Bisq Developer, 13 Feb 2019

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Bisq developer session 13.02.2019 - DAO Q&A

Description

Bisq dev session 13.2.2019

https://github.com/bisq-network/growth/issues/100

A

Hello: everybody yeah. We hello everybody. This week's gross called developer session this week. We want to focus on the Tao and mainly on basically a Q&A and discussion about open questions so yeah there are already quite a few videos out where we discussed all the basic concepts from a developer point of view. So, if you're not familiar with this, please check out those videos first and let's jump directly into any open questions.

A

Gloria as I talked already with Christophe before Florian and too much did you had time to check out the DAO videos.

B

Yes, I enjoyed them very much. I, certainly didn't, but I read the docs yeah.

A

Okay, yeah, okay, then, let's jump directly into it or what a the open question. What is what is unclear.

C

So sorry I, can you hear me Manfred, yeah, okay, perfect yeah, so unfortunately, I wasn't able to finish one at the last video on color coins, as I discussed before the calls I'm, not sure if it's come out there, but yeah. It's clear for me regarding the initial Genesis transaction that we just follow all our transaction outputs to find out what is a bisque talking, what what is not! But if, on every conversation on in each motorcycle, we are creating new psq from scratch.

C

So my question is, and you started to explain it before I call and how how the generation the coloring of these coins is created. So for the initial ones, it's the chances transact transaction, but for the new ones during each. What exactly are the must be also something similar like this eeeh.

A

As a conceptually, it's basically something similar like the Genesis transaction. So every contributor has the right to create a kind like a dedicated small Genesis transaction and the output of the Genesis transaction is bsq.

A

The real Genesis transaction is hard coded and by convention. Well, everybody who use vsq is basically agreeing that they are accepting consensus about this Genesis transaction and from the yeah. The issuance transaction from a contributor will be authorized by voting. So when you make or your compensation request, you are creating a transaction where you put some Bitcoin Satoshi from your Bitcoin local Bitcoin wallet. You remember, one PSU is 100 Satoshi, so let's say you're.

A

You want to issue 10 psq, that's 1,000 Satoshi, so you take 1,000 Satoshi from your Bitcoin wallet and you send it to a bsq address to your bsq wallet and at this moment it's just symbolic because it doesn't has a route in the Genesis transaction or in another issuance transaction. So it's this output, which you, which points to a few address, is not considered well, it'd, be as few and after the voting. When you got voted, ok from the network, then this output is considered as well.

A

It B is Q also with the with the acceptance into invoking this compensation request. Transaction becomes a issuance transaction of the transaction. Technically, it's the same just the context. The context is different. Now it's this output is interpreted as P is Q and you can spend it and you can sell it and the the same rules apply like with a Genesis transaction. All the following output of this one output is considered as PS q, so yeah to summarize again, I predict later concept like Chinese, a transaction.

A

Just a Genesis transaction is hard coded, and everybody agree by using the software that they, this is the well it Genesis transaction and the issuance transaction is decided by voting when you get voted okayed and it from the network. You get authorization basically that you have created new psq for this amount. What you have requested- and this particular transaction output, is now like one output of Genesis transaction and.

C

Information for this are Genesis transactions is done persistently it's B, or is it some more piece yet also.

A

The the Dow State class is the basic or model for all that out later and.

A

Folder folder the immutable the out later, as with every block data get edit, but then it cannot be changed anymore or it's also not really correct, because you have something like stance, meant transaction output map which changes when you're spending some data, but it's basically the current state of your yeah off the towel. Blockchain related data, so yeah there are.

A

There are other data like the peer-to-peer network data like the the computation requests and then the boat into blind boats, interval yeah blind, wrote civil driven on their transaction, so the holes Dow model consists of this Dow state and the peer-to-peer network data like the blind road and computational requests and their transactions which are connected with this peer-to-peer network data like the compensation request, transaction blind, go transaction and the Boqueria transaction, the VOC reveal transaction doesn't have a related peer-to-peer network data there.

A

We are putting the data directly in the blockchain as it's only their private, the secret key for decrypting, the pie, boat and I'll. Sorry I lost a little bit your original question and.

C

That question was just regarding where this information is stored and the follow-up question is we when we create a new version of the client and we also bundle the current state of the adult state. We use the the newest kind of snapshot, but before of bit crunchier, this also.

A

C

A

Not potentiates just a snapshot from our state, basically.

C

No, what I meant are because.

C

Yeah each each SQ transaction is a transaction that is published to the Bitcoin blockchain correct yeah.

A

But we are not using bit country for the Dow, also it's used in some areas, but not in relevant areas, not in a consensus part in the consensus part as when you run the Tao as a foreign node. But when you have your local bit, I think who are running and requester blocks over RPC, then yeah, you get it directly from Bitcoin core. You get all the full block with the 2,000 or more transactions, and you are passing all transactions for well, it appears queue transactions and only the vsq transactions.

A

We persist and we we keep in in this tau state and this tau state every 20 blocks.

A

We are creating a new snapshot and with that, we avoid that when you're restarting the application that you have to pass again older past blocks and- and we are delivering such that snapshots with the binary- so that the user don't need to pass back from the Chinese's transaction everything he can do it. Also in the settings, you can delete the tau state and then you are stopping over again from the Chinese's transaction.

A

But, of course that takes some performance after at least a few months or a year, and when you're at the light note, you get the blocks from the seed nodes. It could be basically any node, but we are using the seed notes for this because for convenience basically- and they are- they are sending you already the filter clocks. So you get a list of blocks, but the blocks only contain the psq transactions, so one block usually contain only one or two or three transactions instead of 2,000 and with that the data is much smaller.

A

What you're transferring- and they only trust a security issue here- is that the seed, node and malicious seed node could theoretically or meteo tater and not send you some transaction, and you don't learn about this transaction. It's similar like which has pinyin SPV mode full note could lie about the that ya. Don't give you some transaction, but you cannot, if you're, a male transaction or put the transaction in the into the block which doesn't exist and similar, it's here and yeah and you're requesting this data from different from several signals in parallel.

A

So it's very unlikely that you get some issue. It could can be some stability issues or some real issues. That seed notes are on a different fork when, when you have went to mine, find blocks of the nearly same time and the blockchain is working, it can be that two signals are on different for accent and they're, delivering different blocks, and it's basically inconsistent to state.

A

But we are, the application, would would show this as a warning and would try to recover from the latest snapshot again. So when anything is out of sync or the state are basically yeah goes back to the latest snapshot, which is 20 blocks back and with normal reworks they're, usually I think maximum four blocks deep and then rebuild again at every new block. So as soon when there would be the event that our two signals have a different block, then you try again.

A

You have basically in this in the UI, then some visual representation that you're not the network, is not consistent. So you have to wait a little bit longer on the next block. Usually this rework scared very quickly after one or two blocks as soon as yeah, with every new block, you will try to rebuild it from the last latest snapshot and then, when you arrive Inc in a consistent state again, then everything is fine. Otherwise you do it with the next block until it's resolved and.

C

And because we will bundle this, these latest snapshots with each release, we don't expect any performance issues. Let's say a couple years from now, we have thousands of additional Genesis transactions of different contributors that needs to be followed, but that shouldn't be a problem then, because.

A

You are downloading, after your even account, you're, just downloading the new blocks, which is missing, which you tell us to feed note. What's your latest block. So when you are yeah when you, when we make a new release and to use it's downloading the new release after three weeks, then he started loading the blocks of three weeks, which is also not a lot of statements, especially because there are not thousands of psq transaction, hopefully in the block. Hopefully.

C

Hopefully there are thousands of uses injection.

A

Yeah then we have the scalability issue, probably because we have to but yeah I think that it.

A

Would be another problem, but that will not happen so yeah.

C

A

There have been some performance issues with that the UI got stuck, but as far as I've seen and I investigated a few weeks ago. It's mainly that the UI is doing too much work and it's not optimized at the moment, like yeah. Some lists get get refilled and get filtered instead of one time many many times and need to be optimized and fixed, and before we yeah when you're getting all this patch of blocks.

A

When you get whatever 500 blocks from the seed node, then you're passing every block and before we were triggering the UI and two listeners as every block when a clock were fully passed and that cost a lot of performance as I think it was something it took for for one week or so.

A

It took one minute to pass and basically the UI was stuck for one minute and after the optimization it was something like two seconds or so, and the only real change was that we were not calling the listeners at every at every block, but at the end of the pass of the batch processing.

A

So after the last block has been finished and you're done with the whole batch, then you're calling all the listeners and and then they updating and then it's only one time hit, and there is still a lot of optimization possibility of surpassing itself is very fast and even the air also can also cause a pass itself can be optimized, but that's very the interest code. It's all consensus, critical and and I yeah I.

A

Don't think that there is a lot of overhead other I think on the UI side, how we are feeling the list and how we are filling the tables and create things, especially in the boat results. There are some complex of tricks. What we need to create just for the UI and the creation of this object and passing through all the later or not the passenger, collecting of all the data and putting it into the new form which we need for this, bring it in the UI.

A

That's not optimized at all in the moment and has to be optimized Sentara. We probably can get out the most of performance. I mean there are some scalability limits, etcetera and there are really a lot of transactions a sinker even after 10 years. It should be an issue when they are not too many transactions in one block, because you just get all the old data it's so you get the snapshot.

A

It would be only issue when you want to rebuild it from Genesis, but then it's like Bitcoin core when you want to rebuild the whole blockchain, it takes you two days or whatever, basically will not take two days, but maybe it takes in one hour and but that's not required when you want to do it. When you want don't want to trust the developers that they are put in the correct that shot, then you can do it and when you trust the developers, you don't need to do it.

A

But it's of course, by default, made easy that users should be able to do it easily, and hopefully some do it to have double-checking and when we would put in a wrong snapshot with some invalid transaction would easily be detected, that some users complain that they get a different result and we need to figure out. What's what's going wrong.

C

Okay, so for for people who update on a regular basis, so each month when we say we have a release each one or two months, then it shouldn't take more than let's say 30 seconds or minutes to get a usage again. I.

A

Think it should be even less I think the pure passing where the UI get more or less stuck because we're running it on the UI frame. Ui. That's right. I tried to use threading for the passing initially, but it ended up in threading hell and so much complications and risks that I prefer to not do it and, as I found out that the passing itself is not, the bottleneck fits quite fast, yeah, I, see and so I think the first thing to optimize.

A

It would be to fix this data aggregation and preparation in the UI I think there is 90% of the performance burned at the moment, and when we still have issues we could we had try to make some kind like delayed her execution, that if the UI some time for a rendering again but I, think it's not the problem and a user who has not used bisque for half a year and started again.

A

He will get stuck anyway from Pitt countryside other than Pitt country will take not only one minute for probably three four or five minutes to catch up and to sync, and then the CPU is 100% for this time at the UI super super irresponsible, even if it's a different thread and so but yeah it. It takes just all the performance.

A

So I would not care too much about those users who don't use bisque on a regular basis. They have their. They get a little bit of performance hit. But it's already now the case and no easy workaround. For this.

C

Okay, yeah thanks the the other question I had mentioned on that, then transactions are sent to appease queue address. So that means that there is a.

C

This kind of address and protections if I look into a regular Pitkin a block Explorer will se. Then the esq address there as well. Yeah.

A

So that in bisque we have two wallets the piece few wallet and the Bitcoin wallet. Technically, it's Bitcoin wallets bit country doesn't know about bsq. It's just on yeah we're using this bit 44 which have so yeah different branches for the coins or the Bitcoin is I, think zero and then any any color coin has a different. The number and I think we are hundred forty two or something, and so with that yeah.

A

Both wallets use the same route and two same secrets, so they are technically built from the same route, but the address is sit on a different branch and the public keys are really different. So you could not, by accident or input psq address keys into into another wallet and then spend it as Bitcoin. That was the main reason why we used to be 44 to get additional security when some people don't know about it and just I try to import it in an electron as another wallet and then spend it as normal it couldn't.

A

They would destroy the psq and can be a lot of money then, and we forgive what it's not possible and.

C

Where is it written down which have the psq tokens on which branch they are positioned in their article, if Steve wallet in the.

A

Yeah, that's in the Bitcoin package. The management of the two wallets and startup of creating the two wallets are two different files, but I said they're using the same seed words or usability, basically that the user don't need to make two times a backup of the cigarettes but because they're living on different branches in the arabic deterministic wallet structure.

A

You cannot make a mistake by importing psq keys into in and use it as a bitcoin key and yeah when yeah, when you send or when you create any transaction, any psq transactions always contains some Bitcoin input and some builds cue inputs. The I mean not always be a skew input like when you create the compensation. I also got to see you need psq inputs, yeah. The Bitcoin input is always required for a mining, see because we don't want to pay mining fees with psq. That would be very expensive.

A

So when you create a transaction, you always start by yeah preparing your psq part, the inputs. Then you give this partial transaction to the psq wallet. You are adding the required inputs from the Pittacus or you're, putting it to the Bitcoin wallet service and you are adding the required inputs for the mining feed on the from the Bitcoin wallet.

A

And then you return the partial transaction and as you sign it from the particular thing, the Bitcoin input gets signed already at this moment, and then you give it back and then at the end the psq wallet will sign the its input and then the transaction get published.

C

Okay, thanks yeah.

C

The in the last class- yes, sorry, the last question: I have sorry the last question I have is then regarding I uncovering, do we have this kind of case? That's the case when we burn them or is the earning of esq a different process? No.

A

Burning is uncommon, it's still not really yeah. We still have not really a good term. Uncovering it more correct, like burning. We are not destroying it. We are converting yeah, we are uncovering, as Pitkin will convert it back. Psq will get converted back to Bitcoin and lose the psq value in that event, but the Bitcoin value is still used for as mining seed, so like trade feet transaction.

A

So when the user is paying the trade fiend bsq, he had some psq input and then, let's say, and there we can burn as small amounts like one Satoshi. Usually you cannot transfer small amounts like the dust limit, that's for a 546 Satoshi and that 5.4 six-piece view. So that's the smallest amount. What you can send to somebody when you buy some psq, you cannot buy to be askew because yeah that's below the dust limit and nobody can send you this, but for burning or foreign coloring or untagging psq.

A

You can use a small amounts as one Satoshi. Let's assume we have a trading here once Akashi, so you have maybe one you take so in bsq with with 5,000 Satoshi. So you put this as input and then you create a change output with 4999 Satoshi to another bsq address of your own wallet and the one and then the rest is just Bitcoin input and outputs change, outputs for the mining field, and this one Satoshi, which is missing as a rule, is that the first outputs are considered psq long as there is enough input.

A

Impious few is covering this. So when the past is going over this transaction, it looks okay, we have an output with four 4999 two appears queue address and we we have total vsq input of 5000. So that's larger, like the output, that's everything! Okay, so the first output is considered as PS Q. Then it goes to the second output, which is then the change-up good for the Bitcoin, that's whatever, let's say: 10,000 Satoshi R. So we have one p. One Satoshi left it on the input side, but the output is 10,000 Satoshi.

A

So that's much larger, like we have available so automatically that considered as Bitcoin and not to speak to you anymore and this one Satoshi. The difference is then, basically about this parent or what this uncoloured and that will be added as mining fee, so that we are not that we do something useful with it and the mind. Yeah, you pay one Satoshi less mining, see from your Bitcoin input and you're using them. This P is Q yeah, this one's a Tasha from the PS q input as mining fee, and that's it.

A

The are blind regarding PS q. For them, this P is Q. Input is just a Bitcoin input and they don't understand anything of this Q. Of course, and by that we are yeah. We are uncovering the PS q value back to the Bitcoin value and we said burning the value difference. I mean, of course, the bond Satoshi. S also is a tiny value in Bitcoin. Let's say it's 1/100 1% of $1 cent, but the PS q.

A

Well, you wanted, or ship would be, maybe whatever 10 cent and to difference between the 10 cent and 1% of a tall ascent is what the money, what was get destroyed, or the economic money and like the mining fee and yeah what the user really paid at the end, this mining, okay, thanks and this I'm coloring. Yet it's a core concept for our economic model because we stepped, we are yeah, we're creating the deep deflation as we are taking out coins of circulation.

A

When there are four million psq initially and you have burned one Satoshi in a in a trade transaction, then there is one sir tor Shilla less available and every psq holder is a little bit more rich because yeah there's a little bit less coincidence of light.

C

Okay I know: I, expect spleen.

A

Hyman yeah.

B

How are you fine.

A

Fine feel free also to jump in. If you want to add anything or if you have any questions, we are doing a Q&A, basically for the Tao yeah.

A

Anything locks what's unclear and.

B

Yeah I'm gonna, listen for now, I. Think I, don't have any questions at the moment.

A

Thank You Lauren and too much I. Do you have any questions.

B

No I'm going to click right clear. Maybe there is something coming up in the next week's, but as of now everything is crystal.

A

Great Thomas do you. I was.

B

Thinking you were mentioning the last Rothko or two before it would be. Maybe a good idea to have does not fall set for bisque and I was wondering which, like what? How hard would it be to do that? That's kind of.

A

Completely, it would be a good idea to what.

B

Have a test not faucet for Beast.

A

Use I'm not sure if it's worth there, but because it's just worth you weeks, hopefully I mean we hopefully can launch soon may not and a sinker. Of course we want to get as many tests as possible yeah if you have any idea how to do it without a lot of effort, feel free I was.

B

Thinking, I'm not sure, so, if I would just have a bitcoind, that's not wallet like it will be enough for it to manage to be asked you if I'm making sure that that's.

A

Quite complicated when you also technically, you could create with any application any softer. You can create transactions which are psq transaction, but you have to follow exactly all the rules and when you make a tiny mistake, it's it's destroyed. We asked you which wouldn't be a big issue now assessment, but I I was a wreck right now.

B

When somebody wants to have tests net, we ask you he has to ask in the slack right, yeah.

A

B

Website do anything besides sending us an email, and then somebody sends yeah.

A

I mean there are more and more people who have already been asked or think it's already quite faster 20 and somebody posts on the on the forum or wines like that. I mean in a few hours at least he gets amperes cue from somebody. I mean get back at Thomas I. Remember you asked also regarding API is at the moment there is no API for that out. Maybe we could edit some at some point- I'm, not very I.

A

Don't I mean basic? Of course it would be good to have alternative implementations and so on other, that's all the theory, but we see in a theorem what can happen week and so on and Bitcoin and Satoshi himself. He said are that it can be very quickly a pain in the ass also for the pistol when we would have now, for instance, when you would work on whatever a pipe conversion for a psq wallet where you implement all the rules and you can handle it, then, with every change.

A

What I do in every optimization I do we have to manage if I'm not breaking anything of you? Your version are that can be some box when I fix this, but to take care yeah. We have to maintain a basically box that it's that, because we cannot guarantee that some uses of of this wallet are yeah working with this or and some exchange would reduced and such software, and we cannot easily update and and change things without a larger coordination.

A

That would just be a pain in the ass at the end and doesn't add at the moment I think value I mean when the Dow is very stable, so I assume, maybe in one year or so, when fully tested and stable I'm more I'd see it less problematic like at the moment at the moment, it would be I wouldn't like to to have deal with this, because would pictures make like not shut up.

B

A

I told you like.

B

When I taught at first I told that that the vsq itself is like an API for the wallet itself, let's make like the faucet really easy. You know yeah.

A

But I mean when you have some music, you, you can send it you have. You always have to take care, how you manage yeah that the outputs like, for instance, with some wallets, where the out goods get get randomly mixed, I, think an electron. That's by default. The case you would already easily burn obvious cubicles well, would be before the PS viewer hood, the PS q would can be, is probably involve it and when it's no.

B

Well, I meant it's like. We have the bsq in bisque right by the wallet itself, so it would be nice to export an API with which you can interact with well yeah. That's what I'm saying so no new! What implementing and just exposing that yeah.

A

I mean the whole API project. This have not met not in the last weeks and I mean when the API is really in place with all the login security and everything it shouldn't be. A big deal basically to to just I mean just to mean two main classes which you have to access from the API and export whatever you need and drop it to some value objects. It's probably not difficult to do it I at the moment that said and yeah when there are I mean, of course, with this.

A

For said, it would be very limited use case, but when there would be an API for for a PSU, we have to take care of the API users, and it's already would introduce already this kind of coordination problem and and risks with update when you fix something or change, something which you're not controlling your environment, anymore, 100% I mean at the moment, I can, when I know that the only users are yeah are using be skewer data over the bisque application. We had all the control there and we can test it.

A

What happens we say old? It's non updated wallet to manage backward compatibility and so on when there would be an API. We have different tests in our is what we have to consider. It's just makes everything much more complex and it's already super complex I mean specially backward comparability, and so it is just a small notice. The address is basically a Bitcoin address. We are just using in the UI, the prefix with the B.

A

So when you want to look up a psq transaction in the normal block Explorer, you just take the address and remove the B at the beginning, and then it's a Bitcoin address. The transactions are normal Bitcoin transactions anyway, and these just to avoid mistakes that they use they sending our Bitcoin to the P, is Q wallet or the other way around, which is West and to help avoid problems.

A

When users are trying to implant their own stuff and play around it's completely on their own yeah, they have to know what to do and really study the implementation and tech. The tech spec is not enough because it's not well maintained at in that time to update the details.

A

So the code is the only real speck at the end, so when somebody would play around with it and create a transaction outside of the psq software he's completely responsible for himself and when psq get burnt, it's his bad luck and it can easily happen as it's quite it's quite complex to do it correctly and I. Don't recommend anybody to do it at the current moment, at least.

C

But too much maybe for for this use case or untested. Maybe it's enough if we have an instant instance running with enough office to to buy p SQ, because you can buy please queue in the the tests that are in the client by using a test net Bitcoin, which you can access I get through a bit confessors. So maybe you can. We can use and could use that and push that that it's that they are office available. So we test also that the trading process more material. It's.

A

Still that the user has to be online, like I, have been at the beginning more online with my notes, but recently not and maybe yeah.

C

There was the data point, so we need a an instance that is available all the time that has enough offers to be taken, I mean.

A

C

One, if no one is sitting in front of the instance and it's it's the same problem again and no matter if it's online or not but yeah. When.

A

The API would be already ready, which the whole trade we could do it basically and even she had access to it. Anybody can click basically on complete completing the trader, but that's not the case and I mean it. It's just now for Yap before until we launch I mean that so and I assume it's not worth the effort and it's for sure not really trivial. So maybe it takes four or five six weeks work and then we're launching anyway, please but I mean if you interested to help us on the outside.

A

There is for sure a lot to do. Maybe it can have a direct conversation and I need to think a little bit more. What can be done exactly had? You are not really a Java developer. Were you more JavaScript and Python guy ever too much John.

B

Iii I do not know John, okay,.

A

Because speed and biscuit everything is each other, and especially though that requires good.

B

A

Complex but there might be some other stuff like with Explorer.

A

Nothing hyper you at the moment, so I don't want to put too much effort from my side, always cues side.

A

But yeah I would think a little bit about it. If there's something else I'm after feeling that foxy is probably not worth the effort and probably takes so much time that when it's ready, we don't need it anyway. I assume to stop you with this. If you want to do it, but yeah I thinking.

B

About the Explorer as well right now, I'm, just playing with you know the setup on rectus and stuff, like that and I saw that the full beast mode is using the block, notify on beat kind and very of thinking about using 0 and Q or.

A

I've never had time to look close into it, how it works and from that Java side and so on, would be interesting, someday someday when it can do the same. But we are I mean we are just getting basically the event when there's a new block with the block notifies with a very simple API, the serum who is much more powerful when I I think it's not really a replacement for our PC from the features. As far as.

B

Yeah, it's not a replacement but I'm using 0 and Q on with my block, explored I'm, double hoping, and it's like really easy. You just expose a TCP socket and you tell to which events you want to subscribe to and they are like free events, new block new transaction I mean for events and I do receive the whole work or job the walk, hash. The whole transaction order. The transaction has these four things you can subscribe to and you will get it and that's okay, so.

A

B

Could have like our nodes having 0 and Q exposed and, like other these nodes could connect to it or not run.

A

B

Own or something like that, yeah.

A

That would be interesting, too I mean so far. We I I mean sometimes I have seen some issues that seed not were not or work got stuck at some block, but I think it's related to reworks that in some real cases, basically trust the Kodi issuing in the Java on the travelpod, but it could be able to that.

A

The block notify maybe at some point doesn't worker so I yeah, but I I, don't know yet I mean it's very important that this is super stable or yeah, that we don't have any risks that are once in a week or once in a month that this notification doesn't work because then the seed not get stuck basically and and misses the block or and I'm not sure if C serum could would be more stable or more risky.

A

It's not I mean it's more complex, so at the end, probably it could add more risk, like the block, not device super basic, but would be interesting topic to investigate and see how it works under from the code side in a small java application as tested out know.

B

I can show it to you, but once if you want to use, it is pretty simple: yeah.

A

I mean I will not have time until we are really stable with the tower everything, because there's so much for yourself and all this stuff what's already work and where we don't have problems which is just optimization well here we don't even know if it's an optimization, it's for me. I cannot donate any time to this.

A

But if you like, you can start make a couple cookies with some would be good to have ten, basically a Java client, where basically, the old implementation is visible and so on, but another thing which might be more more interesting as we would basically like to have a block Explorer, so easy, one-click setup block it flora, where user can install or on on their server instance or local, even and to avoid these privacy issues.

A

When you click on the block girl, many rockets grow, companies are doing business with trainer like openness, at least it can be assumed because most don't have any business model out there on money. So that's probably the way how they own money to sell the data and attend.

A

You know your IP or clicking ten times on different transactions, so they know you are probably interest in the close transaction to probably either your addresses or you have received something from those addresses or so, and we are using block stream blockage by default and we assume and hope that block stream is not doing shady business, but even that it would be when the user can basically just install their own block Explorer very simple at them.

A

Of course, they are already projects where you can do it, but for non developers, it's probably too difficult, so it yeah, that's basically really a one-click script run and then it installs everything and you have your local block Explorer and that you can use them as default block is prefer bisque, so you're not losing privacy anymore on this side, and then it would be good to maybe combine or have a clone of this block Explorer, which is dedicated for aqueous queue.

A

But the best way would be when this block Explorer and the sense bsq as well and shows the full transaction, but can tech yeah transactions at what they are like? The compensation request transactions on and if the outputs are psq or Bitcoin and from the UI and from usability I, find their block spring block Explorer as one of the best. What I've seen so far.

A

So if you're interested in this project that would be would be cool. It's nothing high preview at the moment, but I think it would be really nice to head and I mean the current clock. Explorer works basically and think it has a few smaller issues, but basically it works, but it's not very intuitive to use, and it would be good to have really the fuller transaction in an intuitive way like the block spring block. Explorer were plotting for I. Think has very good.

A

You are UX, but I assumed they are for sure on the shady business side. I, wouldn't trust them to not sell the date them, but a block Explorer like this when it's it yeah my opinion. Many block Explorer are not very into a defend when you really want to understand what was going on is transaction. It's a little bit complicated to to follow.

A

Duty, let's jump back to the to the towel stuff or a Christopher I think you wanted to say something. Sorry I interrupted you.

C

Know it's nothing really directly related to the tower, as it's somehow related it's about the release that we we were planning for for this week. So at the moment, after feeling, it's kind of unrealistic to do a release tomorrow, because we have a couple of pull requests that we want to get in like the the restore of the wallet for Windows yeah, the UI stuff that were Peter just did.

C

The redesign I would be able to finish that off, but I think he is still also working on the website, which would be also available when this release goes out. So what about that? We are trying to have a code freeze and until tomorrow of business and then release on Monday or Tuesday and try to testing, because yeah I think it doesn't make sense to push super hot or tomorrow, because yeah you know it should be. It should be a release where we communicate already in the in the minute client that the Dow is available.

C

So we do expect people more people to test tested pistol on test net and for that I guess everything that we can can push in. That makes it more stable and better to use. It would be worse to to wait a couple of days. What's your take, no.

A

No I completely agree and the trade limit change shows a little bit more complex as I expected. It requires maybe another half a day or whatever, and to get not like to take the time. But we need and maybe try to make a code freeze or tomorrow so sounds good.

C

Yeah yeah, so, besides that nothing spaghetti maxi.

A

It'll be good when we get a little bit.

A

What has support because usually you're nearly down and Devon are the only guys who are really testing and I try to do it all so that last series also couldn't help with testing, and it would be good when, basically, all developers are donating half a day or so for helping testing, because everybody has different approach: how to use the software it tests a positive side effect that you learn more about the software from different angles and get more used and then find more or stuff what you could improve in your domain or whatever so I think it would be good it's because it's not the funniest work for most developers, at least for me, I, don't like testing really but I.

A

Think it's important to and I think this release is still is again carrying quite a lot of changes and a few changes are a little bit more risky. So we we need to have a proper test cycle to not risk anything and yeah. It would be cool if we get a product test. Community I, don't know which, which tool we want to use I, think the tool down like who didn't really work well was more.

C

Unfortunately, and wasn't able to do the the follow-up call with Devon I promise to do it, but yeah just forgot to be honest, yeah, so the tool it worked. Okay, the problem I had is I, had performance issues with the user interface, so it froze a lot and consumed lots of a resources. So it was kind of hard to use for me yeah, maybe maybe there is, and let's say, reduced interface and user interface for our our testing purpose, but yeah I guess Devon has a better idea.

C

If there are any tools out there that we could reuse yeah.

A

Maybe check out with Devon and I will I will.

C

Try to pin Devon and then we should should be able to tell on the first day of Friday which, which truly one want to use for this next cycle. We.

A

Cannot go back to the spreadsheet that was, and this year at the end, I mean I'm, not the super fan of this overhead, where you already need a half an hour to get registered and understand the tool and whatever, and then it was painful with getting yeah. It was an invite-only or whatever, so that was kind of like I firm. It was a bad experience with this tool. Rather I didn't use it, and also, but anyway, I leave it to you and Devon, but to use there and yeah.

A

It would be especially good when every developer, when they're, when they're testing their code change work, they have worked on like Florin everything which is related to tor and network stuff, or for oscar everything related to Bitcoin, with the seed node, for instance. It is this word release or to focus mainly on other, especially on this, because that yeah, that's what you have done. You know the best work could be problems, how to test the best and then make a general test to make a few trades, especially also backward.

A

Copper ability is important to test with one instance from a from the latest release: either run it from the binary or you have as I, usually do it with two different IntelliJ instances: we're gonna have the latest release, checked out and stopped from there. One note and then I have another intelligence which the release candidate and then to trade in between and and see if it works in, both direction can be that other.

A

It has quite a few complications when you do a trade, it has slightly difference if you buy or if you're the maker, if you're, the buyer and the maker or a filter, the buyer and the taker or on the other side, if you're, the seller is the maker or the sell s to take a so basically they're four rolls which are yeah have different implementation. I mean I.

A

Think in this release we don't have much changes in the trade area so beside the payment, the trade limit that has to be tested well, but just to give a little bit of account from a context about the testing.

A

So it's not enough to just make one trade where you were Alice's the buyer and then focus it if the seller is makin publicist, LS taker, you also have to do it in the other way around and another use cases tend with arbitration to make some disputes and check if there are disputes or esta pay artists working from the dispute and everything. So those are the main use cases which have to work all always with every release the smoke tests and then, of course, the current. The concrete changes.

A

Yes, anything else anybody want to discuss or talk about, I mean can be also something known. Tal related not only can talk whatever.

B

A

B

Some now related stuff. The question is: what do we want to monitor for the tower? Maybe we should hear it. Yeah.

A

As I have kind of private monitoring node at the moment, which I use on before myself and SQ, maybe use it for checking the seed nodes, as we have at the moment, I think six thousand eight seed nodes and it seems that they are stable. Now, I, yeah I said what I request from the seed note and what I monitor are the normal PPO Network data like in the normal monitoring and as a number of authors and everything and the number of blocks I, don't remember exactly thing from the Genesis or yeah.

A

So I make a request. Give me all your blocks and then I see how many blocks every node nest. So at one seed, not good, have a problem that it got stuck at one block and don't get updated to the next block anymore. Then you will see I once hit on this behind and you have to check what happened there.

A

Beside that I wanted to implement something like chain of hashes, where the complete dowel state. Also when a passing of a block is complete, then we are creating a hash of this state. At this block of the of the dowel state- and we are connecting it in the chain like a block chain at the end, all this hashes and with the with the block number or the block cache, and to have such a data structure, and then we have Wistar hash.

A

We have basically one or one data item which represents the whole state and when hashes are not exactly the same between different nodes, we see that something is out to sync, and there are us issues. I didn't have time. We had to implement this and I'm not wasn't sure if I find time before the release to do it. But it's a consensus.

A

It's basically kind like a monitoring system, so it's not really part of consensus, but it will help us for normal bisque clients and for the feed nodes and for the monitoring of the seed nodes to detect immediately when there is some inconsistency.

A

Some inconsistency are expected like in real cases when, when you get a different block from different chains, you are a huntress, pure, very likely, inconsistent, actually, not very likely, because only when you have different transactions when the transactions are the same in both blocks, then you'll probably have to sync data, but it can be, and, yes, sorry, very important data. The peer-to-peer network pillow data like the blind roads and competition area, the proposal data, those are those have cost in the last races.

A

Many issues that some people didn't have to say turns on and then vote result were invalid. So that's very important to have its peer review network data. So there's never 100% consistency. Is it it's? Okay, when some notes have less data so, but the seed node should have basically always the same data as we are republishing the state, the very, with some very high resilience, so a node which is always online should have basically always the same data. Otherwise we have to find out what's the reason, as those are very important for that further monitoring.

A

Somebody want to say something Florine to you.

B

hmm So we do the integration to monitoring after the release.

A

Yeah, it would be at chiming. We would need a monitor for the test net, because the state is not available a minute and will not help when you're expecting empty data and yeah. We would need basically a second version of your monitoring environment for testing it, but it would be good to have this before the main net launcher right in place. So maybe you could put this actually a high priority so that we get it ready before as soon as possible.

A

Basically, but before we are going live because it's very it's very important to spot very quickly when there are some problems that we can find it quickly and restart to see those and fix that they share whatever by.

B

Going laughs, you do not mean that release we just discussed next week or.

A

Don't mean that I mean that, or maybe I mean my target is meter of March, but probably would be more end of March. We'll see how, as soon as we are 100% stable on test net, that we don't find any issues anymore, that's the main requirement and assume we have older education and the documentation and communication material already, and that our users that we have reached our users when we go live that they understand the tower and know about the tower.

A

That's a requirement to, but basically when everything goes good in the best case, I think it can be end of March.

C

Ok, one thing: one thing: Manfred use cases like the unlocking of bonds. Where do we want to monitor that? For example, if if there is a bond for a role locked and someone just wants to get out and wants to get out, because he did something wrong and we want to confiscate at the bond, how do we get notified at that? Or or do we just monitor it? Manually? I.

A

Would say we at the moment: I mean the whole bonding and all the advanced features besides the pure voting and computation requests. I would keep a little bit on low priority from the usability and I mean they need to be edit. Better usability like that, you know exactly which form to confiscate and so on, but I mean especially the confiscation is very exceptional.

A

It probably will never happen at all, so I would not put too much emphasis on this, just as basically works on the consensus part, because everything which is on the in the consensus part cannot be changed without breaking the towel, but everything which is on the usability part and how we are connect. The state with with the external data on github, that's important as we need to have some some backups, for instance, for all the computation requests.

A

It would be a suit therein discussed this already somewhere around get up and I think it would be good to have and I think there are systems where you can monitor some webpage and with every change of the webpage. You are creating a new version in yeah git commit and you are yeah. You get the get history of the changes of this webpage, so basically for every compensation request.

A

We could monitor a whole raper there and every change get get a commit, and basically everything is in a history, and when this is a alternative system, besides github, then we are not depending on github as well, and it would be blocked from us and lock up all the data, whatever that this historical date is still available and also that we can verify if there was a change, for instance, when a user in the last minute change the amount what he requested, and it would be kinda like a little bit a social cheating that people review it and say, ok and then, due to the voting, without checking it again at the last moment, and at the last moment he has changed it to a different value or whatever I.

A

Don't expect that I mean such stuff will be handled basically, on the social level. Such a contributor will never receive any any. We probably never get accepted anymore when he tried to cheat the other stakeholders. I think, that's all not really greatly. You guys think everything. What we really need to focus is every everything which is on the consensus area and on the stability area and so far we had mainly issues on the stability area and I mean this box, which the book resolved were not hundred-percent. Consensus talks, but somewhere in between and.

A

C

No, what I meant was the unlocking of bonds. I didn't meant using the face, how to unlock a bond, and this kind of stuff I meant a use case. For example, we have a DNS maintainer, he the maintainer has to lock I, don't know 50,000 bsq. He changes the DNS entry points to a scam site to scam people, and then he tries to UM to get his esq bond out in time before we recognize the the scam attempt- and my question was if how we monitor the unlock transactions of the bonds, so that someone.

A

Christ I think the UI you would treat when something get unlocked and I mean such stuff should be discovered very quickly and the unlock time will be long enough. Something like minimum two months, so maybe three or four months.

A

Eric and I I'm not sure we really will use the bombs initially, because it's not really hardly required, and it adds a little bit of risk when there is a park and so on. I mean I, don't know yet it for me. Not the hyper-v I mean important stuff would need to be, but we to work is the is the computation requests are trading fee payment and the trading of b sqs.

A

Those three RS I think other the minimum, what a minimum viable version, but we really need to get hundred percent safe and stable and all the rest. We can use a little bit later when we have a little bit more time for testing and feel more confident, confident with it. Yeah.

C

A

I just left everything in yeah for that people learn about it and for feedback and for testing and everything I'm, not 100% sure. If I was considering to deactivate all the advanced features, I'm, not sure if that is easy and would introduce even more risk. Probably the easiest is just and we don't use it, and there will be anyway. The policy when our because of packs some peace, Hume get burned like yeah, you make a bond and then there is a button.

A

Suddenly a peace cube at burned, then you can make a reimbursement request and you get it refunded. Basically, for the network, it doesn't matter when some psq got burned and then got issued its tourism and when there is a clear case that it was not your mistake or whatever or your pet intention. When it's a software back then yeah, we will fix it like that.

A

Yeah one thing: what I want to explain a little bit more, which maybe is not completely trivial to understand- is how the peer-to-peer network data and Bitcoin transactions are working together, because that's a core concept of the Dow, which makes it very different to smart contract platforms like etherium as we are using the blockchain.

A

Basically as notary system where we are where we are time, stamping are important events and data, and we are connecting this data which lives in the peer-to-peer network. As really we are using our peer-to-peer network yeah peer-to-peer network as a second layer network where we are distributing all the rich data. The network, payload data, like the blind vote or the compensation request, and to link both together as in the peer-to-peer network. We don't have the guarantee that every node receives the same data. It's eventually consistency and we cannot yeah.

A

We don't have a guarantee for that. On the Bitcoin blockchain, we have the guarantee when a block is at least ten blocks deeper, so we can be sure it will not be changed. It's there forever and it it's yeah.

A

It's it's frozen there and everybody can get this blockchain data and everybody we can assume as we we can guarantee that everybody has the same data and how to combine both is over that the hashed of the pain owed, payload Network payload, get added in the OP return of the Bitcoin transaction and the Bitcoin transaction ID get added to the peer-to-peer network data. Of course you cannot do it at the same time, because you would have a chicken-egg problem. You cannot create the hash or without the transaction.

A

Id is not part of the hash, because otherwise, here you cannot create the hash and then put it in the transaction and it changed the transaction ID. When you would change there up return, so it wouldn't work to the transaction, a ticket and it afterwards that not included in the hash. But that's not an issue, and with that we have the security of the Bitcoin blockchain enter the consistency, the guarantee that the data are there and are the rich data we don't want to spend the blockchain with encoding data in the blockchain.

A

The rich data are living on the peer-to-peer network and every user is receiving this anyway over the peer-to-peer network and yeah.

A

That's the way how this plays together- and there is one thing which is a little bit complex but maybe I try to explain it, don't be too worried when you don't understand it, it's really a little bit tricky, but it's very important concept, which was one of the reasons of the last voting round balers, because we cannot guarantee that all the pure peer network data has arrived at everybody and you cannot decrypt or you cannot. You don't know what somebody has voted without the peer-to-peer network paper, so you will get.

A

Everybody will get all their transaction data, but for decrypting the blind vote. You need to peer to peer network data when you have not received this or when maybe a malicious peer has not published it to other peers or has only publish it to a colluding small group of peers.

A

You would run into different vote results, because some peers can decrypt the state and will include this road to the total result, and that might be a different in the end. Result may be a competition regrets, get accepted or not get accepted because of this one vote, and you have to ensure that everybody has the same fate of you. Are everybody at the revolta at the vote. Reveal is using the peer-to-peer network data what they have at this moment for this space for this cycle.

A

So, let's say: I received ten different blind Road data from the network and Christophe received only nine. So when we would at the end create the vote result, we would have basically a potentially different Road result and would be inconsistent. So we have to be sure that either nine or ten Easter, it's the truth and yeah it doesn't actually matter. It just have to be clear that there's only one well, it's data set and to determine which data set is the winning. If the, if mine, ten, has it.

A

My list of ten blind votes is the valid data set or nine from crystals. We are using a kind like a voting by stake so I when I bought it in my blind wrote with ten thousand psq and crystal coated with 5000 sq, then I have a higher stake and my date, the few would be the winner. Of course, it's not only with you. There are many others and the some let's say that 80% of the network has ten.

A

They talk, checked and 20% have only nine and then the some of the stake which were used by those who have ten get summed up, and when this is the majority, then this is the the winner and then there for a vote result.

A

You have to get all ten all ten blind vote, European Network data to create the wrote result when you don't have it, you request it from the network and when you don't get it for whatever reason, you cannot create the boat result, because you would potentially come to a different result, and that has to be avoided and by yeah, but this way guarantee basically from the Bitcoin side, because this data is in the in the Bitcoin transaction as their the stake.

A

It's it's the input for their client for the reveal transaction and their work reveal transactions. Are. There is no peer-to-peer network data for this, so that's purely in the blockchain. So from there we have basically our consistent voting on baat Esther.

A

What is the yeah, the maturity date of you, and that is the one which will be used for the result, and that has to be to avoid some attack scenario we require that it has 80% of the network has to be in the majority and that we have this issue because of some bugs that are that the most just a minority or that well that the majority of the state of you didn't reach 80% of the total state.

A

It was maybe just sixty percent or even less, and then the whole voting cycle is invalid and that's basically a count like an emergency case when we have like we had network issues and buts and then the whole walking round it's rendered invalid and basically it's no change in anything, and we just have to repeat it's not a big issue at the end, because a compensation request can be just repeated.

A

It's a little bit annoying, of course, but it's at least better like to have inconsistency or possible attacks that somebody with a minority could abuse this to attack the tower and gettin some compensation request sex accepted by withholding some patron. Whatever we had a text in our is I'm, not sure if I could explain it good, it's quite complicated, and so you can read it up in the document.

A

If you want to know again adjustment, because it was now relevant in the last in results and it's important concept to to guarantee that we have consistency from the European Network theater. What about result.

A

Did I manage to confuse everybody completely.

A

B

A

Yeah, is it more or less? Is it's very so I'm not expect that you fully understand it, because it's really quite complicated to maybe a draw it down, to really understand perception and to have it internalized this process with the blind vote and about reveal and which data and the transaction is on. It's not trivial, but.

B

A

B

Is that we don't know which data exists for people in the biskits p2p network, because, unlike Bitcoin, nothing Varys promised? And we want to know what the majority has? And my joy of the here is 80%, so use.

C

B

Chain for voting on what the majority has no chance before was voting.

A

Yes, and here we use only the state because the merit that was one of the parts also the merit- is already part of the PPO network better and we cannot rely on any peer-to-peer here. The pure stake is used so when you and that's a little bit something important- which is a little bit complicated to communicate to the user- that when you're a contribute to this earned a lot of of merit over the time, so you have maybe whatever one or two hundred bsq already its married.

A

You don't need to use a lot of stake for voting because you're married it gives you already a lot of weight, but what is maturity has only the stake is used. So when you would decide to vote only weighs 100 PS q, you might end up in the minority date of you and then you're. Actually, your vote can be still valid. You just need to request the data so it and it's not a big issue, but it reduces a little bit there.

A

Security from as I preferred I would have preferred to have to marry, but in there as well, because we can assume that the people who have earned a lot of merit are not trying to cheat the system when somebody by a million psq on the market. We don't know. Maybe it's a attacker from a state agency- would try to whatever do any, but.

A

Anyway, I think it's it's a more exotic problem. The real problems are probably stability and and box. After box of consensus box.

C

um One question: one fact: do we see it is a problem if the Bitcoin network of the mining fees go up a lot so that people might be reluctant and not to vote or to do anything, you know it freezes them. Does the state I think.

A

It's I mean there are fees to anyway, bsq and those people I mean voting is a management task for the contributors. Mainly, we don't want that any users participating in voting and has no clue what he's working on. So it requires work at the end to read the requests and know what to vote. I mean there might be some popular votes like yeah. Whatever kick out there is can be out.

A

We know so where everybody has an opinion without a lot of research, but usually, if a competition requests important to to know and to understand what was going on and I think for contributors. It doesn't matter if you, if it cost you $50 in voting when you earn $5,000 you basically when it would be really expensive.

A

You're just editor to your earning that you want to earn a little bit more because you know it costs you $50 $100 and for doing the requests and the voting I, don't I mean when bid can become when Bitcoin mining fee become really over long term. Very high, like $50, like we had or a $40 a two years ago in December for a few weeks, I think the best trade protocol on chain we the main problem, because people maybe will not trade anymore when the current trade protocol with for transaction will become too expensive.

A

So I think that's the main problem and they're from the tower I. Don't really see a big issue there and it could be I mean the Dow is not pilfered to last 400 years or whatever, or this current version based on Bitcoin, because I expect that at some point the money will be very high. Maybe not $50 I think something like $20 he's, probably realistic, like Swift I.

A

Think it's similar like Swift, also sweet transaction doesn't cost hundred dollar, but it costs you probably 20 or $50 and- and that might be already long term too expensive at some point and then we have to see in a few years. Maybe there are some kind of side chains or some alternative currencies where we can use. Maybe we get some privacy improvement like using something like manera or so, and to move over to a different blockchain would be such a big problem because we have this way of burning and reimbursement.

A

So when we do would move now to litecoin, which is technically nearly the same, like Bitcoin, we could easily. Basically you have to prove that you have burns and psq in the Bitcoin blockchain, and with this proof you have the right to issue or to get reimbursed on on the new chain that could be done manually.

A

Basically, as with this tool, what we have for a proof of burn, you can prove it that you have burned it and you could sign something and then stakeholders could manually basically verify it could be automated a question also so I think conceptually not a really big issue to have account like a swap by burning on one chain and issuing on the other side, and then you empty out the old chain and you're filling out the new train, and there might be some price fluctuations and some weirdness a little bit in this time when the supply of chain is, we are falling down extremely, but I.

A

Think beside that technically I shouldn't be big issues.

C

That make sense one last question regarding the the Peter Piot data that we are using was the in the beginning of the conceptual phase anytime to discussion to not do it not store it in the beauty, P network itself that is used for trading as well, but I use stuff like ipfs, or something a different.

A

C

But it's kind of data was.

A

The one thing if I started actually with doing the voting on the blockchain with yeah I, was a bitmap where you basically one bit, is for VOC it okay and not voted at 8, then a second beta score. If you're voted or not voted at all and it it would have avoided this problem. With connecting the European network data with the blockchain, but once I seen yeah, we would have filled up top return to the maximum and it would have been limited. It would have only I think something like 20 or 30 requests.

A

You can vote down with this. So when we have 50 or 100 requests yeah, you cannot vote on all you would have need to make several transactions, which would have complicate the whole protest and you were yeah you. You are not flexible at all. When you do anything on the blockchain, it's yeah, it's super limited. What you can do you have compressed everything as as far as possible and any change or whatever yeah.

A

You cannot do usually so actually I'm happy that we moved away from this and regarding ipfs I was when I moved away from the first PhDs, which I was using at the beginning, which never was stable on the on the real internet. I looked in 2000 at Eve's, like ipfs and big message as well and I had the impression that ipfs is not built for a use case, but we need it more. Like BitTorrent for big files, permanent storage, we are usually don't want to have permanent storage for most data.

A

It's only the tracer which are really stored permanently and it was lacking. Many features what we needed for for validating, that you can prove that you're, the owner of the data it has all the THD based peer-to-peer network systems, have serious Sybil and Eclipse attacks in problems. So you could partition the network easier with a flood-fill network like Bitcoin itself or, like our peer network, you're much more resilient. It's much harder to attack such a network I mean it would not say it's much harder because I mean you can t toss it and so on.

A

But at least here you cannot it's not true more resilient regarding sibilant Eclipse tedx, maybe.

B

I think about ipfs ipfs is a good tool to save content that you already know. What is it? What what isn't, because it takes a contact and give it a name? That's not our case. We trying to decide what is the content, so I wouldn't know of the top of my head. How would I PFS help us and once we decide on a blog, we can start an idea first, but we can also store it on a server I mean once the decision is made, it's not very important whether the data actually stopped yeah.

A

And we have some temporary dates like when you make a proposal, you can remove the proposal as well with ipfs. Basically, probably you can also remove it, but it's more yeah, it's more that you're putting some data with the hash in the network and there it sits and we we we want to remove it as after their temporary proposal payload in this time, in the proposal phase which can be removed.

A

That has a time to live for one months or two months and after this gets removed from the network, because in their blind walk phase it can persist then, and then it released get stuck, but the first object. Yeah we talk, we don't need so.

B

I think this is nothing beautiful, don't worry.

B

Beautiful so you kind of I mean this. This student worry was not in YouTube Allah, so what you have a favorite, you put them starting in the States forever, but this is immutability and this does not. This is not a BFS, because stuff can be removed. If you stop hosted it basically, yeah.

A

Yeah and I think it was also mainly I, think it runs over over UDP, which doesn't work with tor I, think. Theoretically, they support transport dock Gnostic. So you could run it also over TCP, but in practice, when you try to run a DHT system over TCP, you get very bad performance, because YouTube is much faster and they're doing a lot of management messages which should be super fast and yeah.

A

When you move everything to TCP, assume that you get performance issues, I'm, not sure, if it's an inter if thing to opera opera, peel it if it yeah, if you, if other nodes which are not configured to use TCP, would work, probably can then only work with nodes which you are trying to spawn and basically the same work.

A

We have now with the bisque user, had only notes in the Pisgah network, so yeah at the end, I'm pretty happy that we build on peer-to-peer network, because it's also very important part is that we are not now super flexy when we need a new feature and all the time I needed. New features like now for the Tau I just coded it up, and and that's it on the center system, with the old peer-to-peer network with this DHT, which were very well designed and everything a good code.

A

But it was a huge big project and I never managed to get to the level to really fix things and extended, or also this similar, like we speed country, it's basically half a year worker, so to get really so deep into the into this library that you're, something like a quarantine I can can change and fix stuff and, of course, with peer-to-peer network. It sells it's the top of floor and at the moment, to get deep into this, and it's also takes some time. But for me it was yeah.

A

At the end, I mean when I would have been an alpha TIFF which would have fit exactly to our use. Cases. I would have used it, but it was mainly the lack of ultinice. That's why I built myself yeah.

C

Sites thanks likes plenty.

B

So maybe one thing about it: if there would ever be some kind of of good decentralized storage, this could mean we could use it, but right now IP. This is not exactly this centralized, though you can just put stuff and forget about that. This is not how it works and all the stuff it's supposed to be like that, like a swamp and see how they just don't work very well yet so it's not an authority if it's a maybe and.

A

I'm you're a little bit more familiar with the etherium space, because I was wandering in this time. Also, you had to look into swarm and whisper I think it was a project but I'm not sure if they ever got really implemented or what's the state with this systems, they wanted to build like the area network for different use.

B

Cases whispers just a message sending system like the thing with you with reuse. We talk, so you can use the friend p2p to send messages between notes. Let's whisper, it's nothing about storage Swamp is in development. It's not ready yet.

A

B

Even close to be mature yet and.

A

How does whisper till wish when they use, is not online? So when I want to send you a trade message and you another online storage, yeah.

B

I don't know I mean some want to do that, but I don't know because I don't think I think that whisper was a concept in the beginning and I'm, not sure I, don't know. If anybody did he use it right now. So.

A

It was kind of an idea for the beginning.

B

Nobody had a p2p Network, let's also you'd like bit message: I, think it wasn't used and then it was kind of left behind. I swam, however, is developed, but it's just not mature. Yet.

A

Yeah a lot of of great concepts and white papers and and great presentations, but at the end not running that's a little bit typical I. Think now.

B

A

Should be coming up this too much bad mood between the Bitcoin and the theorem camp, it attend a little bit stupid. So yeah is anything else. Anybody want to discuss.

A

C

B

A

Rob up I, I, don't know what what we should discuss next week is anything from your side. Would what you would like to bring in I.

A

Mean the trade protocol is one of the main priorities, but depends on crack. How nah man have you got from something from him? He's.

B

Gonna see a doctor tomorrow and you are going to be smaller afterwards.

A

Do you think the tech spec is finished and of the end of the week, so that also that I hear was that everybody is reading the tech spec and has a good idea about the new trade protocol and basically reviewing it, was really critical review and help to find issues. If there are any issues by.

B

Monday, it should be ready. It's okay, a.

A

Boot workers, when manager already and when Craig is good until Wednesday. Maybe we could try to plan it for next week, otherwise we can do it in two weeks when it's don't want to stress so much with this yeah.

B

I mean I, let's see tomorrow, I know.

A

B

A

Yeah, let's keep it open when I mean the other topic. Is the unit test frameworks sink floor and you said that you could prepare a little bit of overview and basic for discussion.

B

Yes, I will always part a proposal. I think when the monitor release is out yeah.

A

So it's also nothing urgent or Christophe is from your side anything. Maybe we can. We could also about the you I thought. Maybe that's also would be good that everybody, because there many areas like especially for crack, he will touch a lot of the UI stuff. So it would be good to get a little bit of that idea about the concept and frameworks and especially with the new UI framework. What were using this article to.

C

Chi Phoenix today Phoenix Phoenix yeah, maybe the one after because if you want to do the release on Monday Tuesday mature, if I have, if you have sufficient time to make a proper presentation on that yeah.

A

Yeah, let's keep it open, I mean we can all just keep it next week or so, when nothing is really ready or just have her myself typical again then just discuss whatever it's up for discussion. I.

C

Could talk about the release process yeah.

A

Yeah, that's a good point and a little bit, maybe with testing and the release, and and because there were some discussions regarding snap packaging. I have not really follow up and don't know much about this, and it's probably interesting your infrastructure and so on, but I think it's important to really understand. There are security issues with releases I mean when somebody would manage to introduce malware either on the hosting side or in their release process.

A

It would be terrible and I'm not sure if that's really well understood by every developer, to the same extent, at least, and basically we would like to get. Maybe that's something else for too much I'm, not sure it's a little bit, maybe too much Java related, but we would like to get a deterministic build. I think the chart phone is already deterministic. At least we are creating a hash.

A

We are not really checking it, but we are uploading it also, so everybody could run it and could create the Java and it should same, but then to create the binary. It's not trivial because, especially on six, it has. The signature, also I'm, not sure how this can work, but we should look into this more as when we have somebody who is very experienced with build systems and such stuff and look into Gideon build how it's done on Bitcoin. They have the same issue reassigning the binary.

A

So then there will be some solution for this and, as the goal is basically to have a deterministic build and everybody could build it locally and get to the same agent that can verify that what we've uploaded is this the same binary and and with that we would add a lot of security for ya.

A

Even if the infrastructure would get compromised, we could even have done somewhere extra can't like a monitor server, which is doing the build and verify that the software is always the same which is under on the download patrons on but yeah. Let's, let's make this a topic for for next week, release on testing and and security issues. Yeah, and maybe maybe we find somebody who want to to start working on this.

A

Maybe I will try to invite no Qi he's very good with this and I think he was interested also in that area and think that was there was a path contributor who is not much active anymore, but maybe I could try to reach out to him. I think he was starting also to look into this Thomas. Also, if you are experienced with this area or to look into pity and build also, would be your welcome as well. Of course, it just I think it requires quite a little bit of Chavan chilean knowledge.

A

How we are, how and package sure to it yet very related how to I know it will be built in our system and yeah I. Don't know if you have tender expertise for this, but I.

B

Was funny looking to the DTM build of Bitcoin, we will see I.

A

Mean I think an important part is to find the conceptual solution like for the code-signing, the windows we don't advocate running yet, but that's also on our to-do list. Basically to get the code signing for Windows on Linux, it doesn't exist, but on OS, X and Windows we would have the problem that yeah, you have the signature from the developer and you cannot create the signature only Kitty and windows.

A

So I don't know exactly how this can work, but Bitcoin has found a solution for this and or maybe it's just that you exclude some some files and ignore basically this data from creating the hash Reza but yeah if you're, already familiar with Gideon build. That would be interested to find out how we could what we, what is needed, how we could apply this and long all to get someday to a deterministic, build.

A

There is another thing: yah-tchi null I had set up the hosting and everything basically for a rocket jet for an alternative for slack and I. Don't know exactly it got stuck Stan I think it was mainly that there was still no mail server or the invites for for the registration, part and I'm, not sure if I'll like Alexei or if Steve were the guy who were trying to work on that part.

A

I think Steve is very busy now, with all the communication part, so so much if fewer kind of experience with server setup and I think it was basically to get the mail server or some system which allows you allows us to to handle the email registration part for for getting into rocket jet from shinola have not heard since a while I can try to contact him, but I assume is very busy wishes day job, but it's a little bit a pity to really three four months idle and not used and would be good to move over to that at some point.

A

But I will not have time to to work on to attend much on this, but basically would be good when somebody takes the lead for this and that get this. This transition done, and there are a few yada dis, webhooks with github and so on. So there are should be basically, everything which work zones like now would work with rocket jet and maybe for Florence and relevant also for the notification system. Ways like that this will work with rocket jet as well.

A

I think it will work I think it's basically a copy of the features of slack, but it has to be checked out and test it before we are moving.

A

Yeah, let's rock Upton, already quite long and I, will update there. I will create a new issue for next week with these topics and let's target maybe the under trade protocol for two weeks that we handle it in two weeks and the UI. Maybe then the week afterwards, yeah okay, then thanks everybody for watching and being in the call and see you next week.

A

B