Fluence Labs Community, 6 Oct 2022

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Developer Community Call #18

Description

Bazaar Zero team presenting the project built with Fluence
Deterministic AquaVM, updated Registry
Tech team updates
Business team updates & upcoming events

Useful links

Bazaar Zero project https://bit.ly/3RAffxv
Hacking On Fluence https://fluencenetwork.notion.site/Hacking-On-Fluence-Primer-28a87754397048e1bec72e3bfc91fd9b
Community Calls wiki https://www.notion.so/fluencenetwork/Fluence-Developer-Community-Calls-2733beddb3ff4e4986c4fcc830fc796d
Fluence Documentation https://doc.fluence.dev/docs/
Aqua Book https://doc.fluence.dev/aqua-book/
Discord https://fluence.chat

A

Right I got here early, so I'm impatient. Let's get this show on the road, uh it's gonna be a pretty full packed uh Community call. So thanks everybody for being on the agenda today is we have uh updates from uh the business and event side from Anna. Then we have a very cool our hackathon winner from the polygon uh hackathon, here bizarre zero, welcome and looking forward to your presentation after that, then we'll have a couple of short Tech updates from the team.

A

I think folix has a few things, and uh then we have a couple deep Dives one is Alex a showcasing the latest release of uh registry and followed by Mike, with a deep dive on all the cool things and changes and upcoming changes in aqua VM. All right as I'm already tired uh other people thought I'd. Kick it off, there's a few things, and uh then we have a couple deep.

A

That sounded like me, and it wasn't me.

B

B

B

All right, um first things for it first wanted to show you um our community calls uh Wiki. So we are just collecting all the uh history of the community calls and you can see the next Community call or today's Community poll you can go inside here and see what are the hackathons. We have, whether it's link and there will be also live stream and recording link.

B

We are live streaming today to YouTube um and in every uh video we put links the references to this weekend, so you can easily find it or you can find in our Discord. uh So let me just walk you. Through the uh the upcoming events and hackathons for the hackathons. We have lumos Labs hackathon, uh that's going on right now and uh you can still join. uh They have a lot of activities today, those Twitter spaces, and they have workshops and a lot of tracks.

B

You can see from all different sponsors as well as from us. There is a 5 000 USD you can, you can get for building the best web, free, tooling, API, SDK, SDK services and etc, etc, check out here or in the GitHub influence, Labs lumos hackathon.

B

um Then one of the upcoming hackathons in the next year we are going to have that's in person, that's in uh in Denver from end of February to beginning of March. We will tell more about that. We're just signing up for this one. That's that's going to be huge next year.

B

um That's and that's Pro, also virtual, that's always in person and virtual and then other upcoming hackathons. We are not necessarily participating, but if you guys want you can check them out. It's it's in San, Francisco, that's in person, if you're in in the US, if you are in San, Francisco check it out uh web3 weekend, that's uh online virtual from if Global end of November uh check out their website git coin.

B

As always, you probably know the git coin platform they have a month of the hackathon and if India that's in person will be huge beginning of December. We are considering of participation there with fluence to sponsor volunteers, but not yet decided and um in October end of October, beginning of November. Our team will be in Lisbon in Portugal if you're there. The events are following that we recommend you to visit and that we we are going to um to participate. That's the lab Peak! That's our partners! That's our friends!

B

um Protocol Labs is organizing. There are lots of different activities, you can check labweek website and you will see sorry, it's a bit lagging. You will see.

B

They have different Summers different meetups um whatnot, uh also ipfs cam Phil Lisbon, the same as it was in Singapore and uh East Lisbon, the hackathon that is organized by one cakes and other partners, and we are going to have our own fluence meetups, uh October, 31st, November, 1st and November 3rd, so feel free to join again all the links we will share later in Discord and Telegram, so you'll see um that and for the other events that are coming up. uh That's computer or data.

B

That's also a part of lab week and other different conferences also do check it out if you're interested, um that's it from from my side. Thank you. So much.

A

All right, thank you. Anna anybody have any questions.

A

No all right! Well, that's our zero.

A

It is so in the recent hackathon polygon hackathon Baza was an entry and they worked on a project and they can explain as much better than I can basically decentralizing marketplaces like openc or air ball, Etc and uh they're, not the first project to do it, but they had some really interesting wrinkles in it and improvements, including uh uh the ability to buy the multiple assets so ERC, 2721 and 1155 into a single product and I had some interesting uh approaches in terms of provability using serial knowledge proofs anyway.

A

Let me hand it over to you guys uh just introduce yourself and uh show everybody what you did.

C

Hi guys, foreign.

C

C

Tailored videos, project.

C

C

Okay, uh we have problem here, sharing our videos can.

D

Someone help us to to play the video. Actually it.

C

C

To share the videos.

A

There should be on the on the left side. There should be on the the left bar. There should be a share screen button in this group.

C

It's okay, we need to restart and the Discord.

B

Okay, so now it should.

B

A

Do you find the button the screen sharing button.

C

Yeah, you need to give permission to the Discord.

A

A

All right all right.

C

C

Yeah, you can see the video no, there is no no sound right.

A

A

A

C

They are building a match up the future Marketplace.

C

Businesses, so it means.

A

Okay, I, don't really see anything and I can barely hear anything.

C

A

All right, the picture's back now.

C

C

C

C

Thank you, foreign.

D

Okay, can you hear me.

E

C

D

Right so, okay, let me play the video sorry about that.

C

D

Can you hear the sound yep.

A

A

How about instead of uh doing the video, why don't you just tell us a little bit how he used the the semaphore, a certain knowledge proof, and how do you use fluence in your overall situation? Maybe that that all.

D

Right, thank you. Okay, sorry about that. So we'll expect the video okay. So the first original idea we had to have is actually that we want to build a new type of what we call a cross chain marketplace right. So there are two things that are main focus that we are trying to look at is that we want to play around with zero knowledge proof, and we actually want to play more on uh fluence using affluence to build a more what we call uh PW Communications, so I mean the project was inspired by open buzzer.

D

So we like the idea of open Baza, that everyone can actually open up their own. What you call this uh can open up their own shops and they can close the shop when they like it. They can open when they like it, and things like that so um uh So, based on that kind of uh what kind of inspiration that we actually found out that achieved somewhere, we are using Baza ex, if I'm not mistaken, for fluence.

D

So that's actually our first foray interview of understanding how fluence actually manage app, how to use floors together with the nft in terms of the market trading but for buzzer is it ex is the main focus is actually to trade. The nft and the nft is actually a physical products wherever in ours is actually to do a transferable like a bundle of a bundle of the projects right. So uh what the first thing that we actually do, that on, that point is actually that.

D

So uh one thing is that we need to build is actually that using we are using uh polygon.

D

Actually, one one thing is actually to generate uh mean for all kind of a smart contract, but we need to have some sort of control, so it means that what we want to do is actually that they can only mean the mean the nft, whatever sorry I'm, sorry about that, the nft is actually a special container nft, so you can actually put more items more assets inside the nft example, you can put the erc20 1155 and uh seven to one into the nft right, so uh so the first step is that that every note will have their own sequence wallets right, so I think because I can't share the video, but I did share it inside the general group.

D

So anyone actually can take a look at the videos to understand more on that thing, so we already shut down the server for for the applications right. uh So uh every note we have their own. What we call a smart contract wallet, so any user they want to open a shop will will need to actually open up the wallet.

D

So, instead of because one thing what we want to avoid is the same thing as openc, because it means that every time you want to sell the product or that the product is not actually with you, the product is HTV openc right, so it defeat the the purpose of a decentralization whereby that we have a full ownership of the product. Once we transfer the openc is the ownership is actually based on the open, C so using a fluence right. So it means that when you create the notes.

C

D

So when you create the notes, all that, so together with the sequence wallet, you can actually create what we call a smart contract on every note, actually, maybe a wallet for every note. So every time you want to sell the product all that the nft are being transferred inside the wallet, but that wallet are still under the what we call the user control. So if the user can take it out the item as they need, but as they take out the item. The item soccer item automatically same uh the the item automatically be canceled.

C

All right, uh okay,.

D

Sorry and I think I've got without the video.

C

D

I guess there's.

A

D

Mean for this they can be one we're actually using a summer for so what we do is just that, so you can only mean this container nft through up the notes. You can't actually go into the smart contract and try to mean it directly. So what you do is that we actually using the.

D

If you look at the videos that the before you can do the mean you need to get a proof from the smart, another smart contract of a summer for smart contract as AKP smart contract right once you get the proof wherever the proofs are being sent through, one of the fluents we're using a fluency as the what we call to store the proof. So once you get that proof, then only you can mean it from your own notes.

D

For that, uh then, after that, after you do the sales all that that that uh container actually been destroyed- and you actually get the item whatever being bundled inside there, so you can actually bundle up uh ESC 2115.71. So it means that you can pay a seven to one and nft and also a token together as a single sales. So in the hope that you actually give more values to the marketplace, so uh one thing that we actually take a look on.

D

This is actually using a registry system for the first time right and uh we actually pull out how we do the search is I see that the search is actually happening on real time. You check out the notes that are live on that point right, so any item I've been published inside the notes. uh I've been put from the network itself right.

A

I I got a quick question here before you go on.

A

Basically, uh proof membership in a group right correct. What's the group is it the nodes is that the group.

D

So the no yes correct! Actually we have two type of nodes. Actually, in this case one is a seller note and the other one is actually what we call uh application modes. So this is actually the idea that anyone can build any application from this marketplace right. So, if you don't like my my application like this, so the marketplace that I open anyone can open another marketplace with a different kind of royalty. That's the example that we look, but every kind of note. Basically, the seller notes are being registered to the summer for right.

D

Yeah I mean it's as a proof to get. The proof is that you need at the first time when you register an account right when you open that aircon, you automatically have been register as a proof, so that notes I will register as a seller.

C

A

When you say node what? What exactly is that node now that seller node? What is it.

D

Okay, so every for this work we're actually using a fluency s, so every first dance is actually that each one of of is you. We call it a notes right, so every influence there is a note, but it's like it's like a slow down all this kind of thing. So it means that every time people want to sell it, they can actually deploy it locally right, their own computer. So they can I think we did actually completely draw the hackathon they can deploy through a Docker or things like that.

D

So it means that you can just uh clone the code and run it directly. The fluence they automatically bind it to the network right throughout the network.

A

C

A

Right yeah, thank you.

C

D

Yeah, so that's actually our first project, actually, the one that we want to try is actually the how how actually we can use. uh Actually, we will do an experiment, for this project is actually try to look in to build a new type of marketplace using affluence right, so it means flow as a base. What we call as a core Communications right, the core integration that is actually directly connected to actually to the smart contracts.

C

D

Yeah I mean I will share some information. You can check out on the videos so uh and actually inside the videos. We have the the GitHub that you can actually check it out.

A

Okay, now before I interrupted, you we're talking about the registry.

C

D

So I mean okay, but that's actually our first time playing with registry I think that it's something that we learned from the last call right. The last Community call. So it's something that we want to look, because how I understand about the registry is sorry. You should ask the question. I answer it before you ask a question right: I.

A

Don't know I, don't I, don't have a question. I just wanted to make sure you're back where I interrupted you you were talking about registry and uh oh, no I have not read I just wanted to hear what you had to say about registry.

D

Okay uh I mean with in the modern agency. It doesn't have much information yet because then, when we play out, I mean we don't understand from the aqua boat all that, but that actually is just our first experiment in registry. So, as we understand, the registry is actually to register the node for you to have the search. So what we do is actually that when when we actually deploy another search node, so basically there are two types of no one. Is a seller note?

D

Another one is a search note right, so the search note is so called like a Marketplace note that they keep pull all the information from the seller. Note the available seller node. So uh we are using a registry that to register this separate type of uh what we call uh separate type of notes. So it means that one is a seller note. So it means that when you do the search they only search the nodes they are being registered as a seller. Only I'm, not sure.

D

Is it the right way of using a registry or not, but to categorize.

A

On the call he's going to talk about registry a little bit and uh definitely we should figure out to establish a Channel or whatever in slack where we can help you out a little bit.

A

Okay and uh the other things I only have one more question for you. You use influence.js as your your peer implementation and have you like ever looked as everything you do? Does it translate into uh webassembly on the rust node, or is it very specific to fluent shares.

D

uh I mean because then uh we, both three of us is our background, is a node.js developer right. So actually, influence.js is our entry point to influence, definitely really study about the rust, the image you have more tools and definitely a more easily easier to deploy and to manage, but because that, when we do all this as framework, we guess that instead of we are trying to focus a lot on the language part of it, we want to focus the beauty of fluence first.

D

So what you can do like what it can be done, but I do understand that if you actually go to rust, you can actually do a lot more things compared.

A

To um I just was wondering uh what your process was and you you answered it from uh from a tooling and uh and back-end perspective, so very cool.

D

All right, okay, so yeah, that's basically about the the open buzzer uh that we worked on for the last second. Hopefully we can do more things. We want to bring out the products out as well to to do to build out the first, hopefully, a multi-chain Marketplace nft for nfts.

C

D

I guess that's all for my side.

C

A

All right, cool and uh we'll we'll stay in touch, we'll, try and figure out somewhere. We can help you out with a registry a little bit yeah. Anybody else have any questions.

A

No all right, uh I think the next part was.

A

What's the time 12 31 um Flex, do you want to give like a brief update on uh all things? Tick, really, quick.

C

I think our main updates are from Alexia for registering from Mike, okay.

A

Then we'll go straight to Alex a who actually has updates on registry, which of course, is very timely for the bizarre team. So all right, Alex, a all yours.

F

Yeah, do you see it? Can you hear me, thank you.

F

E

C

F

I will tell you about latest register updates uh I, as you can remember or not. Registry is a service that provides advertisement and Discovery for services. It creates relations between unique identifiers and services, and we can resolve peer ID and service ID of a group of service by one identifier and it's a built built-in service, it's available on every fluence node.

F

So why is updates was something wrong because uh we have previously web scripture knowledge, like providers, not providers. It was hard to tell about it. It's hard to understand uh and in fact there is not a lot of difference between these providers as we call it because it was misdesigned.

F

Also, it was broken life cycles uh because it was impossible to renew a record without Jazz client if record responds to the service on the raspier uh also, whereas there was no way to remove record, you can register some service or record, but you can't remove it. You just need to wait like a day, and it's strange also. We had a very complicated API with strange namings because uh because of first thing with obscript terminology, so what's new, uh we come up with a more clear technology.

F

We have a resource, it's like key in cable storage and we have a records. It's values in with storage and record represents routing information about registered service like period service, ID and optional related. If we deal with a dress beer, we removed providers and especially not providers from terminology, and we have only records records for services with different peer IDs.

F

Also, we fixed life cycles, we've introducing six service for registry for every last peer, so peers can commit to the record, but with spear hosts some service. For now we have no checking except of uh tetraplets and we can sign anything from the registry. But later we will add some challenge solution based decision solution to it. A record contain two signatures from each year who wants to register some service and from the peers.

F

Where is service hosted so with that raspers can renew local records automatically with ground jobs scheduled scripts and it's cool so for record removal. I added a thing called thumb: Stones. It's something like Cassandra, tombstones or Alberto stones in distributed systems, it just uh record, but it replaced all the record and keep some time and eventually record will be removed from all the network and it's distributed like records, and it's almost similar, also I updated apis, it's more shorter and clearer. We can create resource.

F

We can resolve resource, get resource, ID a register service record Another registered service and find all the records for the sum resource ID. So it's become really simple to create professors and register uh both local service. If we deal with just peer or remote service on the network like class model test net, it's just different uh different uids, sorry, uh also with how it looks to resolve resource and get all the records, and we can go to the period advertise period and uh do something with the service which we want.

F

uh I always looks like we can attach the service and check the result. I'm sorry, my cat doing settings I think uh also I created uh first Aqua API tests. It's the first service with a test for aqua. It's just python Runner, let's run by test, and it checks everything in API and it's cool because we will build up, we will include it in CI and it in a register becomes more stable, also I created step-by-step example, with fluent CLI.

F

It covers creating Service, uh registering service in the registry, registering JSP service, resolving it and removing like it's all API in it.

F

uh What's next with registry, we will have internal hackathon, which help us to test the stability of registry. Also, we will add. We will start that we will. We already start started to add challenge solution. Permission management for resource like obneros resource, can decide who can add records for it also I think about resources. Removal because we can remove records, but we don't manage to remove resources.

F

Also I will add more tests and examples, and it will be good. That's all! Thank you. If you have any.

A

Questions. Thank you. Alex, hey yeah questions. Anybody, foreign.

F

A

Right, Mike you're up next.

E

uh Can you can you hear me yes and uh see my screen in one two yep and now you should, you should see it. Yep is the key. Yes,.

A

It's excellent.

E

Cool, thank you. So I would like to share some progress on the machine Department side so especially on the aquarium site and on this picture you could see this diagram and all feature dependence and feature progress that we had and at the end of August. This is here yes and at the end of the August, our plan was to support first of all, uh determinism on the on the accordion. So here you can see arrows that represent different tracks on the aquarium side and on September.

E

So uh we implemented a lot of features and the most notable ones they are wise on the determinism track. So here you could see that we implemented several features, so our next slides I will discuss shortly about each of this feature. But the main agenda here is that, like determinism is almost here, it's almost an aquarium.

E

um I almost I mean that it's implemented, but one of this feature reduced stream scope is not yet merged, because it's a breaking change and next week will have the internal hackathon and uh we we plan to merge it next so like in two weeks and after that the aqua VM would be determinist, deterministic and versioned.

E

So what does it mean to be deterministic? It means that Echo VM won't break by using an appropriate values in an appropriate size on appropriate basis. So we will see next slide how it could be uh obtained. How could you, how could you write some code on Aqua or well? Not now it's fixed, but we will see you could you could write some code on aqua and then some uh some very complex scenarios it it could break.

E

Also we have progress. We can progress on other tracks, so uh one maybe the most valuable here is the support of Lance Hunter. So it's very very useful thing, uh though it's hooks comp, it looks simple, but from the implementation perspective it's quite complex.

E

Also we uh work working on a hardening of aquarium by hardening here, I mean that we are trying to make a cream as stable as possible and to make it more stable and more uh I know uh more major in terms of some different scenarios in terms of some no operability of having bugs and all this. So here you can see that we implemented so some stuff like dedicated types for air positions and scripts.

E

So we improved tests, especially since test for Canon instruction that was introduced in previous months, and we found some bugs with the scanner instruction and also we are working now on a formalization of Aqua VM. uh So we're working on them specification uh on key framework and then we plan to uh have we want. We want to make uh model, checking with TOA Plus, uh so you're working on the site. Also, we introduced a new instruction called Never, so it's quite simple instruction, but really useful for some scenarios.

E

So it's basically all for, like the whole picture, the whole picture of the website progress. So let's look into details a bit uh so regarding length pointer. So the idea here is quite simple: in error. Previously it was impossible to get size of value in that icon on the aquarium site. You have three different value types. First, one is a scalar they're json-based. Second, one is stream, they are 32 like and they are I would say they kind of array based and the third one is a Canon can provide streams.

E

There are also array based and we implemented once factor that allows you to obtain length of a cannot streams, as well as a lens of array-based array.

E

Type Json values of I mean scalers, so it looks pretty simple so here you could just write dot lens and it's quite similar to what we have with wonders, but syntax a bit different, because so I'm going to start with this sign with the dollar, with DOT on a door time for factors you shouldn't use it and uh but it's quite quite difficult from a internal side, because so there are some issues with stutter plates and, uh like all all stuff is scanning chemicalized streams and all of this so but it's implemented it's merged and delivered so an aqua already supported, but previously from the developer perspective who uses on the aqua.

E

uh It's also was possible to obtain lens of a of a stream of a scalar and all of this, but it was implemented with a service call, so it was kind the hack because like to obtain lens you just you just need to call a service written on job assembly that just returns a return to you uh size and it was kinda uh kind of tricky yeah, not not an optimal cost, and you are working on this light. We have.

E

uh We have like uh do you have some thoughts about introducing new factors that so especially, for example, for math operations, because they're, quite simple and now for math, also service calls are used and there's also not so optimal. uh Next one is another instruction. Never instruction marks a sub graph is incomplete, so, regarding subgraph, uh like air internally in the The, Interpreter is presented as a graph, uh not a tree, but a graph, because Cycles are possible. uh They come from recursive streams and uh because there are two instructions, two simplest instructions.

E

The first one is no, and no is just does nothing. Yes, it's just not Mark uh subgraph, as as whatever it's just nothing and never hear marks a subgraphers incomplete. It means that, for example, for this pattern. Yes, if we have instruction.

E

But to me we use this font, so if you have the following pattern: okay, let's say call here so from the uh The Interpreter perspective. So this instructions reads as a left sub graph, and this one is the right sub graph and the SEC or sequential combinator. Yes, it's it's combinator that comes directly from PI calculus and it works in the following ways that it executes right, sub tree.

E

If and only if, left was executed successfully and in this pattern never marks this subgraph as incomplete and a sequential combinator knows that it's incomplete and it doesn't cause this right. Sub graph is called so this call will never be called action.

A

A

Or did they never Operator just kick in and you're never executed.

A

Right I think we lost Mike.

A

Hey, are you back Mike.

A

All right he's not back okay. Well, in that case, uh I think that concludes the the presentations from the lewin's team. Mike.

A

Well from influence team, so anybody have any questions uh about business upcoming, hackathons events, anybody who wants to meet up in Lisbon any questions for the presentations about registry or uh or Aqua VM, which might just brought anybody anything.

C

Hello, hi, hello, I, have a question regarding the registry just now: yes, oh all, right, okay, so uh that we try to do that. What we actually tried to do last time is to to actually create a dynamic here, so every user we have, uh on the front hand side, they generate new type of beer, and this pair will be registered to the seller right so uh because I'm not sure that we do it correctly or not that actually that we didn't actually attach any key to the appearance.

C

So if all the peer need to be registered inside registry. So at one point we realized that actually that we make the registry become like a saturated by a lot of peers, but not a lot of it that are not as actually kind of used or active right.

C

So I I saw just now inside the the the documentation there's actually no way of removing that bill. Is it.

F

Yeah, so you can create results for group of peers register peers or for this resource and remove unwanted As You Wish yeah. It will it's already live yeah.

C

Okay, so we can actually remove them now, the beer that we don't need.

F

Yeah just update to the latest uh registry and aqua yeah, and you can do it.

D

Okay, but one thing is that the reinforce is I've, got the peer uh dynamically.

C

Generated like automatically generated how we know that the pay are not needed, we reach repair are active which pay are not active.

F

uh Whereas no health checks now you should do your own. Like someone should uh someone should do it? It's not automatic register, just a basic block. You can register and register, but you should decide by your own who who need to unregister?

F

Oh maybe if it's your custom peers, we can unregister by exit of process, for example, when.

C

F

C

Okay, all right so.

D

C

That we need some sort of process or service that you're running to check all the peers like using language. Every building our own house check by bringing back to the notes or something like that. Yeah.

A

You should have the notion of a heartbeat somewhere.

C

uh So it's not building. We need to build that heartbeat run. The health check.

A

Yes, so one thing you could do is for every for every peer. If you have a ping function.

C

A

Like automatically and then you can just go through that right and then you can just do garbage collection on that. If, if you want to.

D

They're the same process that every I think it's every hour, they're actually gonna, do the cleanup automatically there's actually still running all right.

A

F

Can you repeat the questions? Sorry um looks like uh garbage collection. Okay for now, uh if you register some service in the on the raspier it are in use automatically like uh to until you unregister it, uh it will be forever in the network yeah. uh If you register records for JSP, like you create service on JSP, uh you should renew it by your own and you can stop renewing. You can unregister it uh and yeah. That's all so. Garbage collection is actual.

F

F

C

F

Is garbage collected? uh If you stop it renewing to stop the training, you should stop Renew on dress peer or you should unregister it on the raspberry. That's all, and also for now. uh Only issue of the record can remove this record, but maybe it's good to think that owner of the wrestlers can make an May delete this record and think about it.

A

C

We need it all right.

A

Did that answer your question? We can follow up offline if you have more questions, yeah, yeah, okay, all right! Thank you all right, I think Mike is back Mike. Are you back he yeah?

A

Do you want to finish up.

E

uh So yep I I can finish up. If you.

A

Have time yeah we have, we have time finish up.

G

Okay, so space for other questions.

A

Yes, uh who has questions.

G

uh Yeah uh Morgan here.

A

G

There um yeah I just wanted to clarify from the prior question there. You mentioned a ping uh sort of like a ping timer for my understanding with fluence, it's more of a you push to a fluent node, and then you, you sort of like pull the response, and you can't necessarily push from a fluence node to a front end. Does that is.

A

G

A

Would use a scheduled script for that.

G

Scheduled script, yeah: okay, uh that would be a script developed in Rust in aqua. Oh okay, in aqua, um okay and uh I'll I'll probably have some coins afterwards, but um I had other.

G

I had another question regarding uh sort of front-end use case, with being able to use fluence as a type of impression counter on things like a decentralized website.

G

um My current use case is looking at something like uh if you were to have some service that was um basically using a SAS model of some sort for things like UTF for Unicode, um um low or signs or logos. Could you use fluence as a type of intermediary or middleware between websites developed? So, for example, if you had an interface that had certain ETFs and they had certain counts or clicks, or something like that, and there was a let's say, a back end with ethereum, but using fluence as like the middleware for activity.

G

Is that something that's like can be doable or has been done in the past, or something like that.

A

Double yes done in the past, I, don't think so.

G

Okay, cool, um and uh can you I haven't looked at or I haven't uh encountered this in the past? Can you see like? Could you I guess there could be something like um you know, I've seen examples in the past where you've had uh online. You know seeing like how many online users there are for a chat, for example.

G

But could you see how many online connections there are from like thinking about um you know if you're thinking about like the hypertext space of like websites and like their relations, uh could you build like a mapping of these UTF symbols in sort of like a network? um That's like shareable in public.

A

I'm thinking anybody else from a team wants to chime in don't wait for me.

G

I can reframe the question if, if uh basically.

A

You want an impression counter of, uh and, and you want to want to represent that as a graph is that correct.

G

Yeah exactly in the line graph, so you could like see people join and close, but you could be able to see the actual utfs um being displayed in some type of visualization. And ideally this would be yeah. I mean that's sort of like a bare bone test case. I'm thinking about right now for.

A

Eign acquiring the live connections. Right, that's that's a challenge! The rest is just data processing.

G

Okay, is there a Max live connection for fluence nodes.

A

Well, it's with well, it's peer-to-peer right, so you have your your your your limit on the neighborhoods, but uh you can extend it through cash and then okay didn't like uh somebody did something about that. Jura didn't Jura, do something about that in terms of capturing all the the live connections and then visualizing them Folex. Do you remember.

C

I was graphing online nodes, not the connections, okay, but if you want to to track connections for to like to front ends, you can do that by storing meta information about this connection and maintaining it. um So that's possible.

G

Where would the state be held or or is it like? Is there even a necessity to hold any state if you just have like foreign connections being created and I guess, I need to clarify what what is the max for the neighborhood um influence currently.

C

It's a state should be somewhere.

C

You can maybe access the rust peer connection information if we provide the the API for that, but I think it's enough to hold the state either in front ends and share it or on Rust services like webassembly services like have a registry of users for your application, like participants, maintaining it there by pinging everyone who's online periodically, that's how it's usually done on back ends and it's possible influence.

G

Okay, cool and uh I guess the the final question there is like: what's the max uh in a neighborhood, that's the I heard that.

C

Well, you can specify how many nodes you want to retrieve from the neighborhood. You can either even like, say 100 or 1000, but default is 25 or 20 just any. So it just depends on what you want, but usually 20 as a replication is enough for most applications, usually like it's three four five, so 20 is more than no okay.

G

All right, um if.

C

You start developing that you should pin ping us with any questions. It's rather clear uh Target, so it should be easily done as far.

G

As a yeah I'm, okay, thank you very much. Yeah I'm I'm, still like every time, I dig into fluence I, always hit a bit of a wall with the some of the um you know, developing the rust services and then some of the syntax with aqua, um but I do have this application. That I would love to uh test in some type of playground.

G

um I'll be developing some things in relation to uh do the payment like. Basically, these utfs would be streams or super fluid streams that you would that people would like have donators to and in order to use them in your site.

G

You kind of open up a future yield on um with ap wine I'm looking to do that at the Bogota hackathon coming up, but I know you guys, aren't um folks, aren't hackathon sponsors for it, but I feel like there's still this another piece that might pull out afterwards and test ground to actually develop with it. So I mean I've done a little bit of impasse, but yeah, okay, yeah. That was really helpful.

A

Okay, cool Mike: you want to kick it in and kick it off and.

E

uh How many times, how much time do we have? How much time do you need? uh Let's say something about seven minutes.

A

You get seven minutes. Okay, do it.

E

uh Let me try, let me try to share screen, so there was something with Discord on my side, so hope hope you see it yep, so uh I ended somewhere here, yes on another instruction, yep cool, uh so uh so I I said that, like in this pattern, so if you have the simple, a simple Aqua script, so on this side, if we put never on the western graphs that never makes uh this one sub Subway office, he is incomplete. In this case, this red subgraph would never be called so, and why is this this?

E

Never instruction is needed. Actually, uh so the main purpose of uh introducing this instruction was to allow fold uh Works more precisely and have more, you know more advanced semantics, so let's consider how it works now and what what the change that was done and how never is used to follow the change. So, if you see on this, uh this box, we can see like simple microscript is simple fault. So, on this fault we have this sequential combinator as a body, and here we have next and next works.

E

You know how we invaded it, takes iterator, move it to the next value. On a stream or in a scalar that we are used as a heater bone fault and if there are no such value, then next, oh sorry, if there is variant in a stream does it will move iterator to next video and then call recursive with this uh by the third of a body of a fault?

E

uh If no values, then it calls uh generator cursively exits from this like uh recursive uh chain, so why it's needed, because actually there could be instruction after next and uh they they would be a recursory code. But, okay, let's uh imagine that there are no such instructions, that is such a simple body and how it could be executed.

E

So let's say the stream has two values so V1 and V2, and the execution flow would look like this. So first of all our VM executes fold itself, it adjusts the intervals, iterator and so on, and then uh next sequential accumulator would be called, and so actually it looks like a graph traversal and more precisely, it's so called pre-order graph traversal.

E

So in this product, after also, we executes execute instructions. Then uh move to the left, sub graph, execute the whole left sub graph and then move to the right subgraph. So the website, graph of a call here is only one uh sorry over the SQL is only one. This instruction called and it's executed with value V1 from a stream. Yes, this one, because here you could see that we use it radiator as a in the argument of this Construction after execution of this call when it's when it's ready.

E

Next, this right, subgraph and then, as I previously said this next, uh we'll move it right into this next value. Here next value is v y V2, yes and call recursively the body of this fault. Yes, and what's what's about the body so the body here again, it's cool yes, and this again this SQL will be called.

E

Yes, then there is. There would be uh this pre-order tray also and according to it, we will execute website graph. This call with the value P2 from our stream. Yes answer the question: what would be called here? Yes, because here we need to call next and next uh so should like move iterator to the next video. But there is, there is no more valuation stream.

E

Yes, it ends here and, as I said previously, so it will start to recursively come to the chain of calls, but here actually uh on the previous version of Aqua VM, uh this instruction was implicit, so this was simply satisfaction for different uh cases like cases with Spar and cases with SQL and now after the latest changes.

E

So medical fault was extended, and here we have like separate. So we have ability to uh put a gate extraction that we want to call to be called when iterator was exhausted, yep and uh instead of this implicit semantics. That depends on a uh on on a body of fault.

E

uh Then semantic uh now is explicit, and here instead of this, uh this simplistic instruction, like this explicit instruction, would be called so why it's useful it's useful, because so, first of all, it allows you to carefully set semantics of a fold, for example you could uh so uh and on Aqua, for example, it's used for uh different uh purposes uh and different work with streams, especially so we could consider streams as a stream from PI calculus yeah. That's the stream from PI calculus.

E

It should never be ended so and for example, in this case, we will put never instruction here. So if you carefully see, then this the whole, this graph would be marked as uh not ended in not not depending on the size or stream. Why? Because this never will Mark subgraph as incomplete, and if uh runs a graph of SQL, is incomplete, then the whole SQL would be incomplete and the same for this instruction at the head of this. uh This graph, sorry subgraphical fault, so, for example, this fold will never be finished.

E

Yes, it will execute again and again uh according to the aqua VM execution model. So that's why, for example, it's important- and it was the first thing. Yes, it's first thing to emulate by calculus streams. The second one is to allow uh using streams as a uh like a building block for different types and the uh Aqua side, like structures and houses, so streams internally uses for different types.

E

uh So, and uh let me say couple of words at the end uh so like according to this change, also a lot of bugs who were fixed on the aquarium side, so some of them was uh were found by uh more comprehensive tests. Some of them was uh like derived from a kind of instruction, but all of them are fixed and I'll be working on the last last bug that uh comes from a compatibility of fault and.

A

E

Instructions that you found this once and also we have some progress on the Marine side.

E

So Marine is our webassembly, a general purpose of assembly runtime- and here you could see on this picture uh the architecture of marine and we are working on so-called the refactoring that it will allow us to abstract our used runtimes and there could be middleware layer here, and it would allow us to uh change this blue boxes uh in an easy manner, and also there were some uh like other changes that won't be so I know so notable, but they're still needed like, for example, we finally update like I know a horde of crates in the Marine side, for example, gets a lid on that safe and safe code on the aquarium side, and also we have like a special form for here that uh intended for humans and allow human to read it more with less pain uh than just like this break it like syntax of here and uh like in the previous months, the main crates that allows you to transform air to this form.

E

It was uh be changed and situated now it compares to the awesome Target and uh eventually it will be integrated in the aqua run, Aqua CLI and will allow you to, for example, see uh how so actually now you could see how Aqua compares to appear in in which way, uh but after integration of this change, uh so it could be, it could be seen in like OS pin. So probably is that it on my side. Maybe you have some questions.

A

Okay, thanks Mike questions. Anybody.

A

Nobody all right, I have two questions Mike when you say unsafe code, no more. That means rust, unsafe code, right, yeah, okay, yeah, okay, all right! That's once just want to clarify that, and the second question is uh with all this work uh taking place and some of it already being rolled out. Some of it to be rolled out doesn't affect anybody's existing Aqua code. Do people what what do people need to do to prepare for that is there anything need to be done? Is it backward breaking.

E

um Honestly, I'm not sure because, like as I understand now, uh it's not a breaking change in the aqua side, but maybe admittedly I will see more about it because bitter is a developer of compiler.

A

Okay, then uh I don't know if he's still on, then, for the time being, we go with no all right, cool.

F

I'm here and yeah I can uh answer this question. uh It can be some small differences between.

E

Version that was before in aqua, but uh I hope that there will be no changes, not quite all.

A

Right most excellent, that's what we want to hear improvements with no backward breaking changes. That's that's awesome. All right! uh Anybody else have any questions for the TV.

C

C

Okay, uh my name is Joe. Okay, I have a question. Okay, um we are plenty for future development. I think we are going to build tools for uh appluence. So do you have any? Do you guys have any idea that uh what type of tools that you're gonna need, especially on the maybe application site from the development or uh anything else, but um we have anything but tools? Do you have any idea.

A

So but somebody from the tools- dxjs, anybody still down here.

A

huh I guess not uh I, don't know. Does anybody know youth.

A

Nobody got half the team and nobody knows all right. um Do me a favor put your question in uh in telegram or or Discord and we'll follow up there.