Hyperledger Foundation Fabric, 21 Sep 2021

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: An Introduction to FABLO: a tool to generate a Hyperledger Fabric blockchain network

Description

A recording of the Media and Entertainment Special Interest Group call on September 21, 2021 featuring a discussion of FABLO, a simple tool to generate the Hyperledger Fabric blockchain network and run it on Docker. It supports RAFT and solo consensus protocols, multiple organizations and channels, chaincode installation and upgrade.

For more about FABLO, see here: https://github.com/softwaremill/fablo

For more about the Media and Entertainment Special Interest Group, see: https://wiki.hyperledger.org/display/MESIG/

----
Hyperledger – Open Source Blockchain Technologies

Hyperledger is an open source community focused on developing a suite of stable frameworks, tools and libraries for enterprise-grade blockchain deployments.
It serves as a neutral home for various distributed ledger frameworks including Hyperledger Besu, Fabric, Sawtooth, Indy, as well as tools like Hyperledger Avalon, Cactus and libraries like Hyperledger Aries, Ursa. Learn more about Hyperledger projects: https://www.hyperledger.org/use

Case Studies: https://www.hyperledger.org/learn/case-studies
Training & Certification: https://www.hyperledger.org/learn/training
Tutorials: https://www.hyperledger.org/use/tutorials
Webinars: https://www.hyperledger.org/learn/webinars
Events: https://www.hyperledger.org/events
Vendor Directory: https://www.hyperledger.org/use/vendor-directory

Subscribe to the Hyperledger Newsletter: https://www.hyperledger.org/newsletter

Follow-us on Twitter @Hyperledger

Learn about Hyperledger Membership: https://www.hyperledger.org/about/join

#Hyperledger #HyperledgerMedia&EntertaimentSIG #Blockchain

A

Okay, you're alive.

B

Hi to everybody on another tuesday welcome once again both to familiar faces and newcomers. We are the media and entertainment special interest group here at hyperledger and we're very lucky today to have an international guest specifically from poland.

B

This is um mr piotr jawofsky, who is on a larger scale, part of the software mill organization again in poland, but more specifically, is the guiding light behind the new fablo initiative. Fablo projects which you can find on github as well as in pyota, has more official material, so I'll post that link in just a second.

B

If you want to drill down into the engineering of what he'll be discussing and what he will be discussing is fablo, which- and here I quote directly from the repo- it is a simple tool to generate the hyperledger fabric blockchain, uh to generate the network and then run it on docker.

B

It supports raft and solo consensus protocols, multiple organizations and channels chain code, installation and upgrades that initial blurb is what caught my attention. So I wanted to come and explain to us in greater detail how this all works. So it's over to you.

B

A

I'm gonna share my screen and we'll be start talking. Okay, everything.

B

A

I assume that you are that everything is fine, so I'm.

B

Gonna start yep. Everything looks great, okay,.

A

Great uh so awesome thanks for introduction uh as it was set. I am piotr and I'm working in software mill and we're gonna talk about fablo today and.

B

A

It's our open source project and also a part of make hypologist tied up easier trend, and it was previously named fabrica but got rebranded to fable recently, but it's just same thing. So if you got already interested in it and using it, uh you can start call start naming fablo and use the new repository that is fablo and we're gonna talk today about on why it was created, review its features. Do some demos, maybe talk about future plans, answer some questions and stuff like this?

A

Basically, so, first of all why it was created and to answer that question uh first, we must talk about software and it's not about bragging ourselves, uh but only to get the context, uh because we do a lot of cool stuff like like scala, big data, machine learning, apache kafka blockchains, but one factor is in common between all our projects.

A

That is that we are not a product company, and that means that we help in building custom solutions products. So often we work in models such like team, extensions, consulting for the clients, and all of this basically means troubles.

A

It means troubles because we like don't have any single network setup in some cloud or in some external cluster. We don't have stable configuration, but rather we have multiple projects and every other projects. It's kind of.

B

A

And we must propose a different solution like in case of supply chains, many organizations many channels in healthcare systems. Maybe two or three organizations will be enough with common channel and also other use cases.

A

So, basically, our crucial needs are being able to experiment with network setup, so we could easily move from one configuration to another one uh to quickly experiment and also really crucial. Part uh is limited time factor because often we must work in kind of a short time frame to present something working to the client, and all of this sounds nice, but in case of hypo ledger, it's really bloody complicated. Actually because don't get me wrong.

A

Fabric is really great, though, but hard to get and understand in the beginning, because it's got a lot of building blocks.

A

Every organization have certificate authorities, peers, we've also got order organizations orders, we've got channels to combine organizations chain codes that can be installed on the channels and policies to ensure security around the blockchain and everything is nice.

A

But everything is really complicated here and that's great because, thanks for through thanks to these uh building blocks, we can cover a lot of use cases and a lot of scenarios, but it's really hard to get at the beginning and we had really hard problems, understanding fabric at the beginning and I'm pretty sure that a lot of uh guys here, as at community, also had a higher learning curve to understand hyperledger fabric.

A

So, basically, what's our solution?

A

That's why we've created fablo and pablo started as our internal project. Quite while ago there was no good hypoallergen network, tooling back then, and I'm working on this project with cuba or jacob, if you prefer, so he could didn't make it today because he's happy and busy father of two but he's really working hard on pablo also and right now.

A

We are in happy moment that most of our internal projects depend on pablo and everything is work really nicely and we can shift between the projects very easily and very easily changing network setups and after using it for some time we decided to open source it. So it's pretty short initiative like it's been open sourced new vacations near.

B

The summer.

A

B

A

So, what's the main.

B

A

Behind the fablo, the idea is really simple: uh everything is stored in one file that is configuration file. Let's call it json file for now and that file will contain all needed informations about fabric like public version, all organization, organizations, other types, peers channels, chain codes and absolutely everything what you need to get a properly running, hypology.

B

A

And then you give that configuration file to fablo and fablo will transform it into actual running network and the best part is that you do it with one command. So you just say: follow up you point to your configuration file and then the network starts running and let's review some key concepts and what's really nice about fablo.

A

uh First of all, config.json config json is heart of our project because it's like a complete blueprint of your network info and file, and everything should be a part of that file, but it was set so all organizations all peers channels, chain, codes, fabric version, private data, tls, so a lot of stuff so that no later for flags or additions are needed.

A

So config json basically represents complete state of fabric network that we want to have, and, besides of being really complex and big files doing a lot of stuff, we want to be it as concise as possible. So just covering minimum details minimal details. So often we base on default parameters that you can easily overwrite and I'll show it during demo that you can overwrite many things in fablocantic.

A

Also, we are heavily focused on being flexible, just to be basically as flexible as fabric is and it's building blocks. So we can cover a lot of experiments and a lot of use cases and everything in config.json is covered by json schema. So we can do some syntax highlighting to see some public features and also thanks to it. We can do some basic validation, so what options can we use and what it's not allowed in fablo?

A

Also, I'm gonna call it uh on this presentation config.json, but uh just one just wanted to say that we also can use xiaomi. So if you prefer that data format, it's absolutely not a problem.

A

Second part is pablo itself: pueblo can do various things.

A

Basic command is follow units, and it's your starter point, because you can initialize your new configuration file and you can decide if you want blank network without chain codes, or you want some sample one just to play a little bit with fabric and then you can modify your configuration file and if you decide to verify that everything is okay, you can use validate besides of validating it, against json schema. Fablo will also give you valuable suggestions like maybe it's not a good idea to lose.

A

To use solo order type to run your network and raft is more production. Ready. One pair for organization is definitely not enough. Maybe you should need another one. So validate is kind of easy way to get started with fabric and validate yourself and also, as it was said, we have command fablo app to get everything running.

A

But besides of running everything, we have also generate command, so you can generate scripts for running network and then then customize them a little bit for your custom case uh in case, for example, maybe you need max message card for orders increased for five or another default, so you can easily generate files a little bit customized for yourself and then then use them.

A

Also, when running the network, we have centrum upgrade for upgrading and experimenting with the chain code and usual ones like start stop down to manage the existing network and benefits of pablo first of all declarative approach. We are here heavily expired by kubernetes operators and we try to use as less commands as possible.

A

So we try to orchestrate the network or nvr using configuration file and that's why? Because one file is really ee, one file is really easy to reason about to deploy reviews and to check the actual state of the network. When we have multiple files and multiple scripts, everything gets.

B

A

So fabric, thanks to that everything is stored in one file, is really great for pocs and continuous integration, because you can easily download fabro, run it on continuous integration pipeline like jenkins or github actions and said: okay run this type of network containing this many organizations and channels, and just do it and fabio will easily do it and pablo is really great for future, because right now we support only one type of transformation that is docker compose, but there is a cool potential for other transformations like we also can include blockchain automation framework, maybe or maybe something else, because guys at blockchain information framework do a lot of great stuff.

A

So we could easily transform that fablo file again to something more production ready, not only docker compose file to run on your local laptop. It's really also really great that you can transform that configuration file in various visualizations.

A

So it's easy to business, for business, guys to reason about and to see some blocks and machines, and uh possibly we could do also custom plugins. If you have some custom transformation and custom case in your organization, you can try to do some custom transformation to meet your cases.

A

And I'm gonna talk only about vision, plans and then will be the first time slot, so visions and plans uh of pablo first of all, as it was said, multiple times, but I want to say it once once more. We want to treat fablo configuration file as a single single source of truth, like we can transform it configuration file into docker compose, but in future we could easily do a lot of stuff here, because right now we can share it with the rest of team and do some collaboration.

A

Maybe add orders checked why something is not working and we can have common environment between all developers. uh We could also easily do code, reviews and preview. Why we've added another organization, if it's enough peers for that channel or must we join much more peers and in future, this single source of truth also could be transformed into working cloud environment. So really nice option. I think, but I'm gonna pause for now and if there are any questions, I'm open it so.

B

Please feel free to unmute yourself and ask questions.

B

It seems like everybody is listening very attentively. Okay, that's cool.

A

If you have any questions, uh there will be also another slots for questions. So if you have something in mind, don't hesitate and just ask so: I'm gonna continue, so basically fabulous features and further features. First of all, we support uh plenty of hyperledger fabric versions from version 1 3 from versions one for x, basically one all one, four versions up till the latest ones. So two three free, I guess right now.

A

We support two order, types that is solo and draft, and two database types embedded, one, the default, one that is leveldb and the external calgdb, which is good for querying.

A

Also at our heart of our project, is setting up channels and channels. So we try to be as flexible as fabric is and I'll try to demo it really soon. So I hope you understand it and like it also, we have tls support private data, defining copies and the other other useful stuff.

A

So, okay, I'm gonna proceed to demo and after the first demo, slot will be also slot for questions, so we're gonna start really simple project using fablo.

A

So I'm gonna switch to my id.

A

And enter presentation mode.

B

A

Just a second, oh here it is all right, got it now. So, first of all, let's start with absolute minimum like here. We have empty projects, we have only intellij idea directory some git ignore and absolutely nothing nothing else and we'll try to change it with pablo.

A

So, first of all, as a side note, I have already fabric installed here, but if you don't, you can do it easily using this command, which is also provided in our github repository.

A

It will install small executable file that will communicate with dedicated docker and it will install it globally, but if you like, you can install it also in project directory. So this small file is really pretty elastic and when you have this file, you can type fabric in it and after typing it, as you can see, public json was created and that file contains information about json schema and it will validate everything against this g-screen schema.

A

This file contains also information about fabric version, about tls, about root organization, also type of order.

A

A number of his instances also number of usual organizations, and currently we have only one organization, called organization run with two peers, both running on level db and also we have one channel named my channel two and it will be joining both pairs of organization, one as as we can see, uh we currently don't have uh supported anything included in chain codes, but we'll try to change it to proceed to proceed with demo and start and show something more interesting. So we can try also to type fibro init node.

A

And after typing it we've also got for free custom, uh just minimum chain code to test a little bit and now, as we can see, we also have field chain code directory. It's pointing to chain codes chain code node.

A

The chain code will be installed on my channel one. The chain code version will be zero, zero one and it will be named chain engine code, one and, as it was said, with one command that is follow up and pointing to configuration file.

A

We can transform it into actually running network, so pavlo fablo just got our configuration file passed. It generated all scripts needed to run the fabric network, and now it's trying to run it.

A

As you can see our example. It's pretty simple. So it's doing great and fast and soon we'll have working fabric network.

A

And voila everything is running and as a proof, uh this is just minimalistic. Docker.

B

A

Fit for presentation, uh it's like just saying, docker, ps and using some custom formats and sort, and it's showing that we have one certificate authority for organization; one root certificate authority, cli for organization, one two chain code: instances for both peers; also, we have order and two peers and if you will try to execute it, everything is running and v-network is up and running so pretty much simple and fast to run your sample network and.

A

Easily from nothing we switch to this diagram here so root organization with two peers, one chain code and my channel one. I've also got second part of demo prepared. But if you have some questions in mind, I'm opened and it's time for questions. So I'm happy to answer everything.

A

I'm gonna pause for a minute. Maybe so don't hesitate if you have something in your mind.

B

So once again, please don't feel shy. um You can ask the questions just by unmuting yourself, you can post them here in the chat or if you're, watching live stream on youtube. I'm also monitoring the the questions there. So any of those is equally great.

B

Oh, I see one question here: we are in the chat already yeah, okay,.

A

View it, I can't view it actually. Could you please.

B

Sure yeah read it out: oh great, two questions fantastic um from dinesh we have. uh Where can we mention the fabric version in the config json file? uh Could you please repeat the question sure? Where can we mention the fabric version in the config json file? ah Okay, yeah.

A

Maybe I didn't show it: uh oh fabric version is here uh just at the beginning, so it's network settings and fabric version and if you like, syntax height lighting and not sure what fablo actually can do here, you should should have a nice pop-up in intellij. So you can easily switch between the versions so yeah.

A

So here's the version number.

B

Sorry, I beg your pardon. No, please continue. ah Okay and one more question uh from r givens. I apologize. I don't see the first name very simple question: can fablo be used to create nft and.

A

No, actually, no, it cannot be used because nft is more about designing smart contracts and a lot of work to do so. Many of these we are doing only various network setups. So no, it cannot be.

A

Hi so yeah, I have question so: can we use fablo to create a multi-host network uh multi-network like multiple computers in cluster? Yes, yes, yes right and no currently we cannot use it. We are only prepared for using it on docker composer, like single laptop. We have some plans to do something more cloud ready, but not yet probably uh spoiler alert we'll try to be already with it for christmas.

A

Okay, also one more question: so can we integrate explorer into this network? uh Not yet we cannot include explorer but also there's plan for it again, probably newbie christmas. It should be relatively simply because in future we plan to add explorer at this level, so at organization level, just another flag that we won't explore for organization, 1 organization, 2 or maybe global ones, so not yet, but soon.

A

B

And piotr we don't have a question, but we have a comment um in the youtube: live stream. I'll make sure you get the contact, but somebody says I would love to do a podcast with you.

A

Okay, cool, that's really nice! uh I will give our email at the end of a presentation, so we can think about some kind of podcast here.

B

Yep, sorry, we have two more questions. uh Dinesh again, uh do you have the feature to add a new org into the existing network using fablo.

A

No, we don't have such a thing. uh I'm gonna try to extend this example and elaborate on why we don't have such a feature. So just please wait and I will examine. I will show everything.

B

And the last question I think at the moment, from hakeem who says, can fablo be added to a pipeline that redesigns and redeploys a network after an update, um possibly.

A

In future, yes, because we are really keen on providing another transformations and another useful transformation is obviously kubernetes operators, because this whole, because this whole fabo idea really nicely fits into kubernetes operators philosophy. So in the future, you could easily also transform it into something more cluster ready and pipeline should transform your fabric network.

A

So that's the plan. In short,.

B

Fantastic, that's the last of the questions for the moment.

A

Okay, so we're gonna go back to presentation and chapter two with our network design, so we're gonna pretend that we are working in supply chain use case and we're gonna go back to fablo, go back to our configuration file and do some mess.

A

First of all, we want something more production ready, so we'll try to enable tls communication for security, also, as it went as it was mentioned, solo is definitely not a good idea for production ready environment.

A

It's only good for a simple use case to play around with fabric also, let's have free instances of order. Also. Currently we have only one organization and will try to add new ones.

A

So we'll add organization called just called organization, 2.

A

And with domain organization2.com.

A

We're gonna also have some peers. Let's say that we'll have two peers but to show how elastic fablo can be. Let's say that this time you want to have couch db, because maybe for uh for some reasons, organization, two wants to do some rich queries or just like scourge db.

A

So we have two organizations and now we must modify our channel gonna speed it a little bit so in the orbs section, we're gonna act organization, two and we'll only join pure zeros, so like first peers of each organization to that my channel one, and also we can play a little bit more and show another interesting thing, because we can add another chain code for sake of simplicity.

A

Let's just reuse this program here so chain codes, change, cause tv, node, so the sample one and let's go with chain code chaincode2 and let's install it on my channel too, and let's define my channel too. So second channel things are getting really complicated here, but still probably under control.

A

Because we have second channel named channel 2 between organization, 1 organization, 3, sorry and organization 2, and let's join peer once of that organizations to the channel and also, let's add this.

A

Organization, so organization, free and again I will change it to level db. It will be slightly shorter, faster one. So in like just couple of seconds, we defined three organizations, uh two channels and two chain codes and also change everything to tls communication. So our fabric version is secured and last cool thing that we can.

B

A

Is we can say that we want to do that? We want to have private data collection.

A

And let's say that this private data will be visible only for organization 2.,.

A

Oh, let's say that it will be visible or for organization two, so all names should be octu and let's name this private data collection, just collection, one and after having such a lot of mass, we want to type pablo recreate and again we can pass our configuration file.

A

And uh fablo is trying to pass it and again uh try to put everything up and now it's gonna execute for some time and I'm gonna go answer that one missing question. So basically the question was about: can we have some commands for adding another organization?

A

In short, we don't have such of things, because in fablo we believe that everything should be a part of configuration file. So in longer term we won't just. We want just to orchestrate the whole fabric network just using that file. So so, if you want to add another organization, you should add it to json configuration file and then fabulo will try to div it uh to execute some patches of the network and it's really cool idea and cool philosophy.

A

But it's also really complicated and it's not yet supported, and it's not going to be supported, probably for some time, because, first of all, we want to focus on kubernetes operators, integration and some cloud environment and maybe we'll try to add it lately, so it will be provided, but doing so such diffs. It's kind of complicated.

A

Oh, it went really fast. Actually, I wasn't expecting that.

A

I'm gonna verify it. Oh and everything is running so cool. I wasn't expecting this, but yeah. We have uh the network preview here in my console, so we have free, auto certificate authorities now. So we know that every every organization is running. We also have free clies for each organization.

A

We have uh mentioned couch db instances for organization, two, a bunch of chain codes, chain, code, 1 and chain code. 2 here is visible. We have order in raft mode already running and peers for organization, 1, 2 and 3.. So everything is running and we'll try to execute the command. Everything again.

A

And something is not ready yet just give me a second, my channel one.

A

Oh, that's a pity. I guess we have some kind of error at the end of demo.

A

Okay, I'm not going to try to fix it right now, because I'm kind of in presentation mode type, but in most cases everything should be running and easily. You could transform your network, so uh you can try it yourself and I'll try to look an uh on it after this talk uh to validate what actually happened here. I just want: don't want to spend some time on debugging random lines here, so we publish again easily.

A

We switched from simple poc kind of mode just to play with fabric to something really broad and complex, like a three organizations with two channels, uh my channel one uh with one chain code- uh I haven't done it uh it, but we also could have installed another chain code on that channel and fablo is elastic enough to do so stuff do such stuff.

A

We also can join peer ones from organization 203 into separate channel and provide some information about private data. So normally, if we hadn't uh have publo or any other open source project, it will be like work for at least two three weeks to get something running, tested uh and probably ready for for production and from for for some complicated cases. Here, we've just did it in like three four minutes with network execution, so it's really cool.

A

I guess and that's enough of our my demo, if you have some questions after demo, I'm happy to answer it and answer them now, and so it's time for time slot and then last presentation slot. So questions please.

B

We have a few waiting for you. um The first one would be from brett, which cloud solution would you pursue? Perhaps linux.

A

First of all, uh I'm just big fan of kubernetes, so it's another level of abstraction. uh So I'm not thinking about linux machines or something like that, because mostly I'm a cloud developer. So yes, I prefer linux. But on top of that I like to use kubernetes.

B

A

It's pretty useful and you can easily use also kubernetes on the linux machines.

B

Sure has a question in fablo: where do we specify in sorry? Where do we specify the endorsement policy or the chain code policy? Good.

A

Question because I've omitted it because in the newest fabric version we have such thing as default endorsement policy, but we also support older versions. So we could easily write endorsement and just provide an endorsement policy here.

A

We plan to do something more complicated, like to point organizations from configuration files that are mentioned here and try to pass the file, but for now just try to paste the invoice policy here and everything should be working.

B

And lastly, for the moment, uh another question from brett um one: second: uh what about the stack with internet of thing platforms.

A

uh Not supported.

B

A

We are not close in our business and in software meal with iot solutions, so we are just digging into that. So I'm interested uh into knowing more about iot in fabric. To be honest,.

B

That's it for the moment: okay, cool.

A

Okay, so last part of our presentation.

A

I hope that you could see that fablo is really great and it really is simple and it's really flexible and you could easily define such complicated cases with fablo like supply chains, so many organizations with many channels.

A

You could easily switch and refactor your file to cover such cases as two organizations connected with audited channel with one root, audit organization validating or approving some transactions.

A

You could easily also create such configuration file to create free organizations with common channel and then also create small channels joining each organizations. So fiber is really flexible and good for various solutions.

A

So it's nice, but to be honest, pablo as everyone and everything hitler in life also had its drawbacks.

A

It's not yet production ready, as it was answered on qa session uh just minutes before, because we currently support only docker compose on single laptop, but in future we will be ready. Currently, we don't have any good support for linux and we support only sorry. We don't have any good support for windows. We support only linux and mac os and it's very very good to test it on both of these machines and various systems here. But not on windows yet and we can do basically imports of running network.

A

So if we have some custom solution, we cannot easily transform it into further configuration file to like re-pass it to configuration file, but I have some ib ideas on how I can improve it and how we can proceed with it. So maybe in future it will be available.

A

Also at it's what's uh mentioned, we don't yet have uh support of hyperledger fabric explorer, but it should be very fairly easy and heavily pull requests with it. Also when welcomed- and basically that's all for fablo.

A

uh If you got you interested here's our domain, that's fablo io, it redirects us to github repository of pablo and if you get some, questions want to do. Some collaboration uh talk about some open source and stuff like this. Here's, our email, it's blockchain, softwarenew.com, both kuba and I will be available on this email and answering all questions and yeah last time for questions here.

A

So if you got something just hit me.

B

Last chance for questions, anybody feel inspired.

B

Just give people a few seconds to type I type with two fingers, so I don't expect anybody to type quickly.

A

Just a comment here: this is very impressive. Thanks very much for the presentation. No thanks.

A

Thanks actually, to be honest, there was a lot of work to do it's so flexible and we have a lot of battles and then a lot of battles with macos and walking head both on linux and news distributions on mac os. So thanks for the comment we are really happy to deliver. It.

B

Look, I have a question because um one of the most considerable appeals in your project is simplicity. um Have you worked at all on any um user interfaces so that people with minimal knowledge of any blockchain could perhaps interface with your your architecture?.

A

Yes, we are thinking about.

B

A

Right now, just to start with, we want to do some visualization, so transform it into like.

B

A

But we are also thinking about some interfaces, so in future.

B

A

And it's cool idea.

B

So when you say visualizations, which you mentioned in an earlier slide, would that be a form of interaction rather than a you know, drag drop menus. It would be more perhaps of an intuitive engagement with a a two-dimensional de-centered image.

A

For start, without interactions just to present, maybe such a diagram like this one, but in future we plan also, maybe some console to do some drug, dr dragon dropping and configuring schemas on fly.

B

um uh Randy has given us a question fresh question: when do you think fablo might be ready for full production.

A

Oh that's a hard question because, to be honest right now we are doing it after hours in our spare time. So it's getting really slowly there, so maybe next year, sometimes for something more just try to follow us. We are also open for any kind.

B

Of collaboration.

A

So if you're interested.

B

B

um We have a question from jorge on youtube. uh What considerations should the fabric sdk have in order to integrate with fablo.

A

Basically, nothing because fablo doesn't do anything, anything magical or new with sdks, so only one missing piece is configuration yaml for sdks, that's the usual part, and I guess that work on it is actually on progress.

B

A

So soon you'll probably have a chance to transform that configuration file to connection profile for your sdk.

B

Says, thank you very much. This is very exciting uh and asks where the best way, what what is the best way to follow you do you have maybe a newsletter or is it simply the github repo right now only github.

A

Repo, because only github repo, because I'm not a twitter fanatic yet and I'm not trying to post every changes so best way to follow- is to try to use our repository and maybe official twitter of our company and we'll try to make something better soon.

B

And just to remind people the company is software mill is that dot com or dot p.

B

I'll, just oh yeah, I've just got it, that is the central company, so that also is a way to to follow. Yeah.

A

And we have pretty active twitter, so everything uh about pablo will be on that twitter.

B

Yeah, I just noticed brett said he's already joined you on twitter, so there's that's um exactly what you wanted people to do, so that appears to be the last question. um Thank you very much. Indeed, it's been really an eye-opener and I'll certainly be using it for one, and it's also very kind of you to give your contact emails and also an open invitation for collaborators because, as you said, this is being done at the moment in your spare time.

B

So thanks once again for helping us after the the end of your working day in poland, thanks thanks.

A

Yeah and thanks for warm introduction and some compliments, and hopefully we'll be also be seeing uh ourselves induction community and on next talks.

B

Fantastic well thanks again and please stay in touch thanks, bye, bye, all right, bye, everybody bye, everybody.

A