All Core Devs - Execution #23

Transcript

try this now [Music] [Music] [Music] are we waiting for anyone else yes there's a couple more people for semester I'm hoping someone third are these bubble people or friends from

credit or us or our friends from intimidation or so one of the most friendly [Music] this is the other person [Music] [Music]

[Music] [Music] [Music]

[Music] [Music] [Music]

[Music] [Music] [Music] okay we can get started welcome everybody let's see oh I didn't update the agenda oh well we'll start out on the agenda and then I'll update it so this is the core dev meeting number 23 let me just make sure everybody can hear

us I'm not sure the thing working though yeah the YouTube stream changed it's now here if anyone wants to go and see the troll box and I tried to update everybody okay great they can hear us so let's go ahead and let's actually start with the note that Peter wanted to add and go down to that so okay so Peter you actually are fine with 684 it was just a bug yeah yeah so clarification along with notification I'm fine perfect okay and so if you could just go ahead with the first item we'll make it

the json-rpc updates needed by the receipt status change EIP oh yes so it's really just a minor thing so basically we changed the receipt format so that instead after the root stage we throw that out in the intermediate route and we'll paste it with the status byte and the only updated I think we should agree on is how this affects the json-rpc itself since it's part of the JSON specification and it's for my point it doesn't matter my suggestion would be either to add a field flag or succeeded flag which is a boolean and the focus on what it wouldn't return reroute rather

it would return this so I guess the question would be is everybody okay with that is that pretty much it Peter or yeah so the only question is that since it's a standardized API interface it would be nice to have it work the same way across all lines yeah and there's a link in chat by the way for people wanting to review that so the relevant groups that

would need to respond here I'm gonna guess our aetherium J and parity is that right Peter yeah more or less so it's actually consensus related issues just lyin to user they will be fine with whatever maybe someone could add it to the entry as well so I'm not sure if that something like this would be an EIP necessarily since it was a I guess this is just everybody agreeing on a JSON RPC standard so maybe that would be an e rc so yeah Peter maybe we should make this into an EIP just to say here's what

clients should do can't we just edit this and those do the EIP that's not good and yes that's a good idea yeah that's my suggestion oh sorry I thought you meant make a whole new IP sorry Arkadiy yeah let's add it to the current IP okay then I'll add the column here and in fact if it's finite everybody then we can ask Nick or whoever owns the IP to edit that should work fine all right let's see all right everyone has confirmed that they can hear us in the chat so that's great the next item on the agenda real quick before we get to the next item well you

see real quick if everyone can refresh the page I've updated the agenda just a little bit oh well it must have not updated I'm on top of it today you guys oh here we go alright so there's still just one eye still just one item on the agenda right that has like the ABCD yes and now I've updated it to include Peters and then yo Ichi I saw you made a con and I wasn't sure if that was something he wanted to add to the agenda or not oh you know make a comment the gas consumption issue oh oh no problem okay and then the talaq you said unifying the

behaviors of sufficient spender balance was that what we talked about before the meeting started yeah so it was I mean we ended up making a different resolution but it's resolved now okay perfect all right so the only other thing on a would be e IP 646 49 so there's been a few members of the community and a few reddit threads and a IP discussions about the issuance reduction so this is kind of two-pronged one of the prongs are when did we decide this and why and the other one was you know should we be doing this in the first place there's some few different opinions on that so for the first point from my recollection

and I'd have to actually watch through or read through the notes I believe a few core dev meetings ago we decided on this does everybody else remember that so my memory is that I participated into deathcore meetings where it is on the first run the the main the main opinion was people didn't want to touch on the subject they preferred to not not change issuance because I don't think the the discussion was metallic broad point and I think the I think the main point was about and the consensus was to keep this

one the same but related to the amount of become either per day that was already deposed on the ice age so I believe that the the amount of tree eater was a rounded thing because mainly because that was too late how many eater per day we were producing and that's sort of how we came to the teacher is that correct anyone has a difference that's exactly what I remember as well okay so now all that being said there's been a little bit of a revival in the discussion and we don't want to spend like a heat like the whole meeting on it but I didn't invite Kai Barnett or from

the internet his real name is Keller he's a CPA so Keller if you just want to give a few paragraphs of your argument for it I think it's also on the agenda page and on various AIPS and right at post but if you just want to give kind of a summary and then we can discuss that okay yeah that sounds good Thank You Hudson for inviting me and for the others for listening regarding it the debate there's two kind of issues that keep going on one is in terms of security or the word security is used and whatever the word security is used then there is a dollar cost associated with that so July 14th in batalik comment he

mentioned that 3e sounded like a suitable value for security of the etherium blockchain which amounted to 1.2 billion dollars and the reason that it's referenced in dollars and not eat is that the reason that you're paying miners at all is because they have expenses in dollars if the miners did not have expenses in dollars then you would not be giving them units that could be converted into dollars to pay for these expenses so on July 14th Vitalik mentioned that 1.2 billion if you calculated it out on the annual rate was sufficient to pay for as security for the ethereum blockchain which is

roughly the amount that I agree with frankly but since then aetherium has gone up to roughly 330 dollars which amounts to two billion dollars are 800 million more and then likewise I my speculation is by the time that the fork is officially announced there's a possibility for aetherium to be let's just say 500 dollars and then you're paying 4 billion dollars or something that for this security so sometimes the reference is security and if if if you're using security as your reference then you should always have a dollar value attached to it which many people have suggested roughly a billion dollars and then the other issue is ownership or

a minor bonus or minor ownership of the blockchain which in Bitcoin the miners own approximately 4% of the blockchain or they're given 4% of the blockchain every year and in a theory of the objective is to make it six point seven percent of the blockchain and I will suggest that if Bitcoin is able to only give their miners four percent of the ownership then I think there's no real question that etherion can also give their miners only four percent ownership at their blockchain especially considering that they've made no direct financial contribution to the ethereum blockchain that they were not paid for immediately you know ignoring all the

mining that went on one in 13 was 10 bucks ok so Vitalik I think you had responded on write on the agenda to that would you mind responding to this I mean like speaking for myself personally I would actually be quite happy with further with further reductions going below 3 ether and like if we were to follow app says like take it or the like the kind of medic your wrist ik that asks are brought up literally then we would probably want to even go down to three point a to point five instead of three my main concerns against going down going down much much further are probably twofold one of them is that I feel like it's like if we make a hard

fork that actually reduces miners rewards and on net in stem and not in in connection to adding proof of stake then that at least kind of an infringes on an on a norm that one that might at least in some people's minds be there and my potential agenda legitimized further kind of a that's whittling with rewards though I mean on the other on the other hand a lot of this is all kind of moot because because of the prover stakes which in the long run the other thing is that it if we go to if we go too far below then unlike bit well ok actually now Bitcoin the situation is a bit

different because Bitcoin has Bitcoin cash to deal with and that's been lots of fun over the last few weeks but in general we have to deal with the possibility of competition from ECC competition from Z cash competition from Manero pretty much this is a fairly substantial array of other block chains that are also using GPU miners and if the ratio of block rewards all by etherium to block rewards by other cryptocurrencies goes lower than that does negatively impact security to a greater extent then it would impact of watching that only done that basically only has doesn't really have substantial competitors for hash power so next those are those are the only real arguments

against push it up pushing it for theory Theory far down so like if someone were to suggest that zero point to those who would be the obvious would be the obvious reason started to reject that so other people in the room anybody else have opinions on this I believe so I so that was deviate fish and they were not able to join but a lot of the arguments the Keller correct me if I'm

wrong but but you you and D via fish kind of shared some of the arguments as far as reasoning but then you all differed on whether to increase or decrease but basically both of you said that having it at 3 is wrong um I sort of any and I just want to emphasize I am supportive of making aetherium more efficient and ultimately going from 5 e to 3 eath is a 40% increase in efficiency so and I would like to reply to metallics arguments in a second but sure so is there anybody else in the room that had had anything before Keller replies to Vitalik well just so-so Keller was mentioning these dollar

values and the fact that miners have kept all our issues I remember when when we were having the discussion I do not like the notion of so the doubling of price from 200 to 400 or 300 that is not a relevant something that should be relevant to the discussion of how much it's it's how much agent is right a price change from 200 to 300 or 500 honestly it's it's irrelevant in crypto we should only be caring about about what is what is the result if something goes a hundred times over or I think whatever whatever the Easton's

is what we should be worried is okay is it good enough so if the price is cold ten times less if it goes back 30 or three dollars is the security good enough and it's not really about oh if we have we have a fifty percent increase and plus I'm Stefan therefore we should not the compilation from that premise that Qyburn had I believe I would agree with that I mean the market fluctuates but this is core protocol we can fluctuate with the markets like that so I guess the the

thing would be we shouldn't be pegging it to a dollar value necessarily because that would mean that we'd have to hard fork every other month that things go wacky in the crypto economic world with prices and everything but I guess the point Alex and Martin that you were making is it shouldn't be pegged to that but we should be looking to adjust it based on the overall security and based on previous stuff we've done so like whenever ether was one dollar you know if we lower how much would it be compared to when ether was $1 and we had less minors and stuff like that is that what you're saying yeah but one of the things that the conversions of dollars is is on the minor side we what the protocol has is already the mining

the adjustment for hash power and then it's it's up to miners to figure out how much hash power per dollar they are they want in how much of this their cost and to make this decision if they want to mind eater or or some other GPU coin right so what what the protocol cares about is the okay yeah so yeah Keller what was your response to the talaq okay well you know a comment to the dollar issue so Vitalik said that 0.2 eath would be insufficient in the assumption there is that would lead to too small of a dollar

value in the competitive market which is a conversion to dollars and I do think anyone who's spending billions and millions of dollars should recognize that eight billion is far significantly more than two billion and abstractions of how you determine security values that don't use dollars I think is a little bit questionable regarding the talax arguments he said one argument which was regarding 14 and minor 14 is the primary idea and I do agree that miners are a very significant economic in social marketing block to block chains they are not merely just

providing security as some people like to reference but I investors for far easier by supporting alternative block chains that you mentioned far easier than miners and then the second thing you mentioned is the competitive marketplace for mining so this goes back into the idea that the total amount of reward for security should be a dollar value in a dollar value that is relevant to the overall market place these are ideas that I fully agree with and I think the simple way that you speak clearly when you're stating this is if you say something to the effect of and I'm just making out these numbers but

aetherium should pay for no more than 50% of all GPU minors or aetherium should pay a minimum of or between one and two billion dollars when you ignore those dollar figures and you only use it percentages then it becomes abstract as to what you're talking about and people get lost because when you said three eath before that was 1.2 billion but now it's - well now it's two billion bubble well no my argument is that there's that we we do kind of cross over from a zone of high security to a zone of lower security at the specific point where the amount of dollars per seconds that ether pushes out is greater

the goes from being greater than to being less than the number of dollars per seconds that other ball chains push out so the the main block chains that we have to look at right now would be as others in aetherium are aetherium classic and and on arrow and as cash and so with my arrow for example the block reward it looks like it's 6.66 one arrow divided by a hundred and twenty seconds and multiplied by a price of ninety eight ten which is around five dollars a

second with aetherium we're gonna go to three ether divided by fourteen transfusion to thirty which is around seventy a second so right now with yeoreum it's probably Manero has around five ET cie has her own five and when a Jesse cash has around five so altogether maybe if he REM is around that has around three or four times more or wheel with three it with three thora blocks they'll have three or four times more mining or water than the other cryptocurrencies if the block reward goes down by a factor of by another factor of ten then that would basically put us into the position we are Manero or if classic have more hash power than we do anthesis it's it's totally it's not fatal because Manero nice classic themselves

survive totally fine but it is a lawless security zone yeah and I agree with that I understand the lower security zone and and so you're right your references that aetherium will still pay roughly four times more than the competiting competing block chains with three eath versus something else's is that what you're stating roughly roughly that with three es etherium will pay or something likes three times more block reward per second dollars per second a block reward than all the proof of work all other non asic before qua chains combined right okay it oh then all others combined yet that's that's roughly what I thought as well and so this amounts to for those

who are curious this amounts to roughly two billion dollars versus five hundred million for all others combined this is about right I think so and it it I'll just say one other thing that I think it's worth considering and this is where others in the developer community they like to they get nervous when you started attaching dollar values to things well you're paying two billion dollars for this block security I don't really want to guess what the expenses are for the etherium foundation but are they less than two billion dollars a year no they are vastly lower and it's I mean it I actually know I'm actually totally fine with the idea of putting

dollar values on how much we're burning on how much we're burning on mining I think it's a definitely very good way of kind of keeping it real in terms of like letting people understand how much money is being burnt on this I have consensus security expenditures so like as I've said I'm actually like in principle totally not even opposed to further reductions no the thing that I will say though surge and two is that I think there's while I feel like there is like good community consensus for reduction either two three ether or plus or 22.5 just two there's not much commute there's still not enough consensus for

reductions they like either like 1.25 or or even lower yet and so I feel like the kind of the the right thing to do would be to at least four but for Byzantium stick to three its stick to three or two point five and then people like yourself and end of lad will you know be free to continue to make your arguments and if there is community consensus for further decreases then that's great and that can happen in later Forks uh I just did like one statement if you want to make this statement now and or if you want to think about it this is the main issue the main issue that I have is that it's still unclear in terms of dollar values

is it so July 14th you said one point are you indicated one point two billion would be a suitable dollar value for security now it's two billion so let's pretend this is just a number changes we all know so if the dollar value of aetherium is six hundred dollars on the date that you're planning to make the fort and if the expenditure goes from 1.2 billion from the time you announced the change or suggestion to to a four billion dollars is there going to be any consideration that at the time of the planning it was 1.2 billion but at the time of execution it became four billion um I mean it's definitely something that

should be considered though as people said before like differences by differences in the market of a factor of 1.5 or just compared like compared to the general scheme of things that could happen in the crypto market fairly small and totally could reverse themselves within the span of two weeks I'm yeah and the other thing I think is that at the time that that happens if there's a huge community push within the etherium community to lower or increase or anything like that that's when it would be necessary to look at it again and even for this kind of thing and for the stuff that you're arguing Keller it's it's more we're discussing this at length because the community wanted us to so I think that there's

that would there would definitely be consideration but you know it's one of those no one can predict how everyone will feel if that happens especially if people are rolling around and you know gold coins ducktales style whenever that happens like if people will want to reduce issuance or not or if there'll be a real community push I probably agree with that as soup and I mean Hudson and I yeah it's so that's why I feel like basically more or less I will bring up the I will only bring up the issue when a theorem is troubled or whatever because this is the only time that people are concerned about improving things when things are going well people don't they're not concerned about improving things yeah but but but what

but what we can do to make matters clearer going forward it's just you come up with some dollar value any dollar value and I agree that you know maybe 1.5 is too small in terms of in terms of blockchain variability but if we just say between one and three billion dollars like if it goes over three billion then we'll consider that a bad thing you know less than 1 billion is a good thing that our I mean I mean not less than 1 billion but you know 1 in 3 billion so I would actually still say that like this idea of like a fixed amount for secure a fix the dollar amount that's not the best heuristic in the long term because if we imagine for

example of the price of ether goes all the way up to 3 million dollars then we do not want the 10,000 factor reduction in the reward and the reason is that basically the size of the attackers that we have to deal with is all and the incentives of the attackers we have to deal with will also go up and so security expenditures probably should should go up always somewhat with their kind of economic activity size of a blockchain okay now in that issue is another one where it becomes quite complex when you consider the price of aetherium because I knew investors must pay for this mining reward as well as like Don and expenses but they're they're very small compared to the mining reward relatively I think foundation expenses

are probably one percent of the total expense expenses of the etherium blockchain and 99 percent of the minor bonuses so so when you when you talk about the price of aetherium and right now aetherium has to create roughly two billion dollars from outside parties into aetherium so if the price of aetherium was to go up to i I will say a market cap of one trillion dollars which really isn't that big in terms of the price value increase okay okay 30 increase yeah sure so this is the biggest increase we could imagine so if so if a theorem goes up to one trillion dollars at the current rate you

would have to get new outside money equivalents roughly 100 billion every year and eventually the rest of the world runs out of money we know this is so killer just just to clarify a few things here a lot of people are not a foundation thing and it's not it's not it's not our job to worry about sustainability of the price we or or or job here as a dibela art is as developers just to keep the network secure and ultimately I think everyone believes that the easier issuance should

at some point who the only target that he wanted at some point in satara to should drop to zero or near zero when when we have force take working fine so the reason we I really support the issuance is in the end because or go it's long term to reduce it to as low as possible when we can have promised take on so it's not I mean so you're talking about the foundation as if it's this is this is a foundation meeting and we are discussing like this yeah I think that I think the language does matter and that and I I think that that's so partly the other language does matter so that's been a I guess that's been a

number used in a lot of the examples and the e IPS and I think that might be why he was bringing it up Morse more so than to imply that this is a foundation only meeting but I agree that that's a good point for especially for the people listening that this is not a foundation meeting and this is you know all types of teams across the ecosystem so Keller thank you so much for your comments on that anybody else in the room do you have any rebuttals basically what we've landed on it sounds like is for you know the purposes of you know buys baizen tinium or is it bison tinium am I saying that right yeah byzantium gosh that will keep it at three that is a partial reduction and that's a good medium between you know

people wanting to keep it the same and people wanting to lower it to you know much lower values so anybody else have any comments on that I totally agree about it being a good medium hopefully not too controversial yeah either up or down no yeah and the grand scheme of things like I've seen about six to ten people total raise alarm about this and some of them are like Keller who actually wanted lower and that's what the alarm they're raising so I would I would say that by and large not even looking at carbon vote that there is consensus to do this and it's at this point it's not just for some people have

suggested for investors or against miners and I I have not seen that as a reasoning for doing this either I think the arguments that have been made are good ones so any other comments on this property water effects or blockchain tests in Byzantium and Constantinople so any change here would imply say ma'am week or 10 days tbh yes and effective of testing yeah so everybody might who's listening might not know this but Byzantium is very close we're at the

final stage so any change we do at this point needs to be read in all clients and all the test cases basically yeah and there hasn't been any change from the three the the issue its reduction to three is that what the current tests say basically in line with a IP 649 yes so I'm in the course of preparing all the tests okay she's with this I'm running a commander that takes fifty five zero hours and when the book report the changes I have to do it again show me the texts free Wi-Fi to zero hours this might have where this might succeed it failed twice yesterday under the three days ago but this time it

might succeed so after this time so considering this these kinds of procedures I said I takes one week or ten days to the flip to it hold on why does it take two days to create tests like not fifty hours a study this is for Heineman I became not to run general Stata tests so we have to translate the general state testing the blockchain tests for that we need the mind blocks that's why it takes hours oh it's here okay so this is a good segue actually so I think I think we're done with that point thanks color and everybody else who participated so the updates to

testing so we have jared in here who's a recent addition to kind of the etherium j/s and testing team and also I believe I believe Tim has been helping as well us is that accurate Tim and Jared to say that you would be considered a part of the testing team yes yeah yeah I've been trying to make some efforts great okay and then so the other people who've been heavily involved would be Demetri who isn't able to make it today yo Ichi and Martin so I'll let either you know each year Martin whichever one wants to kind of give us an overall update on testing and let us know of any road bumps or sorry yeah that are in the way

that they would need help with or anything that we can do to help okay so I can't do something so basically it's done except a few known problems one of the main problems is the one I talked about the high test heavily out-of-date so I needed updated with this 50-hour come on I mean the course of doing that next week I hope we can see this type test updated and we will start seeing a number of errors decreasing on this height board and then we can actually start measuring our progress there are a few more known issues these are very

very tiny there is state general state test fading on all clients including if it's a mine this one annoys me a little bit because the test generation for the frontier network fails mm-maybe SiC occasionally implementation we might have made a mistake I mean I'm investing in investigating this there are this few more known small issues but these are much much tinier than the hybrid test regeneration okay thank you for the update and by the way

I forgot to include Matthew and Casey they're also part of the testing team so the testing team has definitely grown since we kind of embarked on some of the stuff with this hard fork a few months ago which is great so let's see here so thanks for the update yuuichi so is what can we continue to do or what can we do to help as far as testing goes or are you guys just pretty much still on track so why'd Demetri and I busy fixing updating existing tests there are still new test cases

to be written it affected many of them while debugging in kids and other clients so in the Gita Ezreal testing channel I will post the distal this case is to be written and I'm happy to provide instructions so yes gift I am Adriene tests I will post some instructions for those interested okay great so we'll probably revisit this in a second once we go through all the updates from the client teams but Martin Casey or anybody else any other testing updates regarding hive or fuzzy testing yeah case if you want to mention your

state test CBM thing yeah sure we I posted a link to the current test results in the testing Gator room I posted in the Google chat show these are the trace equivalence tests so it's running on four clients all the state tests and then checking the EVM traces to make sure that they're equivalent and the ones that are failing the the checking for trace equivalence does get it shows more failures than just

checking the post state hash so this is a better way to find consensus bugs then the way state tests are currently checked for by just comparing the just checking the post state hash but it is also trickier because failures may just be differences in the in the format of the of the EVM trace and not actual not an actual difference in functionality so we have to maintain a patch on each client until until we get merged that that uh

changes the way that makes the tracing the the trace format equivalent so that's been a lot of work but it is it is just about just what I've done so then and once this is once it's complete then fuzz testing can work is if we start fuzz testing before the traces are the trace formats are all equivalent then we just end up with a lot of false positives where the differences are in the trace format and not actual consensus bugs wow that's like really cool I would be lying if I said I understood everything from the link I

just clicked on but it looks really awesome but yeah I just threw the the files in a in a folder because I'm still working on the table to actually make this easier to to see as a scoreboard as you can see I post a link there to one log which fails and if you go to the end of the log you can see the actual comparing of the traces and you see that the forth operation in the execution there is a consensus failure because the top stack item differs that's really powerful and really cool great work Jason yeah Casey

this has been great ok cool any other updates from testing thanks again Casey yeah I had some questions as far as I give you guys a better a better accurate idea far as like the top code performance across different architectures and across different operating systems so what what would you guys like to see as far as it being tested on and different scopes like specifics on that and then I could try and capture that for you back to you guys so you have a more realistic closed eyes personally I think that we need to to decide here now kind of the what

we're going with for bicenthium without necessarily looking at different architectures at this point however for the long term of course will be very interesting to have a good set of benchmarks for different architectures and and different dunno different versions of go or different compiler settings and stuff like that but as so as for deciding what to do with bicenthium it's kind of it can't it's too late relax that's my view okay sounds good and did you guys just get

try to give the C++ benchmarks or just focus on go or what was priority for right now so right now yeah we're done at that point I think we should I think we should focus on go and parity and not just one client okay so I don't know do we want to skip ahead to covering gas prices I want to talk about that or do we do blended updates go ahead and skip to gas prices that's just fine so for the best price in Europe coats nothing post your link yeah so so this analysis that I posted

to it contains in this in this report there are raw beta for TPP ethereum and for Jeff and I think there are five but gets benchmarks and one CBP benchmark and which follows it and the yeah the C one is the top one except for mod X where C++ they are not that hard to see it was possible I expect to get to be the slowest of the CPP at CBP parity and gap because various things Peter mentioned that it might be that goal

does not optimized for seeing the instructions so this this analysis mostly based on yeah and instead of focusing on the raw meaning gas per second which of course varies depending on your architecture I was hoping I thought it might be a better idea to aim for about the same gas efficiency as easy recovered and if it turns out that they're all too expensive and then at least at the same

level and we can do kind of raise or lower them all in one place when I want to so yeah so this is just in this analysis you can check out the raw date down the scripts there and say more Foster is use the gas was a thousand for tradition modok's to make it bit more about 1.5 more expensive than earlier I can say an exact figure because molix depends on the input the cost the council members well just after just some constant there scalar

multiplication should be around 30,000 guests according to these measurements and pairing should be roughly double is there let's see and so those are for see those are forget and for c++ right you said yes what about parity do we have those no okay I think I think we should really

try to get those benchmarks for parity I think Arkadiy had mentioned that they had already ran some Arkadiy do you have any updated benchmarks well yeah have benchmarks that being run on my machine idea if it should be run on the same machine okay would you be able to work with Martin on that or with Tim on that I think Tim's actually might have more resources right now how do you talk to Tim at all yeah sure I mean it's not that hard just to build them and run them I have an instruction to their scent to mark in our they're cool yeah I'd be happy to help you on that I think

I just working on that a few weeks ago so I'm a little familiar yeah I would agree but easy recover should be used as basis because all the major clients use the same library and the same implementation of UC recover and it should be a good basis for calculating average guest costs for other instructions arsenic do I don't know what suspect areas are are implemented in parity but it will help if it is same inputs as yeah at least for harmonics I think the most was the same and the other

shouldn't matter I think okay so I guess the question is when when do we when we want to set the price is definitely by setting the gas prices will that cause a lot of the test to change dramatically or is that something that can be set in right before we start the testing phase yes I expect all the tests when they about that then let's set let's set a hard date then and I think that that dates gonna depend on when we want to launch the test net which sounds like we wanted to do that will maybe early September to late September right you

know I'd say or it seems like we could get away with early September quite possibly and actually that's a better question just to skip to that just because I know we're hitting 10 o'clock and some people might have to leave it looks like and we'll go over with each team but it looks like a lot of the teams are getting by as far as implementing byzantium updates so yo Ichi Martin everybody else would early September be too rough of a timeline if we in the next seven days are able to get these gas costs down in stone well the test network yes so we should have definitely tried to do that okay Martin do you agree for testament is what I

mean yeah I'm a protest that can do whatever my oh yeah because I guess that can be changed on the fly since we will just have the test net clients working on that note let's go ahead and run through unless there was anything else from the testing team the second I was going to run through the updates from the client team okay great so get any updates or relevant info from you guys I think Peter would yeah yeah sure so we've implemented everything everything is merging master and all necessary so from our part they're ready okay parity you know the tests already will take some time to make sure they are passing

so take about a week or two okay great sounds good and then C++ I'm a study done but I'm seeing maybe so yeah but it doesn't maybe maybe when we could should be fine to some piece okay great yellow paper the product Christa already it's now about much in them but I guess this

can't be the date okay and then also were you able to talk to GAAP since the last meeting about some of the issues you were having last time fixing spurious dragon so what I want to do is I want to check what happened wing out three particular blog I so many comments on with discussions on github and everywhere but I wanted to check it in the capital I it remains disappeared appeared and then I will talk to him but this is you know like a historical investigation and my first focus is no

more attention got it okay great so it sounds like that's going along pious app Vitalik yeah so PI a theory a mess I think I mentioned is passing the state tests it's still failing a few people a few block tests though I think that's due to a combination of me IP 669 not being like solely merged into the tests yet and and these some of the tests having empty accounts that shouldn't be there and that's something that I've that's been fixed for some of the tests in one of them in one of the recent releases but not all of them so you know just waiting on that okay actually we have a

theory of Jay here today I believe Anton are you here yep I'm here awesome well welcome and is it still called aetherium Jay do you want to give kind of a reintroduction and let us know where that where the team is that as far as getting it up to speed we are roughly 70 percents ready with VIPs so let's shortly great thank you for that are there any other clients in here I don't believe so or actually you know what Kasey or Jarrod if you guys have an update on a theory MJS I don't really

have any updates at this time maybe Casey it's something more yeah it's it's progressing and I'm optimistic that well it'll be fully compatible with metropolis by the time the it's released stock by the time metropolis is activated on it HS net uh-huh robson okay great so let's see we already determined the GAT or we already had plans for the gas prices all right so now we get to the point where we talk about the time estimate for testing and release so we just said early September's when we could probably start for a test net that if my memory is correct would run for

between two and four weeks something like three weeks would be a sweet spot from what we talked about last time as far as deciding that we would decide by block number is that correct I think that's how we did the last one yeah mm-hmm tends to be the easiest way to make the decision is this something that we should figure out in this call a block number and just give like rough day I feel and I feel like we kind of know when at this point what the ballpark ballpark is so like the super up to the absolute flying colors super optimistic would be four point three million which is September 22nd and the pessimistic Steve is four point four million which is October 27th

so somewhere in between those two depending on how quickly we can gets to everyone passing everything and all the tests being done okay well we could have an unconfirmed block that we pick today that's not like official and then from there we can come over the next week like be looking at that date and then make a determination but at minimum we can make a testament block number decision right now it would that be would that be accurate for me to assume to switch the test net over I should say I mean if I get once geffen parity or right I would say like we should just do the test that switch over

maybe it's something like three days after a gap and parity both give a green light okay that sounds good to me it would be a good idea to if we're gonna reboot Robson with a with the clean a clean chain so that it's easier for people to sync up and reach the metropolis activation block then then guess and parity may want it we may want to coordinate on the new Robson Genesis blocks if we're making a new chain shouldn't we pick a new subway station yeah actually what do we be making a

whole new chain or would we just be clearing out Robson and using the same ID hello wouldn't it be a good idea to keep ropes them during the transition since we're changing the block the word base block number yeah I think that would be that would be wise so we wouldn't clear robson then well we could clear it but still start with with a week or two of five ether block three words so what would be the benefit of clearing Rob's Tania's I know enough people use it that like you know some history is gonna be swept from nose you

get for testing naps well just make it a lot easier to think oh I see it's hard to sink it today yeah there's quite a bit of spam like passing us attacks okay well but but but the same thing can be solved so if you care about syncing up and planning so mist is already using Kirby so I think there is an advantage of yeah I guess we could do both I mean I could have two proof-of-work chestnuts

mmm I think that's kind of complicated I think that if the main reason we would be clearing it out is a sinking issue we can just find creative ways to make it sink faster so for instance having a few people put up semi quote official you know state you can where you can download the blocks basically you know I'm talking about like you can export the blocks and then download them you can put them on like a BitTorrent or something by the way just my two cents I would probably vote against restarting robson before the couple is hard for if we have all the I think it's a good thing that we have a mature tester fits a lot of junk in it just helps people

basically ever since everybody's using it for testing at if we if we fork that into a trifle is at least everyone will test using the new code whereas if you just launch a new test not and gets to guessing most of the developers will just keep using old clients it kind of beat the purpose yeah I agree I think I'm four so just to clarify I would be really glad to clean out the test nut and start a new test that just not poured before I see okay yeah I think we should keep our obstinate is there a so everybody on the call what is the best way to sink Rob Stan

that's like the least painful I thought it would be exporting all the blocks and then just importing all the blocks from like a downward probably not but then you beautiful fully important little bit better to do a life either light client sink a fast sink or a snapshot you know warp sink so anything besides a full import I see yeah so if we want we could we could export just for example I mean forget it we got itself or we could do an export of the just a lightweight empty database with just the recent state and then we can just shove it up on the big grant

yeah that would still make it functional right like people can use it yeah oh okay yeah can we do something like that metallic where you saying something I think I saw your mic pop up no okay so yeah that yeah that sounds good to me so I think that as the agreement that we're keeping Rob stand and we're just going to decide a few days or so after parity and death are agreeing on both gas prices and other stuff that we switch that seems to be the agreement okay sounds good and then we're gonna shoot for somewhere late September for the actual hard fork late September

could mean anything so just via the second half of September okay brick does anybody else have any other stuff that was non-agenda no but I would like to I'm hoping that we can accept the gas prices in stone early next week and I hope that this some of you guys are available the public are you going to be have a good a very good thing Monday Tuesday Monday Tuesday I'm gonna be at conferences then not like a voice

I fail ability but like Skype text availability depth on and off definitely good I guess that works yeah I'd love to help with that too but yeah maybe one you guys can help me figure out the logistics on how you guys were calculating that or to decide though calculate it yeah perfect okay and then yeah well we'll get with Arkadiy and I think our Cadiz main on that so what we'll we'll make sure he's in on that too sound good Arkadiy yeah sure great one more thing I'm gonna post this link in here so crypto mines is a group that represents some of the miners and they felt like they weren't really hurt in

the discussion earlier so I'm just gonna post a link to their comment we don't really have time to go over all of it but if someone could address this and the agenda that would be great just with your opinion we can't change it from what it sounds like because we pretty much we would have to do a lot of stuff with the test cases so we probably won't change it from the from the reduction of three but I do want to take this chance to say we need a way to signal for what the community really wants because some people are feeling left out so let's start to kind of brainstorm some of the ways we can do that whether it be on blockchain or not especially since people felt like they weren't being

heard because even if you know they don't get what they want this time as far as like they think that the production shouldn't happen at all we need something that will make sure everybody's heard next time we have one of these in a into a more controversial issue you're asking one of the hardest problems in the space Hudson which is to be able to measure community or make some sort of hole that's very hard yeah you got this Alex I believe in you go forth I wait for the multiple ICO that some some dollars trying to do a identity and KOTOR key and prediction market than that yeah yeah definitely okay cool well yeah but uh sorry about

that crypto miners I know you felt left out of the conversation but I posted your link we're gonna try to respond to you but at this point because of the way the timing of this and how soon we have to make the fork there's no reason we can't adjust this in the future it sounds like so as long as we get community consensus around it but yeah I'll I'll be looking into working on the getting minor communication to be in the forefront in the future anybody else have comments yeah I would like to said I think one of the important things is that we have this open us about the eats which eaves are currently under discussion for being included in in fork X Y and C and what has been decided regarding the eaves so

that people know which eats to keep track of and start to comment on if they have input okay great yeah I agree Martin any other comments but uh if you don't mind me asking the question I know this is a hard one there's been some talk about that the eath will get readjusted at some point in the future I am sometimes unclear as to if this is some time like let's say like June or later 2018 or if there's some other plan that I'm unaware of that might occur

like in January or something like that so there is going to be a hard fork after byzantium called Constantinople but as far as issue issuance changes go like I personally like see all that the wisest thing is to just basically keep it at three it for now and when we do the hybrid proof of stay exchange that's going to be an opportunity where it actually makes a lot of technical sense that you are very substantial for their decrease and it's you know we can use to all work together toward making that and making that happen right in regarding the hybrid change is that some something that might be expected somewhere early 2018 or is it lovely not ugly more late than early

got you got you okay thank you guys just asking yeah the research team has been killing it lately meaning they're doing really well I'm using metaphors way too much they're doing very well lately and adding people so I really feel confident and I mean metallic would know better than me that the hybrid proof of work proof of stake is gonna come earlier than people expect but however not to say you're throwing yeah Oh perfect yeah so to keep up with progress there's a link it's just github.com slash aetherium slash casper so that that's really cool how fast it's going i've been following some of the progress and I should say everybody else should be looking at it too because it's really cool stuff alright I think that that is it let's see yeah I think that's it

anybody else have any other comments great we're ready to hard fork not the second but we're gonna be ready really soon so I'm real excited about that everybody I'll talk to y'all in two weeks goodbye [Music] [Music]

[Music]