and Foundation has been a minor that's I'm not sure we're running that these are the guys who made them I don't know if they're using miners I might blast out a tweet today from the etherium account that just says if hear

it on Rob Stan update your an update your node okay so we'll go ahead and do that is there any other comments on that okay next the Istanbul hard fork roadmap I think we talked about that last time and it just stayed on the agenda so we might not need to go into too much detail on this but offeree do weed is there anything since the last time we talked that we should discuss about it no I don't think I mean a lot of things are happening right now Istanbul is like low priority for me right now I yeah that's what I was expecting

sure and then the next item is the proof of stake finality gadget on the proof-of-work chain for serenity that was on there last time too so I kept it on I don't know if Danny or Vitalik are available because they're at the Stanford conference and they'd be the best people to answer that so we might just need to keep that on for next time unless someone wants to speak on that okay next we're going to talk about the prague pal audit there is a group forming or multiple groups forming to perform a third party audit on prague pal with the intent to get an audit done before we stop accepting a IPS for

Istanbul the reason for that is to test the viability of the claims that were made about Prague pal such as the fact that it is that it is even across all types of forks or sorry all types of like cards like AMD and NVIDIA and a six and FPGAs that it has the same like hash rate and performance and also that there's not going to be any problems once we launch it on main net so like testing things about it to make sure

that it like works correctly yeah I'm not like in tuned technically with what all is gonna be tested quite yet but we're working on that and we're also looking for funding for that so if anyone's listening on the call and interested in funding a prague pal audit will probably be calling for that soon and you can reach out to me same thing with any organisation on the call so yeah that's pretty much all there is about that but I'll go ahead and Martin yeah questions about that so that sounds to me like a pretty diverse go and I'm wondering will the auditors

be hardware experts and experts in designing a success yes that's what we're going for and Martin I'm gonna get you more looped in this was a really new thing but I'll get you directly looped in on this since you've been following this pretty closely because yeah it's it's pretty new so we don't have all of the stuff exactly down yet but we should have that really soon yeah any other comments any update on that they're gonna be clarifying some of the specification part because I tried implementing it a few weeks ago and

halfway through the spec it turns into implement this code block there's like no narrative no unit tests and it seems very heavily weighted towards please these are patches which is not something I'm wild about so any update on any of that from them or I haven't heard anything personally you might want to visit their get her channel which I think in the all core devs chat was a link to their Gator channel was posted very recently yeah I was there two weeks ago I listed my concerns and issues and okay hmm I will make a note to ask them about that so because they're so for some things like in the spec they have

some undefined constants like dag bytes I figured that that's the size of the bytes of the dag but everything else is presumed in word sizes 32-bit words there's also some readability concerns I have with the weight specified makes perfect sense to people fluency but things like JavaScript and Python aren't gonna be able to you know easily employed the subtleties and get you know like any reference types of of numerics which is hard to read reference types of structures are you know everyone gets that instinctively but there's just you know some readability of amenability things that I think should be improved if we're gonna call this a spec that's gonna be implemented because right now it just feels like that here's our patches go for and parity again which this is not something

I'd rather see some generic specification it's easily nineteen's necessary easily implementable some of its skilled in the art can implement without too much difficulty ok yeah that's that's definitely a legitimate concern so I will let them know to join the Gator chat and address specifically some of the things you've been saying in there is your name dan affarin on and the Gator chat or is it like what's the nickname sh e what was that again Oh got it

ok great thank you alright any other comments on prog pal does anyone have any further ideas on how we will even decide to implement this from AB or what group should decide to implement this because I know that was kind of a sticking point is to just implement it and basically let the world decide do we want it or not that could be done if we implement it in the clients I mean merge it into the mainline lines and decide on

like some block number and we could beforehand we could decide that well if the some random thing Bitcoin norms on a particular number if it is even then the clients will default to we will put the default option to enable it if it is odd we will put in it not by default not enabled and otherwise users will be expected to set their preference go with it they'll go without it that would be one way to do it then we don't really make a decision we just lift up for the

world one drawback that is that there will be less hashes per second with Prague cow so if we compare the I mean the the difficulty is expected to go down on profile so you're breaking up yeah Martin I think I heard all that offeree what was what were you saying you were breaking up oh you're

muted aw free it's just a roundoff what I was saying so even if we say there's double the amount of hash power on the prog power fork only then will they appear to have the same diff will they reach the same difficulties yeah that doesn't sound good this is something that they maybe eat cat burgers should have a discussion in the community because I'm concerned if we do

stuff like block voting that we risk splitting the chain unnecessarily I think you know the cat herders would have a better pulse on the needs of the users of the needs of the miners rather they relied on the animal system like my cache numbers yeah that's a good idea I can bring that idea up to the cat herders and we have off rien Lane in here to that are part of them so we will we'll bring that up and see if there's a way we can find out how to equitably make this decision that'll be good yeah well I mean we've been we've been trying to come to a decision for several months and we can't just run it off to some

other group all the time I hope that they will make decision for us I'm wondering really how will this be resolved it's resolved by us making a decision thank you thanks Greg yeah that's probably true do we need more data to make a decision or do you think we can just make it we can just make it it's our job I agree I just I'm glad that that some folks have strong opinions about it I just feel that a

number of core devs have kind of said the opposite right that they don't feel like qualified to make this decision so and we have like a maybe we can look at our menu of options which is pretty short right one is sort of decide not to decide a second is decide one way or the other a third is I think it was Martin who suggested implement both and maybe have some randomness that would you know Institute a default and a fourth is maybe some other course of action like you know auditing or bringing in some experts or something I guys sorry I just want to be clear like that was sort of our menu of options we missing like huge options here yeah I think you're conflating a couple them so basically what I said it's the same

thing as I mean if we decide to go with bow and we put it out there to the world it's still up to the world to decide no well that's assuming yeah okay you're right even even if the core devs were to decide in favor or against then you're right people have the option um not for sure but then maybe it would need to be in its own hard fork right so there's no other problem yeah I think that would be prudent to make it its own hard work yeah I think the one challenge or risk there is just somehow okay so all right I'm gonna shut up never mind

and the last one change said to leave it to some expert group that just sounds like we're trying to defer the decision even more yeah I this is Piper I honestly don't know what to do here I've largely stayed out of the discussion because my expertise in the area is not adequate to make a decision and making an informed decision from other people's research is also hard because it's hard to know who to listen to which informations right things like that and my claim said it's a decision that's frankly I'm not excited about

making for the network so I think personally that having a third party as neutral as it can be audit of hardware experts and others who will assess some of the claims of prog pal if we can get to a point where we can say this will work and here's why it will work if we put it in I think that would help a lot if if we can write an algorithm right to put this in computer science terms right it's a bit like let's say we decide to do this audit and the output is X right so it turns out that it is or is not you know compatible across different hardware whatever the what might be if before that process we can commit to a course of action given

the output of that process I might support that does that make sense yeah it does okay perfect sense yeah that does make sense another another way of putting it is light I mean to put it in layman's terms right what is the purpose of the audit I just want to be very explicit about that let me pull up what we've written so far in notes on it and if you give me a second I can do that so I think the only questions that the passes what's our decision because if the other fails it's an obvious no no I'm saying we have to we have to pre commit right we're saying if we do the audit right and we you know barring some completely unforeseen circumstance and

we you know there's some criteria that the audit would be used to prove and if it proves true then we've already committed that's my suggestion so one thing about the audit we met you mentioned whether it performs differently on different types of hardware and that in my it sounds very different from the investigation whether it's difficult to make specialized circuit for it and it sounds like a typical mining firm and in mining operators current GPU owners who

actually have a sorted set of GPUs like they would test these things and determine does this work for my type of GPU and the pools should really kind of already have this data based on the test net which they've been experimenting with am I being naive I mean is it is it reasonable to think that some auditors would purchase ten different types of a you know a million Nvidia cards and try out the hash rates on each and every one guess that depends on how much we pay

them yeah I'm just trying to clarify what the audit is what the focus although out of this if it's no I I'm with you I think well said I'm worried that we're still just asking for more information when really we're in a position where we this isn't a decision that it's it's a decision that I don't want to make I

can't speak for everybody else but the decision I am NOT excited about making but I also acknowledge that it's potentially our responsibility to make it and that there's not really anybody else to make it so I'm not sure what to do here I I think that we still haven't explored all the possibilities about getting community feedback before we're forced to make a decision including stuff like this audit which I consider a piece of community feedback so I think that we don't need to make this decision the the latest we could make a decision

on this would be or I guess I'll put it this way is it is it possible we would have a prog Pao fork before Istanbul because for right now I haven't heard anyone say would happen before is stand bull accepting like Martin saying we might have a small one I think it could happen anytime that was isolated right if it's just this one change and it's thoroughly tested so I think it could happen before as well and I think by thoroughly tested it needs to be like danos concerns need to be met it needs to be implementation set there needs to be I mean I was looking at some stuff on the prog pal minor on the

gangnam tests net branch the other day and there were people saying it was making AMD cards fail even yesterday so like it doesn't it just feels like it's not ready which is why which is why it gives me pause to say that we could you know fairly quickly go in and make a hard fork for this thing I think he needs to undergo the same rigor as the eath one expert bones we've been talking about let me let me copy and paste some of the notes we put the other day from the meeting I was in and the reason I don't want to give away too much about the

meeting is so that it stays neutral until we get enough stakeholders in there and make sure we secure funding so it's not we're not calling anyone out but I can I should be able to post some of this stuff right now and the zoom chat oh it didn't format it great but it's basically question after question and that's only that's only part of it there's other things we're also deciding there's like non-technical due diligence as a piece of this so given the current Hardware how would this affect GPUs participating as minors is there an alternative approach what's the timeline for switching what's the

expected impact on the current miner ecosystem is what they replicated is what they replicated what the specification says does Pro groundbreaker fishin' see an ASIC chips etcetera so it's a pretty thorough it's a pretty thorough audit spec that we're expecting one to two groups to complete and it's one of those things where there's probably not a group that's perfect to do this but there are groups that can come up with most of this and that would be better than nothing or having two groups do it so that it completely takes all of it we'd need at least two minimum yeah oh it does look

truncated I think there's like a limit for how much we can paste at a time and zoom but yeah you all get the idea from what I've bought it what I posted so far I think is anyone very opposed to the idea of an audit or doing more homework assuming that we have a specific set of questions like this I support the idea of audits and I support the idea of seeing your clients can implement it at they have time that will give us I think more data points that we need to proceed or not receive calming down a little I'm happy to get more information I just want some

decision that says this is the information willing to get and then we're going to decide then I'm very opposed to the notion of we can't decide so we'll go to the effort to let the hash power decide I don't see risking a split in the network just because we can't make our mind up in the u.s. system at least you take a jury of ordinary people they listen to the experts and they decide yeah you know yeah but we haven't heard the experts yet is the problem until a lot it if we're gonna have some meeting down the road to actually make this decision this

might not be a popular opinion but I think that should happen behind closed doors and then we can open up exactly what was discussed after the meeting because that's one of the problems with trying to make these types of decisions something our core devs call is that we're all on stage and it's not entirely a safe place to try out ideas and try out opinions and this is a contentious topic contentious jury trials happen in public you can have a jury is deliberate in private right we can I'd support a smaller group to you know do the work for us of evaluating the audit and you

know helping us understand it but we need to make the final decision here in public so the people know the basis on which they're making we're making the decision quite clearly well take a very important voice we're missing as the voice of the miners the people who are miners we're talking about it where only those who are rapidly in favor of prog pal so I think we need to have a meeting and these miners will probably want to be behind closed doors we ask people mining on aetherium especially the ones who have a voiced an opinion on prog pal I think they have a very important stake that we need to discuss if we can't do that that's the way they run their business hmm I've read

random ideas last question I want to throw out this may be impossible but as I'm sure you guys all know like Bitcoin does things with minor signalling right where they have miners that particular signal bits on the blocks that they mined I suspect one reason we haven't done this historically with etherium is that miners play a different role in our ecosystem but given that we're looking for input specifically from miners here when people consider doing that in this case like to see which they support so that what sexers suggested by I think was Peter from the play pool Peter Fechter and the thing is so if most miners are in favor of power then it shows they're in favor of if only a few miners vote in

favor pronto then they can say oh look it's already taken over by ethics so does that really I mean additionally well I already know the result of what we do miner signalling it would be if we did miner signalling it would be favored in favour of prog pal because Peters pool itself takes up thirty to forty percent of the network and I he already said he'd signal prog pal okay all right so turning the question around assuming you guys are right if if we're not looking for feedback from miners then who are we

looking for feedback from I mean we're just sort of saying that community loosely defined let's not get stuck on this topic for an hour because it's complicated yeah um that too is something that is gonna be maybe like so a lot of the reason people are against Prague Pal has been conspiracy theorists kinda ish from my perspective so to cut through that would require having the audit come out with results and if that quells everyone's concerns then to me the community looks like they're in favor of prague pal that's my opinion yeah I don't think we're just gonna have to live with that oh yeah no

it's not that it's gonna turn out to not be controversial anymore it'll just the loudest voices are the ones either from ASIC manufacturers which we can just say you're biased generally or from people who say that Christie's a bad actor if we're being blunt so just speaking for myself the audit would give me more confidence to like make a decision here it's sort of a symmetric in the sense that like if like if something turns out to be really wrong with cog cow then that would be as someone said earlier like pretty decisive but if there are no red flags then we're probably still in the same situation we're in right now which is to say it's still controversial yeah okay that sounds like a good plan and that

sounds like something that's pretty important so hopefully people will step up to get that done more quickly and I'll talk to a few people I've been already been talking to a few groups who have volunteered to help so I think red yeah go ahead Lane sorry to interrupt I think that well I don't think that the cat herders are like magic and could solve this on their own I do think that sort of facilitating the audit is definitely something that group could handle yeah I agree okay so we'll try to get to that as quickly as possible now the audit may take months so I guess the biggest thing to decide here is is

that an is that gonna be too too long of the time for the for the threat of a 6/10 realization etc claim should go ahead of implementation while the audit is going on and hold off on committing until the results if that's a concern yeah this is a big enough deal that I don't think it's a waste of time to keep implementing while the audits going on because if it happens in parallel then we can say okay everything is implemented as the audits ending we didn't find anything and if along the way there's a major red flag that's

found in the audit we can just ask the auditors to immediately let us know so we can stop implementation did you have a comment Lane I was just trying to respond to your question about you know would that be too late and unfortunately I think we just don't have data on that it's really just speculation at this point yeah okay and there's no way we can really get data on that from what I can tell so we've decided that the audit is something that we all want right yes I'm on board okay sounds good so we're definitely gonna go ahead with that and then from that audit let's continue to implement prog Pao and

make sure that we have if def else give very clear specifications and stuff like that so that we can continue to implement it with the idea that if there are no red flags found and the cat herders come back very soon to say this is probably what the community would want then with that combination that should give us enough confidence to have the decision kind of be made for us almost should we ask for miner signalling yes or no I'd say no I don't think it wouldn't matter but would it hurt I think it could hurt it just is

one more like conspiracy theory signal where we don't really get a strong we don't get strong information out of it and so the information that comes out just something that's easy to spin and hard to actually build from okay does that make sense like from what I have thought through no matter what comes out of it it doesn't actually give us a strong signal that were like one way or the other but it does provide fuel for anybody who wants to manufacture controversy to build off of well I I guess if it shows that there's very strong support it does tell us something if it doesn't show

that we don't really know that much but but that's my assertion is that we can't actually see that it shows strong support like we've got a pool that has a very large majority we've got like there's all kinds of ways in theory that that it could be gamed like it's and in either direction it goes there's a story that you can tell about why it went that direction and why it isn't good information because saying we don't want to being afraid of if we have signals that we can work with them if we don't

have them saying we should have as many singles as possible and if we had - signaling that would be one we could use why not so even if it causes more confusion yeah changing my mind - I'm gonna retract my my statement there yeah me too it's better to have more signals yeah okay how fast would it take for us to do bit bit what's it called signaling bit signaling wouldn't even require any form modification they can just set the

extra data to like support or don't support and if they do either one of these they are signaling their votes if they leave it to whatever it is that we don't count it's something like that to make people running scripts easier should we come up with like a three letter for letter like support or not support kind of thing is that presenting does that make sense so people who are calculating this they don't have to like read through every one manually say that again Martin yeah it's just one signal

per block it's pretty what I meant is when you're collecting the data if someone says I support it exclamation point versus support if you're running a script to try to calculate all that is that is that something that would be happening is someone running through and trying to count the number of times they say the word support or should we have like an acronym that is meaning explicitly support or not support I think that it'll be an exact 32 bytes extra data value for both and and people doing the mining can set it to exactly

that value or not okay so it should be not no support versus support right well some of some of these might be use for different purposes so but we'll break that if we require all of all of the bytes to be used for it so I prefer something much shorter yeah we don't I'm sure we can yeah yeah and then we'll spread the word let's do that okay so just to be super explicit here this

does not require a hard fork right but does it require client updates no they use a bit better but it doesn't require them okay great cool alright so we're gonna be going forward with the audit or consider if we get the funding which I'm pretty confident we'll be able to as long as some people step up or some organizations step up and then secondarily we're going to have minor signaling and third we're gonna have the cat herders do some investigation or come up with some ideas that people haven't come up with before because we've kind of been stuck in an echo chamber of all this stuff in a little

bit hesitant to make decisions and then finally let's do a timeline for when we should decide stuff I think if we don't we're gonna go with this forever I don't think we can do a timeline until we know when the audit is coming back oh that's a good point never mind then okay we can move on in the meeting then that only took half an hour yay three minutes because I'm not sure I heard that so we don't have any notion when the audit will be done not at the second because it's like I literally had a meeting about it the other day we have even a guess yes April is what we talked about but that was

because they asked me to put it they asked me to put a date on there and I said well if it's not going in still Istanbul April but we haven't told like explicitly any of the people who are doing the audit that it has to be April or anything like that or that it it was still kind of a little up in the air okay I mean obviously I really want there to be a date on which we say the decision will be made yeah but we need to figure out the audit time first and I'm thinking March or early April is something that we can shoot for realistically Greg would it make you happy to say that we will make a decision within one month of the return

of the audit one month or with what basically a core dev meeting from within the turret return of the audit yeah I'll give it two weeks sorry the one month was arbitrary just essentially using like we are also very tired of this you know I would happily decide today but I'm not the expert I'm happy to wait for the audit but I'm not happy to make this decision in May or something you know awful lot of people would like to know so they can get on

with their business okay sounds good I gonna spend thousands of dollars to buy a rack or CPUs today or not yep yep okay let's continue on with the agenda so the etherium 1x stanford meeting it went really well there was about 30 people it wasn't just core devs it was also virtual machine experts it was DAP developers from a few major daps there was people who were like unaffiliated core developers and just experts in aetherium in general so there were some good conversations let me just

pull up my notes real quick so the objectives for this were to discuss the longevity of aetherium and decide on early and incremental changes that could be proposed in order to extend the longevity of aetherium and make sure it doesn't break down or anything so that Hudson means the eath 1.0 network absolutely the eath 1.0 network so let me go ahead and just copy/paste what the objectives were so I don't have to read them all out loud

and again these objectives were something that were brought up by Aleksey they were not something that was explicitly agreed to from everyone in the room at all and we didn't even really get to a conclusion on number four so there hasn't been a social contract made specifically that I'm aware of but that was an idea that Alexi had so the general thing was is the etherium network we need data so that the etherium network we know is not going to crash and burn and about a year or two or three or however long it's gonna take we're gonna figure out you know if the if the state is you know this big or has this many accounts or whatever if it's

if it's going to make the network unusable if we don't do anything about it and we have this unbounded state we talked about ways we could combat that the he wasn't team that was there talked about their improvements and what they could be doing and then we had simulation and testing team which that can go over in a second talk about what they're gonna do so we had the format was like five to six short presentations in the morning and also in the afternoon followed by a breakout session followed by individual work from everybody it was very laid-back it was very ad hoc there was a lot of good discussion

happening and I think that it was there was some really good takeaways all the videos for from the presentations or online I'd say there's about let's see maybe 10 hours of video from the presentations and so if you want to look in on some of that it's it's interesting stuff so let's go over some of the working groups Alexi isn't here to talk about state fees which by the way we renamed it from state rent to state fees because it involves stuff like lock ups now so it's more of a framework of ideas then specifically rent so we're calling it state fees now when we're referring

to Alexie's work and then we have the e wasum group the testing simulation group in the chain pruning group and we don't have Peter or Fred here today to do trained pruning either so the only working group we can talk about is like the E was a working group and the testing the simulation group that Shaw Hannah and Zak came up with so let's go ahead and start with awasum was anyone who was at the event from the ewaz i'm group in here I was there on the first day only yes sorry diem I think you were there right yeah I can hear you sure so the during what what is your what is

your takeaways and your like perspective from the II was um working group that happened at the etherium 1x event or your general impressions from the etherium 1x event yeah well there's interest in in getting a lot of people call so we we have to just think how many compossible would really want to add people expect that they will be able to add their own plate composite like vertically so yeah that's that's not

going to happen oh actually it's not nothing immediately and yeah like we had a I mean I remember an interaction from Frederick that said yeah we we have all those of t minus three compiles for example she recovered that is called very often actually sorry that was Alexander said that and and we we just we're probably going to keep the pre compels that accurately there in and in gas in parity as as they are we just want to address them as a because as a way to easily extend and add some freaking piles but we're not going to be crazy and accept every single

pre-compiled just because of some rent or persons from the internet needs it yes roughly yeah what I what I remember from them for three days cool thank you and then Layne do you have any other comments just from the kind of one-day interaction not on you awesome specifically I think you uncovered that quite well the day I was there the first day I think we spent more time talking about Alexia's proposal than anything else so I just wait for that I guess okay yeah we weren't gonna really talk much about that today because Alexei is not here and I'd rather get it from him

then yeah so just a just a quick general update um I've been I've asked everyone who took notes during the 1x meetings to share them and I've collected I think well definitely collected notes from several people I have my own notes as well from the first day so I know that the video streams have been published I'm gonna publish all those notes publicly as well if if anyone has notes that haven't been shared that would be amazing if you could share them with me and I can add them and for the folks who already gave me their notes if for whatever reason you don't want them to share publicly or if you want them unattributed or something then let me know but I hope to do that in the next few days okay great and then we have Zach from white block who was in

attendance and Zach if you want to go over your simulation working group and kind of your thoughts on the event and what you guys are planning to do and as a result of it yeah so Vanessa presented their work on simulations I think that their that their group was kind of working primarily on focusing on on Uncle Ray and I like meant certain environmental conditions that arise and may cause uncle ray they are had aggregated data from the main net through various sources and ran some simulations and the K framework so that

was a pretty pretty cool thing to see and then on our side we were kind of working with Alexia on establishing a test strategy a test series or a test plan at least rather on validating the hypothesis and observing different environmental factors and the effect on sink failure in relation to the increasing state size so for the past week I mean pretty much since I got back from Stanford we've been attempting to actually generate this large state size within our within the white block

framework so I think we agreed I was at first I was kind of unclear on what the definition for state size actually is what what how do we define that specifically and what we said was well let's just start out with a count let's try to generate 240 200 to 240 million accounts within the network on on white block so it's in a controlled environment so we kind of that's kind of I mean we've generated thousands before but never millions so we spent some time trying to figure out the most efficient way the optimal way to actually generate

those accounts within the network and not just actually generating you know public/private key pairs or private keys or rather so I tried a few different methods the first was generating a genesis file and then just you know bootstrapping off that genesis file that proved to not work and i didn't think it would work anyway but yeah so if you do actually generate that if you can actually generate a genesis file 240 million accounts in it guess we'll cry if you try to try to point to that Genesis block so what we ended up doing was essentially kind of like an

incremental counter that's just generating random strings for account addresses and just sending transactions to those accounts so we started the new network with a difficulty at 0 and just multi-threaded it so in wrote the script and C++ so now we're about halfway there and everything seems to be working and my math says that we'll have that state generated by Wednesday so the idea is that will generate the state save that state as an image so that way we can run various tests repeatedly within our within the emulation platform and

that'll allow us to like test various environmental factors conditions one of the main one of the main hypotheses was that there's a correlation between bandwidth and sync failure rate so what we think is that essentially there's a race condition because as the as a new node is bootstrapping into the network and syncing it's never going to it's essentially never going to catch up with the actual state because as the state size grows pruning becomes much more aggressive and as that pruning becomes aggressive they're going to lose there's going to be like holes in what they are

able to in the state as they're sinking so I kind of have been documenting the entire process and I'll share that later but where the plan is to start actually testing this next week and I'm my team and we're gonna be working with Alexi and Andre and kind of validating those hypotheses and taking it from there so we're gonna just start off simple and observed and with but I have a I have a suspicion that it's not just bandwidth that's it would just that's um that's a simple simple answer I think that there's probably multiple conditions but will isolate each one of them and observe the effect so yeah so that's where where we're at

but yeah I got a lot of the event it was really really good I mean I was really tired it was kind of like being in college again but I mean we were in college and everyone was talking for a really long time it was nice cool thanks Zach Brooke do you have any feedback or anything from the conference you wanted to talk about I mean nothing in particular I got some good feedback on you know the state of awasum and you know great to connect with the team there and have some proposals for improving the EVM and a transition for

IBM to Eve awesome out of that but I'll be typing that up over the next few days great and then finally I think the only other person who was there was Anton did you all go over anything oh you're muted oh I could hear you now oh no I was talking to and sorry Anton I was talking to Antone or oh cool yeah go right ahead and tone no you should be

able to share that state publicly that you majored the what publicly once you have oh I think you just asked cans AK share the state after he's loaded up the accounts right it's hard to hear you so I think that's the question that was asked they're like good sharing the state yeah is it like a proprietary state file kind of a thing or no it's just it's entirely test related but I don't know what what would

be the I mean yeah we could share it I mean it's probably going to be really big and prohibitively Oh somebody sent me a message how did you generate a chain I didn't take time from the call but I was just curious about that yeah when I say zero I mean as low as as low as possible yeah so what you're saying here I actually had a lot of had some issues figuring that out I didn't really know that right off the bat yeah so like I was getting some errors to guess but I've like never seen before so a pretty interesting experience like block lost okay I've never I've never seen that

I've been working with but yeah we were getting some areas I mean this has been an iterative process so I mean there were a few different approaches that we took like initially we had several nodes and then I found that it was actually more optimal to have less nodes and then after we reached a certain point it was actually more optimal to have two nodes because one node can't be sending these transactions we're flooding it essentially so it's it's harder for it to actually send those transactions once the difficulty reaches a certain level while simultaneously mining but still at this point as of yesterday I I printed

out the data or the stats and the block time the block time was at 13 seconds so it's getting a little more difficult so I'm anticipating that as time goes on it's going to take longer to do this here I'll share this link in the chat you can check out the stats if you want and I also post it a lot in the this Twitter thread so it's kind of like venom pretty so so what I noticed is that the process of actually generating in a state is kind of also testing the hypothesis as well because now I need to have another node sync on the network so I should be observant of this and taking

notes and you know like documenting the process and the outcome and all of that so yeah great wait am I on me nope I'm not does anyone else have any other comments on the 1x stuff or any of the working groups okay let's see next our testing and client updates so for testing updates do we have Demetri or Martin in here Martin's here yeah so Demetri generated the new test very quickly and yeah they work

fine I guess now and then why later I pushed it to master and they also think to work fine in high disparity and guess is the all the test if expecting just to pass so all looks good one thing though is that the fussing we can't do fussing because parity has there's a feature missing imperative for the new status yeah with the test harness for parity so unfortunately we

can't do fussing on Constantinople fix right now okay now we can go through the clients let's start with parody yeah we released released us today with 2.8 stable and as far as I know and that's I guess okay thank you

yes yes we made a pact release on 1.8 22 for the concert no to fix it also contains some security fixes in the peer-to-peer protocol stack meanwhile we're continuing to merge changes on to the master branch which will eventually make it into 1.9 release which will be potentially breaking changes but that's not yet for for a while Cheers okay alas yeah the important information is just we merged the Constantinople

exchanges so it's the master branch is where it should be sorry I had to yawn Trinity there is a work in progress to get the st. Petersburg fix in probably get a release out with that next week and Tuesday Oh was I needed that whole time

yeah I was mated that whole time wasn't I anyways harmony didn't give an update we have released Constantinople fix likewise the others and that's it great can anyone speak for aetherium Jas okay then we have Pantheon yeah we shift topics and we just did a release where we reworked all the CLI and one of the options we added was support for the new corley test that's so not the old one

the one that launched yesterday article started work on fasting even though that's an older standard the new 1x standard isn't ready yet and whenever that happens awesome turbo get Alexie's not here do we have anyone from Nimbus or anyone from mana ok and I've been meaning to do this but I haven't gotten around to it I'm gonna invite the mantas people here that's the IO HK aetherium client written in Haskell I believe and I'm gonna start inviting them to the

meetings since they are able to sync to mein net ok is there anyone who can do research updates I don't think there are because it's usually Danny or Vitalik okay offeree congratulations on Gourley con I heard it was amazing I think I think a phrase enroute to fall stem and he may have like spotty internet access or something but I can attest to the fact that it was awesome great thank you IDs boy awesome I wanted to say that I

won the battle with the board of directors so we're gonna be open sourcing white Bois our framework for testing and that allows us to do things like you know act like the state is large or we can simulate or emulate network conditions or like p2p networks and it's pretty it's been pretty useful for one direct stuff I think it's also applicable to 2.0 I was speaking with maker L yesterday and we were talking about like helping out with their client initiatives because they said something about the emulation stuff yesterday on the call for two dotto and we could

actually emulate like DTP networks so if you don't have a stack ready you could actually test so I'd encourage anyone to to get out to talk to me if you if you want to collaborate or work on that so we're also incorporating like stuff like like a kind of presenting like this entire framework for development and design and implementation that includes like formal verification kind of like Petri net and like state-transition modeling and then the idea is like you come up with a mathematical proof formal verification for your system then you develop then you run emulations to generate large data sets that are going

to be indicative of real-world performance and then based on those data sets you run simulations so you can test and validate performance before you you know ship anything so I hope people are excited about this cuz it's cool so in fact congrats Congrats in the decision to open source I think that we believe thanks and if hopefully the community can return the favor by contributing so let us know how we can do that just use it and talk to me and let me build better tools for you because I think that these processes will make everything much faster and more efficient and accurate and like you know deterministic so I

mean just collaborate and tell me what you guys need and let us help and build the tools and so we're gonna announce I'm trying to like do some sort of like marketing push around it so we're not just like dumping a codebase I want to like we need to write the documentation and and do like clean some stuff up I'm rewriting some modules and rust so any of you rust nerds should just hit me up because I'm actually rewriting some modules in rust because they needed to be rewritten and I'm also I'd also want to write thrust so and yeah that's it just you know hit me up so we could emulate all that traffic you can

automate transactions you can generate accounts you can deploy your client each node exists within its own private network so it has its own IP address so they're logically separated from one another so it actually recreates a real world environment so you so when you test within this framework it's actually going to it's highly accurate right so it'll look here as if I mean what you see within our testing environment is going to be very similar to what you would see in the real world so anyone else questions we don't need to talk about this here just feel free to hit me up and I will oh yeah by the way so right now we have a SAS product it's in beta it's huge but can you stop shooting

your products here [Laughter] that kind of thing I'm sorry okay yeah I understand husband we had a couple updates you see the last text message here about yes yes actually truffle ganache is here it's a work in progress with the Petersburg release and then ganache core has an experimental Petersburg release published to NPM David are you still here oh okay cool any elaboration on that or is it just an update just an

update we we just had a fork of the MJS p.m. that we kind of packed together last week so earlier this week changes are still in progress thanks so much for doing that that's great that'll help with some testing stuff I'm sure okay and thanks white block for open sourcing their stuff that's gonna be super helpful and thanks everybody for coming is there any other announcements I think we're at the end of our agenda oh I think we didn't give an e was an update Alex oh yeah Alex was here looks like he

dropped yeah I mean I guess just very briefly like we've been pretty focused on the etherium 1x stuff so I don't think we have a ton to report the other thing is um benchmarking there's been some really interesting work on benchmarking looking at life which is a very interesting compiler I think it's an ahead of time compiler written in go just as kind of candidates for you know I guess for for for speeding up execution looking at different options considering the question of trying to do just pre compiles in 1x versus like full sort of user deployed contract support cool all right yeah I also have two more announcements by the way so the more

advanced Lib cluster based buzzer has it's been is now being worked on again by a guy from a good point grants crypto mental this is get up handle and also there is another person who has started working on more focus post testing of solidity and it is now on the Google open source fastest testing framework so that's been put into a lot lot better into production than it was earlier that's it okay anybody else just questions that remark about apparently prior to fall stem

there's need to be a meeting that I will not be able to attend about w3m I think it's called a replacement to whisper I wanted to ask your everything you or anybody from ferry if there would be some videos or any notes tickets that meeting that would like to have a look at it I didn't know that this meeting is happening and I'm one hour away from Blizzard so I can't yeah sure if you yeah yeah so thanks cool anybody else thanks everybody see you all in two

weeks I think I'll be there if not someone will take my place bye everyone have a great day or night well thanks [Music] [Music]

[Music] [Music]