Swarm Summer School ‘21, Day 5: Demo day

hello everyone and welcome on day five of the swarm summer school today we will be a day filled with presentations by our participants of the projects which are currently being built on and around swarm i'd like to welcome our guest speakers already joined we will have a full program today packed with demos from all these exciting projects which are currently culminating on top of swarm and for that i would like to ask every participant if they have any questions with regarding these projects feel free to raise your hand during the talk or ask your questions in the q a section and on discord in the channel we have our first speaker which will be alook niruka who is working on the bfs project alek are you here yeah hey hi can you see me you're coming through loud and clear so thanks yeah so thanks i'll just share my screen confirm each of these presentations uh uh will last uh maximum 20 minutes and there's some room for questions afterwards okay take it away yeah thanks uh yeah so hi guys uh so yeah my project is basically bfs which is like a backup and recovery solution using swarm um so i'll just start with like introducing myself so i'm basically part of the b team um i've been working with b for the past five to six months before that i have some experience in like the storage and networking industries uh i'm located in india and you can reach out to me on github here so talking a bit about bfs so what i wanted to build was basically a backup solution for uh for users to sort of back up their personal data or you you could even extend this to having like a a fully deployed solution where you could sort of back up like files directories and they'll all be stored in the swarm network um the additional part that i wanted to provide was having like a snapshot functionality so if people are aware about the snapshots like a point of time image of your entire file system so with this with snapshots what you could do is you would pretty much go back in time and you know choose pick and choose what version of the files that you want to destroy the other thing would be in a backup solution is definitely you should be able to restore the snapshot and the files so that's sort of one of the important goals of this um and also provide like a simple management interface so by management i mean like uh so when i started thinking about this i modeled this around like the google drive folder or the dropbox folder that we have you just have like a simple folder on your pc you put stuff into it and sort of syncs it with the cloud from time to time right and you can access it from any other machine or give access to other people and things like that so bfs sort of it wants to be that google drive folder but for the decentralized ecosystem so why use swarms disks so so the main thing in swamps disk is content addressing so as you all know so basically a piece of data is represented by its hash of the content so it's always going to be unique right based on the content uh what this provides is like global deduplication so the background is like me being from the storage industry these are sort of some of the buzzwords that you have right like global d and things like that so like systems that implement this uh but essentially with swamp we get it out of the box because as long as you're going to store the same chunk it's going to have the same address so you know it's not going to be stored twice so it's it's sort of provides you that out of the box again another buzzword would be like incremental snapshots so what this means is uh like you keep backing up stuff right on your drive say things are changing like you're sort of editing some of the files so you don't want each snapshot to carry the entire contents of the file system right so we want only sort of the changes that are happening as part of snapshots that way they could be light and they could be easily stored and you know pushed around the network anyway so again this this is this comes out of the box because as you change files like as long as like the chunks are sort of uh they might get a change here and there but overall like if you have same like parts of data in the same file you're pretty much having the same chunks when you back it up so only the newer chunks like the newer data that you're adding will sort of be part of the snapshots the other thing is like as we all know like the bunch of talks that have happened this past week so privacy is like a very big concern to i guess most of the folks that are part of this uh you basically if you use google drive like you don't really know what is being done to your data right so you don't really have access to what they are actually doing but with swarm as we are like open source you exactly know how your data is being pushed to the network you can encrypt it and make sure that no one you don't want has access to it so things like that so all of these sort of the reason why i sort of thought of this backup use case could be like a very useful thing for us um and the other thing is the manifest so as being part of the b team so i would also like to introduce some concepts that we have in our sort of ecosystem so a manifest is just like a a data structure we use to represent like files and directories with their metadata and this sort of translates well to having a whole file system built around it so i'll probably talk about this more later yeah so being like from a back end i'm from i'm a packet engineer so you know user interfaces is uh not really something i pay heed to i would say but yeah so i wanted to be able for like this to provide some advanced functionalities of backup like all these niceties but i also wanted to make it as simple to the user and also some take some of the concepts that i've learned from my past in this industry like how to like how to set up like all these backup relationships and things right so uh but i i would introduce some concepts here which is like relationship so what i want is uh say a folder on your machine is in a relationship with some uh you know backup target which is uh you know and you sort of define policies on it so you can have like hourly daily snapshots so think about it like apple time machine or any other service that does this so you could even configure like how many snapshots and things you want to keep so this is sort of what i wanted in the application so like pretty simple just set up a relationship don't worry about it and you can sort of query information like the snapshots you like the ids of the snapshot and you should be able to restore stuff from that so this could be restored on another machine uh anywhere in the world so this is what the swamp network sort of gives us yeah and so like for me like the easiest interface for people to deal with it is have like a file system interface so people know the interface you know you have your finder you have your internet explorer basically everyone is sort of already familiar with the interface so this is what i wanted to provide like have like a simple file system interface which can even integrate to say other services going ahead so yeah so so what i i'm not sure if people are familiar with fuse but fuse i just want to introduce this because i've used this in the project so fuse is like a you can think of it like a library which enables you to build user space file system so as you see on the right like it sort of hooks into your uh like operating system like kernel space and sort of allows you to implement some of these file system operations right and that way like users can just look at it like a regular file system they won't know what is really happening underneath and so the other thing is there is a lot of support for fuse in golang so there are a bunch of libraries out there with support like which provide you the functionality and so seago fuse i just want to point this library it's an amazing library the good thing about this is it has like a lot of platform support so it sort of supports windows which is uh from a file system area people will understand the gravity of it so yeah and then also some of the concepts that uh i want to introduce are like from the b uh sort of land so b has this concept of chunk which which is like the basic unit of data for us so it's fixed size and it's basically it has an address which is sort of the hash of it so it's unique for the content uh then there is manifest which is sort of a prefix try implementation so we we sort of use this already to store file and directory metadata right and the good thing about this is it's optimized for like space saving so that's one thing that helps here then there is like splitters and joiners so these are again this is part of the bee ecosystem like how we split files into chunks all these sort of libraries are open source so people can sort of come and play around with this you know if you want to build like tightly integrated applications and then we have like the get and put chunks so as we know the b api sort of is a chunk store uh right so you can sort of get and put chunks so the different the the difference between what the api is with files is and chunks is like you sort of uh create the format beforehand so if you know how to create a chunk you can use the chunk api uh you know to interact with it um and then sort of uh build other things on in like your logic to sort of you know create these things and also there is this postage batch and stamp so i'm not sure how many are familiar but if you player played around with it basically like a way to attach value to chunks so this this is the unchained component in the so you sort of buy a batch and then you sort of upload data using that batch and that's how when people get your data they you know know how to prioritize uh and get rewarded incentivized on it okay so uh i mean this is this is just a simple as you can see very simple diagram it's just how the project is sort of structured so there is like a daemon process which sort of provides an http api right uh to interact with right now there is a cli interface but this could be extended and then there is a mounter sort of component which can be used to mount multiple fuse file systems so what i wanted to do here was you could be able to set up multiple folders so it should not just be restricted to one and each mount as you can see it's it's going to run like a process which is going to you know trigger on the fuse uh calls and then that has sort of access to a local storage right and a cron library so cron library is sort of these policies that you set up right to set up snapshot set interval so these are sort of cron jobs that will run and local storage is basically so as this is a backup use case so i i didn't want so b is not at the place where i think we can do like an active file system on it uh because uh that's why i thought like what we would have sort of a hybrid file system where you could have some sort of local storage and then a snapshot would be a place where you would push the stuff to the network right so and this mount sort of provides you the file system interface so this can be used to integrate with a bunch of different and and yeah and then when you basically do a snapshot like what is happening is it's going to be talking to a b node on on its http api right using like the chunks api and then these chunks are going to be pushed to the swamp network right based on their address basically so yeah i mean it's a as simple as i could make it but we will go through the use case uh the demo and then we'll have more yeah so i'll just move over to the demo so one more thing i think is this sort of visible i think so one more thing i wanted to introduce was this dev mode that we have now uh so it's basically just a small b uh you know node running locally not connected to the network yet this is again this is a good tool for people who want to build apps right they could see how the apis function is good for testing locally before you you know get into like blockchain and setting up everything so and then what i'll do is i'll start the bfs daemon here and so i have this uh directory here bfs right it's empty directory now what i'm going to do is i'm going to create a relationship that sort of sets up a backup uh you know relationship so okay before that i think what we need to do is we need to create a postage batch right so again this is running on the dev node so this batch is just like a dummy batch right now right this is can't be used outside of this node but yeah so for the function of this demo i think we have enough stuff yeah so there is so bfs so this is this is the cli utility so there's a command to start a d1 and there's options to like mount uh restore and like take snapshots and things so i'll go through each of them so what we'll do first is we just create a mount right yeah so and what this does is this associates this batch with this mount so whatever chunks are going to be pushed for this uh you know mount point they're going to be using this batch this is again this is sort of like you can sort of estimate you know how much gb is are you sort of looking for and create this batch um and then one more good thing is there is uh the postage stamp indexing so the newer chunks will be sort of automatically used so i mean it's sort of it's uh i don't want to go into too many details but yeah and then there is there is this local store that we have right so there are a bunch of them actually there are two of them because it's a simple get put so i've sort of created a bunch of them so badger is pretty fast so i'll use that for the demo so okay so some logs here so if you can see like all these are sort of the file system operations that are happening which are sort of implemented in the daemon all right and then you have this bff so now i can actually go into this bfs right and what is happening is as as i go in there are like all these commands are sort of getting executed here right i can create a directory say let's create d1 and if this is visible and now just create another directory right and i'll also create we can also cd into the directory right or create a file say 10 mb right then so you see badger is pretty fast so the writing is pretty yeah so if you see we created this structure now we have directories then in the directories we have files uh we have a bunch of metadata here even the timestamps are there so all this is sort of the file system metadata and all of this is sort of going into this fuse implementation which is writing things on the local storage so it's being written in the local storage as you would have it in a on a b note but not actually pushing it right so now say right now what has happened is i haven't set up a policy so if i see right now so that i can list some exist amounts right so yeah if you see that the default is daily so this is sort of going to run every day and create a snapshot right and then keep count to five so it will sort of keep five snapshots so your entire week worth of stuff right so this is there by default you can also change these things right and then there's also a way to force creation of a snapshot so i'll just do that right now right so okay so we've sort of created a snapshot just let's look at it now right so right so we have the snapshot this is the reference now what this is is is basically a manifest which is uh which contains the entire file system info so it has all the metadata stored inside it and it also what it has done is it has created the files made into chunks and also stored the root reference in the manifest and it sort of backed up the whole manifest to b so if you see now here it sort of uh sends some chunks here right on people this is that mode but yeah but and this is sort of the manifest like i've just kept it in log so people can see like how this looks so you have your root and then you have the directories and then these entries are basically swarm addresses and this is sort of the file system metadata this is sort of getting pushed to b right so okay so now what we'll do is we just add another directory right so let me just create a d3 uh then go to d3 and create a file it's a and you can always like repeat this process right and create more snapshots and then you can see each snapshot and the synth here right so what this does is this will also query the b node and see how many chunks are synced to the network right now as this is uh running the dev mode this is just saying that okay everything is synced sort of but yeah it's in this this sort of creates a state in the b and like keeps track of the backups so you can actually see how much of it is synced and things all right now okay so now i have this snapshot right what can i do with this so let me ah so okay one more thing i can show you guys so there is this uh file browser so you can always see this in your finder right so so if you see this right so this actually comes into your finder right and you can actually even browse this on your machine so this this looks like an exact directory for the system and then i also wanted to show there are like few apps there which uh you know you can sort of expose this local uh disk onto your network like on your machine as a so as http so you can even provide like access so this is like open source project file browser um it can basically show you you can even upload stuff on it so alright so yeah you can see it here but all this is being run on the fuse mount right so if i go here now and i'll just stop it right now just this image is still going to be here right so this is all going to be there and then you can always take a snapshot of it in case okay so now what i'll show is uh the restore functionality so right so you should be able to you have the snapshot right so what i can do with it is there is a restore there are two options you can either restore like a single file from a snapshot right you say one this particular movie that you had at whatever time right then you deleted it or whatever so you can always have that one file back or you could just restore the whole snapshot right i'll show is and you can also provide a directory for it to restore to but the defaults will sort of create a tar here right and you can so if you see this is from the first snapshot so we had only d1 and d2 right and similarly we can yeah see that all these chunks are getting retrieved from the b note actually so these are all pushed by your bffs right yeah so uh that's and yeah we can even see say so now we have uh you can do this so i'll just show you the file link store and i have a third file in the second snapshot so i can just say yeah so it just restored this f3 here right so all this is again coming from the b node here and what it's doing is it's actually just looking up this manifest so it doesn't really need the whole info beforehand so it's manifest has all the information to look up each file and get that particular reference and you know pull it back from the network yeah so i think uh pretty much wraps the demo um i think you can go back to yeah so uh some future work right so basically so backup is not really an ideal use case for b uh because like the network sort of prefers like popular data right so and your backups are not really that popular you're going to be storing it and not looking at it and things so the but what what i want is like since the disk is going to improve our time right like there might be a time where we could actually just have this as like an active file system right we just um and we can even use like other features that we have in b to you know sort of communicate like new stuff being written uh to other places where it's mounted and sort of update them and things right so all this is possible uh once things improve right now i still like that there is like use cases for it like people can always uh use it as a uh backup uh you know if they want even want to share files or things with others they can use it and i would love some feedback uh also and we also think of integrating it with fair os so i'm not sure if you are familiar you guys think there was a fair drive presentation day before so so there are things that are being planned here as well um also i'm looking for more people to you know get involved like this is a good place to start uh you know learning about different concepts around b so like although we wanted to be you know transparent to people but it's always good for people to know stuff and you know even make improvements if need be so um and yeah one more thing is uh i was thinking of making we can sort of use this as a docker registry for like having your is informed right because like you can have a docker local registry over a file system right and that could be this bfs which is backing up to swarm network so this is also something i plan to try out so let's see how this uh things go okay thank you okay i think yeah yeah so yeah um we have some questions uh coming in i'm gonna repeat them one is from the q a section uh from um i definitely b and he's asked uh first of all he likes where you're going and he will check out your github repository uh but he's wondering have you done any testing with large collections via the b manifest and he gives an example about a collection of 22 gigabytes and he says yeah large files don't see anybody issue but large falcons could be a problem yes uh that's true like uh yeah there are like limitations with the manifest uh definitely but yeah we this is this is these are the things that we want to improve right so if we have some need for it i think people are going to step up and improve things right i haven't personally tested this these large collections because this is this was more of you know a side project i was working on but i i think how the ecosystem is developing and i think this is sort of important to get out there and like have more people work on it and i i do see like these numbers like will cause issues right so because right now i think the manifests entirely in ram as you said so yeah but it's more of opening up to people i love it if people can you know come and solve these problems okay wonderful i uh yeah i i really like the functionality uh i like the apple time machine function as you just described it uh and it looks like it's has the ease of use of being combined with like a feature like mega i used to have but then open source and distribute it but what more is that the like yeah yeah okay um thank you for your presentation and uh if i know further questions i'd like to invite our next speaker which would be dimitri paltrow who's working on the deplots project and one of the projects within the dappers project is the swarm search project which he will demo dimitri are you here are you online yes i'm here should i whoops i would like to switch off my video and share the screen is it okay sure go ahead uh do you see my screen get a hearing yep coming through loud and clear and the screen looks good okay screen is good as well so uh i would like to present you our project uh dotnets project and some kind of sub project we have implemented inside of so what is our mission uh and what we are doing as a doubles project uh we are building an augmentation augmented web and our mission is to provide you with a tool set to shape your own information ecosystem and we can also enable fundamentally new workflows and popular websites and essentially we provide them three technologies for legacy web so essentially we are decisionalizing uh ux so how it works in general so if you have some web layer as you know it so you know you see the dog on the lowest layer we put some kind of adapter on the top of it this adapter defines visual elements and also this adapter makes some kind of passing of visual content website content and pass into some kind of semantic tree yes we have some uh a question that there's some background noise where you are could that be the case this may be so i can change my location because i'm on the road better this is better yeah much better okay not yet okay so then uh that uploaded an application on the top of it use this visual audience like mustaches and then what you see what you become in the augmented uh dog and augmented content so why do we need strong because our platform is mean to be decentralized and unstoppable and we do things that probably not everyone will like we need uh between decentralized and unstoppable and we need to store our modules our applications in this way and this is not stoppable and that's why we need swarm uh to make it um yes and that's why we use to store our data but not only we need small although we hope that we can create a huge added value for this one and that's why we have introduced slam search well swan search is an opportunity to inject the results of search in slider clock inject this the results of the search results into the website where the user makes his search for example this is youtube so and then for example if you make a search we can start our application we can create we switch on swap search and then our plugin our doublet injects the results of search in the swamp from the swamp into the youtube so that the user who performed the search on the youtube uh will see not only youtube videos but the content which is usually stored we just start on uh swarm as well no this how it works uh but how we can find these videos essentially uh we need to index these videos with videx metadata to this videos and we have augmented swarm gateway yes we have just augmented swarm gateway to collect metadata and to index it now we can just see how it works so this is usually small between the space but there is a doublet running on the top of it that goes inside the swarm with ray and collects metadata from the user about about the file he uploads now you see this droplet a smart index running on the top of uh gateway and this is how the applets gets injected out of copyrights gets embedded into the strong gateway we collect metadata from the user which is currently not the functionality of the strong gateway but we need it so that's why we have extended the functionality of smart company in this way and this is essentially the way how we extend the functionality of any website so then we have uploaded this metadata and then if you if we make a search on google search then we can see that some google search results were augmented with swarm search so essentially we have injected on the fly the results of this research into the google search this is how it works essentially so oh you see this icon on the top of google search means it's not the native google search results but it's injected on the fly from the small how it works what the architecture of the sound search the uploader uh uploads his file and inclusive metadata into this rom gateway and then we have an index adapters that uh collects the metadata and uh pushed into this over the search api and it stored into the elasticsearch engine and then after the elasticsearch engine indexes this metadata it is possible to make a query and get results from from this index and based on this index we can get results uh from the search and inject them into the youtube so this is what the searchers do uh you can see that this elastic search is not really the centralized system some kind of mop and we hope uh it will be able whether we will be able to implement a decentralized a truly decentralized search uh on the top of uh swarm and uh this will be the next step elastic search is a mock and we will change it and we hope to change it together with the swan team another interesting part is the search api and there is this open standard open search the standard and we suggest swamp team to implement open search api on the top of the swarm and once it gets done it becomes possible to for any website for any browser user uh to define swamp search at the standard search engine in their website in the web browser like google it's not about doublet functionality anymore but it's just something that i suppose would be very useful for the smallmouth so once more about uh components in the spam search architecture there is an indexed doublet that collect metadata and send it to the indexer the search tablet essentially makes requests to the back end injects such results into third party third-party websites like and search api is essentially a currently open search api and the server implementing it transforms the search to their request to the back end and make essential research uh the elasticsearch is an indexer and it's a centralized smoke currently we need to replace it with a decentralized implementation on the top of swarm and yes we have augmented swarm with way to collect the metadata but in the very similar way we can augment any upload website so we can just intercept any upload and collect metadata and the file from the website yes and b notes so that is if you share a values let's talk about cooperation and feel free to visit and subscribe to our news and let's talk um about how we can decentralize the world on the visual layer and this is that we are currently doing oh that is thank you dmitry what are you what are you currently looking for you're looking for additional developers on this project or projects to integrate it uh if i understood it correctly i ask you about what we are looking for yeah yes uh in this project in this one search project we are looking for um for support uh how we can implement the indexer on the top of this wall in decentralized tv because currently it's a mock server it's full functionally so but it's centralized it's just one server running elasticsearch it's not really what we would like to do just a test of it's a test suspension but we believe that it's possible to in to implement uh this functionality on the public school but the question would help us to do it [Music] and then another question we got was which browsers do you support which browsers so currently we uh test it in our plugin in chrome but essentially it [Music] was developed to run in major browsers firefox as well in opera but okay to be honest we don't test it every day maybe it's got broken because not released okay yes and uh perhaps redundant but uh if people want their files not to be indexed they shouldn't use the swarm gateway in that case sorry if people would prefer their files not to be indexed by the application by swarm search then you advise them not to use the swarm gateway okay we need to collect metadata in [Music] any other website that we we just we just create an interceptor to show how it works we can create an applet intercepting uh swarm uh gateway uploads just to show that how how we think about websites we just we don't ask to the website holder we do just uh create a doublet and uh my arm in some sense uh break into the functionality and extend it so this is what we did with the swamp search we have just uh in extended uh swarm gateway functionality to connect the metadata thank you thank you it's uh intriguing because it is imagine it could be hooked up eventually to other decentralized networks as well perhaps over arching search um we've got a question from uh victor and does the gateway augmentation injects a form to the page and send its form data directly to the indexer i'm sorry i can't hear you clearly certificated sorry let me repeat that does the gateway augmentation inject a form to the page and send its form data directly to the indexer here's the question it's yeah this index currently uh the index is currently the server and yes the dab let's collect the metadata and send it to indexer and index make it possible to map the results of the search to their swan cache can we answer the question uh yes okay uh there are no further questions thank you for your presentation good luck with the project i would like to invite uh today fuse if he is available hey are you live yes hi to present the uh energon x project which is currently being built on swarm um today do you have a presentation or a video you'd like to share okay okay see it yeah so it looks good so what we are trying to do is uh build a marketplace for hydrogen producers and hydrogen service providers and of course consumers so they can offer and find green hydrogen mostly in hydrogen on the market so we believe that uh the big energy transition is coming and it's already actually happening and we believe that uh in near future the the energy market will change drastically because of the hydrogen and how the hydrogen is actually can be produced uh so these are the fields where hydrogen can be used is already used or will be used some of these fields will actually be profitable some will be profitable because of the co2 to reduction of emissions uh and some uh will be uh non-profitable right for some we already know which will work good for some for some uh there's still a lot of research need need to be done but basically as you can see it goes any eat it it will influence majority if not all of the energy markets that are now dominated by oil and gas industry so where are we now we are a startup uh we are trying to solve this convertibility of the green hydrogen we use swarm for immutable storage of so we track everything right uh and uh we we publish all the cancellation statements and swarm uh we are looking for funding of course and looking for partners namely the liquidity providers the hydrogen producers of course the consumers and other service providers from the hydrogen field and of course we are looking for the developers to help us build the product and expand it even more what are the challenges so the challenges the market is mostly wholesale limited which means that the power purchase agreements are mostly done for longer periods of time and for fixed cost and price and it's very difficult for small producers to to to enter the market or for small consumers to to buy small quantities of of this commodity uh we want uh to establish access to large-scale producers so you can as a small user you can get smart small small quantities and if you're a small producer that you can deliver your your product into local storages for a normal or competitive price what we also aim to is to reduce the high operational costs because right now if uh if you build like a big hydrogen and plant you know production plant it requires uh quite a lot of legal expenses and other costs for stakeholders and we aim to to reduce this part of the operational costs and of course we want to change how the energy market is working because we don't want that the same thing is happening as is happening in oil and gas industry where the speculation and the manipulation of the market influences let's say whole civilization the market is huge these are the estimates from the european investment bank what we are building now and what we are working on is a allocational trading platform uh marketplace uh for for the hydrogen we also have the other different products that can be like a side product from the from the electrolysis or the pyrolysis or or recuperation or something uh it's uh very inclusive you don't have to pay a lot of money to to get in uh basically there's just like a need for the verification of the of the parties that are on the on the platform and that's it so there is no no there are no fees connected with the with the if you want to use the market because for example if you want to to to to do electricity exchange or if you want to do the the the gas uh buying directly from the producer you need to pay like huge fees for for just to to to be part of the the network we promote regional connectivity uh and uh locality of the of the of the um users and it's completely transparent and traceable and there's like no middlemen in between we connect consumers and producers and of course distributors and as i said we reduce operational costs and this will make and makes the hydrogen much more accessible so we still have uh quite some problems uh regarding legal framework and and you know the guarantees of origin and all this stuff still waiting for some eu commission what will happen on that part and how the certificates will go and all this stuff uh we did something on our own but still we will have to wait for this to become the part of the agencies for the energy from each state that will be part of this energy transition uh it's uh for now it's unique we didn't saw any other solution like this uh that's so inclusive uh and to be like a market enabler that enables also other you know like smaller parties to be part of this energy transition so where it can be used and how the the value chain goes uh so the gray hydrogen is where you produce a lot of co2 right and the blue hydrogen is uh so the majority of the hydrogen right now is actually blue hydrogen it's like so the hydrogen is like a by-product from some chemical reaction or something right uh and then you have like a green hydrogen which is completely renewable it comes from either wind or photovoltaics or some other renewable parsers so how the rollout is happening so immediate next steps now it's going into the transportation and light duty passenger vehicles and of course distributed power and green balancing and all this is already happening and diversification in till 2030 and 2030 and beyond it's very ambitious but the hydrogen will be mostly or let's say including inclusive in the industrial world and the mobility and transportation world and and heating and of course green balancing so the market as a set is huge we are not in it only for the market but we really believe that if we do a system that can trace everything from the energy renewable energy source to the supplier of the power for the electrolyzer and tracking the production of the electrolyzers and then storing and calculating all the capex for the storage and then of course the distribution and the fueling uh or consumption we can say consumption of the hydrogen if we can track all these then we can do uh lots of other stuff like analytics and of course prediction markets and stuff like that uh so where are we now so we have like orders we have auctions we have distribution and we have tracking of the distribution of the packages and stuff it's a hybrid uterus centralized architecture we use ethereum for for for value transfer we use swarm to store immutable data and of course everything is ready for tokenization why swarm because we know exactly who can write and who wrote something into it and the immutable records in this let's say segment industry segment is uh very important and uh with this you get this trustworthiness that's otherwise quite difficult to to to reassure you know your clients that your records are truly truly immutable uh because we are location aware we can do like transport route optimizations we have some apis ready some are still in the works for third-party integrations especially on on the part of logistics and statistics and analytics and all this stuff it reduces like a lot of risk management and uh of course it it it increases the usage uh to the to the consumer um so our our vision and plans so we are mostly ready with the marketplace except the the the capex calculation for the for the for the storage this is one of the things uh we plan to roll out the the the first phase of the marketplace in 2023 the start of it and the fully functional location we're trading in 2026 so at that time we will have the whole chain from smart metering of the electricity production uh to the automatic rate execution to the calculation of all the storage crafts so the transmission the settlement the insurance the reporting all the parts of the of the order and tracking inside we will have the option to deploy like a local autonomous marketplace that is not connected to the global part of the global network so there will be like countless small little uh autonomous marketplaces so basically tracking from the production to the consumption uh with the automated billing and micro payments uh of course the market intelligence on top of it and service providers and of course third party integrations so right now we are working and trying to find uh hydrogen production facilities from some we already get that got the interest that they are interested to to offer the hydrogen on our platform we are looking for equipment companies and we also have some partnerships with them also and of course the distribution companies uh and the companies and the partners to that drive adoption of green hydrogen uh we plan to incentivize the local production of the green hydrogen this will require uh being compliant with the financial loss but this is also something that we are looking for because the the end game is actually the commodity exchange for the hydrogen green hydrogen and then by the end of 2030 uh when the broad rollout of the hydrogen will be available we are hoping and targeting that the marketplace will become completely trustless and autonomous and decentralized so from that point on we just tackle and do the data management on top of it because there's like enormous amounts of data that that happens from this uh from all those deals and working on prediction markets and of course trying to solve you know the combinatorial problems when you start to do uh orders when you have multiple sources and multiple destinations and this this the complexity of these combinatorial problems problems just explode so after the the the central decentralized system is established this is where we will focus um our energy on okay so the revenue streams are coming from the marketplace uh something from withdrawals right now we have crypto only but we need to build the the aerodenominated market for the hydrogen we are selling the transportation information so the distribution companies actually option for the for the for the distribution uh and we have like a hydrogen index we can we can have like a global hydrogen index and we have like a local hydrogen index from the different like i said uh autonomous marketplaces uh of course we we we do we have some of the market analysis uh this is also like that as a service being sold we charge for the unification systems and of course for the procurement network so the in the future what we are looking for is also the income from the validation validation services you know from so you know that the manufacturing equipment is being built by uh trust uh uh trusts were trustworthy um so the not the trustworthy by the trustee parties by the trusted parties and the the they validate all the equipment and they vouch for it so the reputation is also being included in the system and then of course the issue of the certificates and tracking of the and of course become the exchange with the term contracts and of course for futures trading so these are let's say future goals what we see with this is of course a lot of growth and of course of jobs most importantly reduction of the co2 so something for the future of our next generation and the generation after them we believe that the climate neutrality for these to to to be achieved we have to play part of it we all have to play part of it and of course we we think that this will bring uh broader social and uh environmental and economic uh change you know and uh and uh and of course with the adoption of the clean energy we believe that the the the the world will be a better place so so what are we now where are we now we are part of european clean hydrogen alliance we receive the sworn grant we have some interest from uh equipment manufacturers and of course the hydrogen producers and we have uh operational prototype so i will do the demo now if you have any questions we can do this so this is the the the the parties involved okay so just log in uh so maybe maybe i'll start this so who are the users as i said the producers of the hydrogen the distributors the consumer the buyer and then the analytics guys that take the data from the platform and do the analytics and sell it back to the to the users and then only the providers of the data that can provide their data and the procurement network is imagine like amazon for for equipment manufacturers or the buyers that already have established like electrolyzers or production equipment or something like that uh so yeah so we have like a a marketplace you have different uh hydrogen so we have green blue gray white turkeys so all these depends how much co2 you have as a side product or how the how the process goes we also added ammonia which is much easier to be transported and then we added oxygen because after the analysis we found out that there's like a lot of oxygen that can be used and of course sold and carbon carbon is a by-product if you do the pyrolysis for the hydrogen there's a lot of carbon that can be used as a also again for the fertilizer or producing nanotubes and stuff like that so there are different like qualities of the of the hydrogen for example green gaseous and then the quality is 99.99 and then we have a liquid hydrogen again with different qualities so so this is the it's like a double auction system with a little bit of uh the turnaround trick is uh so everything is uh geo-located so you know who is the closest one to whom uh and you can publish like a spot price and the buyer will buy the closest one or the most appropriate one then we have like an auctions when the producers can just say okay i'm willing to sell two thousand kilograms the next 10 hours and then declare the auction and then other customers can beat for it so the difference between the auction and and the marketplace is that you can only buy the whole plot and then of course we have the distribution where distributors beat for the for the distribution and uh um then after delivered so in the background it's actually like uh it's like you make uh escrow accounts or the the the the distributor until it delivers uh he actually owns this uh dispatch or the lot that needs to be delivered right so the the usability of this uh delivery of the hydrogen is up to 450 400 maybe 500 kilometers right so that's a local local distribution and not like long term for those the pipelines will be used so so we have like vendors okay so i bought some before uh and i will show you so so the difference between the commodity exchange in the marketplace what we are doing is you actually know from whom you bought so there is no like middleman in the between you actually bought this from the from the from the party and you know from which party you bought this right so what we are storing on on the swarm is actually this information along with uh cancellation statements so we know which transaction and which transceiver gate from which certificate to reach the certificate this was executed and all this stuff is then stored on swarm and it's like this you can you can think of it as a recipe or it's a cancellation statement yeah but you know from where the energy came what was the device that produced this and from whom you bought and of course uh all the the the details of the order that you did so we also have some of the statistics we have like an index and then you can see the market and you know exactly who bought any something from the car country or this per country or prices per country and all this kind of stuff that you can calculate because we have like a geolocated market so yeah i think that this concludes my short demo and presentation if you have any questions thank you today i love how this is uh a project which has a real world impact and has built its premise on top of this new uh infrastructure um let's see i was wondering you mentioned the aspects like location aware trading how are you solving the the oracle problems do you have like when you work with sensors on these uh oh so yeah so weapons so the the centralized part is uh so what it does so the majority of the centralized parties so we have one thing called like multi-transport protocol layer inside and this protocol layer is actually a thing that you can embed into like for example into the nozzle or into the electrolyzer right so so the user interface that i showed you this is actually in 10 years time it won't be used because the machines will do all the trading so if you imagine having like a nozzle at the the fueling station and the refueling station is running low on the hydrogen and the machine says oh i'm running low and it just goes from the marketplace and says deliver me 300 kilos and the trade is executed and the truck is loaded and the truck delivers these 300 kilos to the refueling station and that's it and the same way goes from from the nozzle and or electrolyzer right so you know that's why the analytics part is so important because if if you do proper analytics like you know the prediction markets and you know the weather is good and i'm uh i have like a photovoltaics electrical source and uh my weather today is bad so maybe i will just buy the hydrogen or do my weather today is good and the neighbors uh weather is bad or whatever so i can produce i should produce more and not sell the electricity directly to the to the to the green but i should produce the the hydrogen right and all this stuff will influence and basically be done completely automatic by the machines right so the electrolyzer will say i don't need to produce because it's better if i don't produce or i have enough or i have not enough you know on all this uh um all this uh you know the equilibrium will be always be trying to reach right and the the marketplace just like enables all this stuff to do analytics and predictions on top of this and especially if it's like a global or let's say like a wider range then you can do a lot of the optimizations regarding the production or you are sending into greed and stuff like that so yeah let's see we have a question from true [Music] will your marketplace add any other additional energy products in the future perhaps related to hydrogen well but the the energy product you mean like the electricity because you know the hydrogen is actually not the energy product it's like a compound chemical chemical compound you know it's like a vector for the energy still it's a compound that you need to change and then you get the electricity up right so you need a building block and stuff like that so not directly the energy like electricity no no we won't we won't do that but as i said as i've shown uh the parts or the the the the different compounds that are connected with the hydrogen production or transmission those we will include and we already have like i said carbon and oxygen and this is like a quite um a big part of it right and very useful there the carbon the the black carbon is useful in a lot of parts and it it is it has its value right the question from church was will hydrogen be the fuel of the future for vehicles or electricity well i think there will be there will all be there will be electric cars right and the range of them will be you know to go for the shop or something like that the heavy duty vehicles and trains where you don't have like electricity the trains will go on the hydrogen if you don't have the the electricity the long long vehicles uh you know it's it's uh it's not useful to have four tons of batteries and you your vehicle is 20 tons right so that that that will be taken by the hydrogen also and uh so i think that also like light trucks up to range of 500 kilometers will also be on the hydrogen yeah but there will be like cars like personal cars with electricity yeah and another question was uh do you now have the the data for it of the the movement which you track and place on swarm will you in the future also place potential other data like analytics once one so so the bar that we are developing now is uh like some kind of a data place where you can ask for this data from the marketplace and get it out and uh yeah that's the idea that goes into onto this world too but we are not there yet yeah but uh all the daily data the indexes data and assessments data they will have to be stored somewhere so we are targeting for the for the swarm can the geodata be linked to openstreetmaps to be part of a global geodatabase well right now we [Music] we use like um only like a vector data and we only have like support for for eu and the the optimization root of optimization we are using uh third-party solutions for it probably in the future somebody will be will come and say ooh look we have the better or cheaper or whatever and they will integrate their part into it okay um about the project what stage would you say the project's in right now just uh so yeah so we have the working prototype we are trying to find the solution regarding legal frameworks uh because it's still quite difficult you know to so every country in eu has different uh regulatory requirements for the transportation of the hydrogen for the production of the hydrogen for the usage of the hydrogen so this is something we are waiting for uh but our plan is to roll out the marketplace with the discovery for the hydrogen and especially the green hydrogen in 2023 at the start of it so this will be without the distribution part with other options for the distribution uh this is it it will be like uh so the buyer has to to arrange the the delivery of the hydrogen yeah i noticed you mentioned you were looking for funding did you apply for eu funding oh yeah yeah yeah so we are in the process for the eic okay and my final question would be i know you're a and as you mentioned you were a part of the um swarm uh grand wave you received a grant from swarm earlier on uh how has this helped you during your development how has swarm been helpful in this development of your project well it's hard to accept a lot except a lot we had some problems regarding the documentation and all this stuff but majority very good very good results very good communication [Music] whatever i needed i got the answers immediately so all good i would advise everybody that uh that are trying to [Music] to to to work on swarm and use the swarm uh to to apply for a grant and that's uh that's something i think that the community is good people are great and i would advise everybody and yeah the more people work on sworn the matter for the ecosystem thank you i have nothing to add to that okay i would uh thanks again i would like to ask uh our next presenter to uh uh to appear to petra wilson petra are you live hey hi i gotta figure out how to share my screen here just really quick i'm not sure [Music] i have in my notes you'll be talking about interoperable standards for data process flows yeah so what i want to talk about actually is kind of uh math um yeah so uh it it really starts out with um this concept of the minimal case and so we have different kinds of schema kinds of relationships um you can have just a simple linear relationship a scale if you will and that's a one-dimensional system is there a name for this project by the way could you describe it uh this is the fundamental basis for the interoperable schema that i'm developing uh and the interoperable contracts that i'm developing for microsparc fate which is a media distribution platform thank you um so there's lots of different ways to categorize data and we have data itself and then relationships between data and these are arranged as nodes in most information systems and edges between the nodes but there's this concept in mathematics called the dual in graph theory and in graph theory the dual of a node is its edge or is an edge in a graph right so um for example if we had a set of just these two points then we could construct two edges and connect them and that would form a dual um so a little bit of this is uh i'm gonna try and move as quickly as possible through a lot of math and just kind of point out the relationships that we're looking for the idea is here that these are the minimal subset structures within a structure that has these numbers of degrees of freedom okay within the information set and so we find subschemas that can be used as the minimal element to produce a schema that has three dimensions for example you need one point and a triangle as a repetition of those shapes will get you um to the minimal shape in three dimensions uh the minimal polyhedron uh in three dimensions which is um a tetrahedron and in four dimensions um we come across uh a new shape uh this uh octahedron the regular octahedron um when we put all of these shapes together we get something called a tesseract this is also the projection of orthogonal relationships between entities in a graph and so the tester act in four dimensions is like a cube it is the minimal set of unitary relationships of orthogonality within the grid that covers every uh point in a four-dimensional space or a four-dimensional graph which is not the same as a space per se um anyway when we combine all of those things together it forms the tesseract and the tesseract is interesting because it also demonstrates some of these uh qualities of pascal's triangle where um each node can be represented as a specific binary within the binary pascals triangle and the sums of those binaries actually create the logical operators that we know and love from coding and we can actually apply those not just on binary sets but on any any base in fact when we do it with hexadec hexadecimal bases it resolves to the actual colors and these colors follow the same color theory where um white is red minus green and blue or where black is the sum of cayenne yellow and magenta and these actually work out mathematically the same way so when we're building structured data that's based off of the hexadecimal pattern we can bifurcate those decimals using an operator called xor which is this operation here uh in order to bifurcate those streams into shapes that will conform to this shape and of course we can do this in higher dimensions i'm focusing on four dimensions because this is the relationships the nature of the relationships in most data models that we have in the world today so in the world today most data models are based off of a user sort of the source and um and the data sync uh where that data sent um or i'm sorry i think it's the other way around the user being the sink the data coming from the system being the source but within that system we can split it up there is a process that occurs over time and storage which is an embedding of the data in some kind of medium spatial relationships um the interplay of these spatial relationships over time actually form the inner operators so this is spatial relationships these are temporal relationships within the schema um this is the user this is the system uh source this is the sink and these inner relationships actually form logical operators that we can perform directly off hatch and we can integrate over those hashes and define a volume within this space so it's a visual way of representing sums of schema over uh integrated over uh this minimal case of schema for a four entity for relationship for entities for relationships um when we decompose the hypersphere our i'm sorry not the hype the hypercube the tesseract uh it's actually composed again as i was saying of these tetrahedrons and so actually the minimal subset of the minimal schema in four dimensions is the tetrahedron from three dimensions and so the idea is to construct a schema that will follow a tetrahedral pattern and that this would be the minimal case for interoperability between cells in a schema subschemas within a larger schema to allow us to do this integration um fractally right so that we can cover the plane um and fit the most information into the smallest amount of bits this is the idea and so when we do that we have these four basic arrays which are all tied to the inner array in the dual and i just combined those into this arrangement here and so this is just any color directed acrylic graph and the position in it uh and defines the context of the document this is a did and it did is just an address and we have versions of the same did this defines the volume of the space in the center of the tetrahedron there's a time relation and the actual data itself the content contained inside of the object that is being stored within the minimal schema and then there are the relationships some citation or dependency of the code that's contained in the hash and the authorizers or signers of that code or information or data and then in addition to the idea of differentiation there's the idea of composition and so um anywhere where this information is cited as a part or or another composite contains so this is a contains relationship and or is contained by and this is a contains in relationship and those are the minimal sets of relationships that we'll need in order to compose subsets of text super sets of text but maintain the relationships between the text and the origination the user source of those texts and the agreements that bring together the combination of those texts for representation and transmission um again this is all based off of some research that i worked with um uh with michael zarghum um and he's a professor out of mit um he wrote a paper with uh michael stevenson color directed just a cycle are acyclic graphs for multi-class attribution networks um this all has to do with uh again these concepts of fractal integrations and the idea that we want to be able to create uh quasi crystals and the reason for that is because a we want to fit the minimum uh i'm sorry the largest amount of information in the minimum set of bits but we also need a architecture that is going to maintain enough symmetry to be parsable uh over the maximal number of cases and so actually these tetrahedrons add up across um this is a e8 symmetry group and if we cross the e8 with itself uh we get essentially the bhagwa if anyone knows what that is and these problems are all based on again just the idea of close packing of spheres in a space it's exactly the same problem right and because of the nature of numbers found in the pascal's triangle there's a predictable pattern by which we can index these relationships and when we lay out these cells in sequence in series they tile the plane and so this is very pretty as well especially if you consider that um each one of these classes these are classes right a refined class and as long as we have a minimal set of classes that will fit within this geometry then we know we have the minimal set of relationships necessary to relate the system that has four entities and four processes yeah so this is an approach um right to developing schema and again this is a generalized information theory so uh in fact it doesn't have to apply for example to a media source or media storage or media processing or the entities that produce those things or the media itself but could be applied to any number of different uh processes or entities and uh the process that's being developed here is done across four and i do this for a lot of different reasons uh these quasi crystals in four dimensions are becoming incredibly useful in a lot of different applications across industry this has to do with protein folding for example or autonomous control of uh uh multiple uh devices across a grid network uh so if you want those those drones that fly around uh in patterns in this in space you know and they're all adjusting to the same wind currents and and simultaneously sort of thing the way that you manage that again is with these sort of fractal distribution grids uh in the nodes and they play out in these patterns but you can actually go beyond four dimensions into five and six and on uh all the way up to uh the quasi chris uh quasi-crystalline spin network uh it turns out that with this level of resolution right because every dimension adds an additional degree of freedom and therefore increases the resolution capable of being stored within the schema right so um at this level of resolution uh we can model quantum chromodynamics and this will be the method for construction of quantum nodes for quantum computers that are stable because they evolve and form these quasi-crystalline uh structures um so it's it's not just an approach um that will work uh at 4096 kb um in a four-dimensional system uh but uh for for video resolutions in in color uh combinations uh six-digit hexadecimal parsed uh digits right um but actually can be expanded uh the same way to derive these kinds of quasi crystals uh to the maximum resolution we need in our universe anyway that's that's pretty much the extent of the um presentation of the ideas here uh that that i'm exploring and uh or have been exploring for the last 20 years and that i'm now building into these schemas for representation uh i would think a user interface that works in four dimensions is something that we really need to establish uh because we are using four-dimensional information systems at this point uh most ai ml and other uh systems are running off of these kinds of degrees of freedom models um yeah yep so that's the the idea of future applications of this thank you petra very uh very insightful i um see check if we have recommendations or questions i uh there we are i wonder if we have participants who could uh respond in this way uh we do have a question from uh again from edef and b um i confess that i'm not this theoretical but is there any application of this to resolving version feeds in this form so actually i definitely you had mentioned in your um comment on uh i think it's a develop swarm uh channel in the discord uh that you were trying to set it up so that you could go back and insert records into feeds um i think this is this is kind of the problem right that we run into and the idea is a way to create schema that will um resolve uh a way to insert and and keep the and keep the schema intact i guess and in order to do that um it's necessary to classify uh but once you have uh the sequences that are produced out of these uh symmetries are predictable they're algorithmic right so one could theoretically reserve a sufficient space for the depth of resolution that needed to be formed within an unknown set of variables and then that would leave enough space uh to fill in but there's there's there's a requirement here that you have to define the bound because when you're dealing with non-abelian groups uh discrete groups uh they have to be finite uh unfortunately otherwise you resolve to there's a paradox of ambiguity is the term if we go back to pascal's triangle and the binary development of it then yeah um within the binary of pascal's triangle uh let's see we set up the grid we have oh yeah yeah as as we sum off of each row we end up with the fibonacci series and this is like an infinite series but it's predictable so if we know the depth that we want to arrive at um then we can cut it off at a certain point and build a geometry out of that but uh if we don't know the depth then it's just an infinite series right um and eventually it resolves to ambiguity because in the construction here when we add the set containing the empty set there's two ways that we can differentiate the dual but when we decompose the set containing the empty set we always get the set containing the empty set we don't get the empty set and this ambiguity paradox uh is uh part of the basics of of whether a problem can be considered mp complete or um or hard uh or whether or not it's trivial um so there are some fundamental limits in computing that require us to have a defined finite set but once we define that finite set then if we know our resolution uh we can define that minimal case and in this case uh in swarm we have 4 096 uh bytes right which is a finite number of bits a finite number of rows in the triangle so we know what the geometry is again i'm not necessarily saying that this particular geometry is the one that you want to build out of the geometries that are available at 4096 bytes but um this is the simplest one in four dimensions there's not necessarily a reason that you want the simplest one you might want a more complex geometry because it will take more time for an opponent for example to decode but theoretically this can be read directly from hash it doesn't really have to be decoded because it's constructed of hash uh it's like we think a lot about security uh in the web but i'm actually trying to construct uh things that will be just visible if we look at nfts for example this is just a series of color codes in a frame 256 by 256 those are all hash so we've got hashes of hashes it's all predictable right if we know the parsing length then we can xor through the parking length and derive whatever the hash is but we need to know the passive parsing lengths and we need to know uh the integer space so if we know that the interface integer space is six hexadecimal numbers then that's the maximum range then um we can parse out that distance in within the hash and decode it but that's if somebody would want to find out what picture was stored in a hash right like other kinds of information won't resolve to a color space they will need a different schema on the other hand if you choose a schema that follows this then it will be decodable that is certain thank you for confirming the uh that i caught the gist of where you were headed with all this i'm glad there are theoreticians like you out there so that practitioners have can springboard off of your coattails i mean honestly i'm springboarding off of a very brilliant man's coattails and this concept of the document id uh in ipfs was was not even his uh so there's a lot of brilliant people who are building uh in this way or or thinking in this way and uh if you really want to learn more about this i would suggest to look up uh michael zarghum and especially this paper on color directed at cyclographs because i think if you think about blockchain and data sharing and data applications in blockchain cdags are where it's at most everything in a merkle tree is a dag so once we get colorization of those dags then we discover the holes in our dags which are considered this idea of clicks or it's any set of functional relationships between notes which are unique between those nodes a click so it's like if you have a click within a a hexagon it must be a triangle is this the same sort of algorithmic uh mathematics space um that you can use to resolve a lot of utility um off of hardware uh across code in in blockchain thank you very much um in case you um you'd like to discuss more and validate perhaps assumptions i recommend to uh of course bounces off the swarm researchers uh i'm for sure um perhaps we would also perhaps recommend taking a look at the people in the state box project because they've been working on graphs for quite a while yeah state box you think yeah yeah they're a different uh project but they uh they've been working on the graph three for a while and then they're uh they might be nice uh uh partners too to uh well i would definitely check out what they've got up on the on the web anyway thank you very much i'm going to go ahead and dodge it thank you petra there we go let's continue see it is uh almost five um let's see i have alba alberto open social network hello hi hello everybody uh thank you for the opportunity to present i'm going to share my screens okay uh i'm going to present open social network um it's really simple compared to the whole explanation before um we want to build a social network where all opinions can be expressed it will be a kind of twitter but based on swarm our team is composed by two members we are alva and jordan and jordy and well we work we work as web developers um we are building a open social network which is a censorship resistant permissionless and decentralized social media platform of course based on storm it will be a front-end fully based on the on the back end of swarm it will be a non-profit but it is a non-profit project and the motivation comes after all those project problems that are appearing related to the main social media platforms such as facebook or instagram or twitter maybe um that are acting as censors of our time we consider that those companies don't have the should not have the power to censor any voice and that's why we thought that building a social network based on swarm would solve uh these problems uh well i do not need to explain what is work here um [Music] but we chose swarm because uh it offers us a whole backend and it's also a cheaper way to build a social network than compared to build a social network deploying contracts directly on the ethereum network and why you would or should join the open social network well one point is what i said it's a censorship resistance [Music] network but also it is not based on the extractive attention economy we do not want to win money with the attention and we want the user to be owner of their data so i think it's some of the main differences with the the main networks uh we do not share the the polarization that is promoted with all the other platforms and um i wanted to share just uh some technical aspects that it's written uh a new framework and host it on form the it communicates with the swarm network with bjs library apis we also assume that the user has access to a swarm notebook i think it's quite important [Music] and well we would run with our browser and now i would like to start with the demo as i think the idea is quite easy to get just one second i want it okay we get in here and we log in with our metamask um second i think so we enter and we will with metamask i will change the account because [Music] okay so now perhaps you could turn off the video to save bandwidth sometimes you you break up a little bit okay um perfect better maybe thank you okay so now we are a fully new user we have a welcome page where you can see a bit of the presentation of the network and then we have the first step we'll be connecting to a denote we have to write the address and we'll check our connection the next step will be buying a bunch of stumps for the note um and then well well we it's a bit slow but we wait for this one and then we create our profile and we won't store many information and so jared can keep as anonymous as he or she wants to i am using uh for a whole frozen scheme just to keep a funny user experience so i will keep with uh anna and i like flowers okay so i have the batch stamp let me store it just for a second in the because i will need it later one second that's you will need to do this but it will make it easier later okay so now we start we have to sign because we are posting the feeds uh and so we come here we can edit our profile um here i will choose an image i have prepared um now we can just pause the image by an a link maybe in the future we will use drive to so that you can post your images and your videos with project but now it is we are not using it still and we will save it okay well i will also add an address from another user so that i can follow uh no sorry well i i can follow but i will do it the other way around now i just want to store my account address and switch to the other user so now i will switch to the other user okay and then you will see [Music] effect okay when once you have logged in and registered before in the social network you will have a different message you have a welcome back and you just have to check the connection and buy the badge if you do not have bought it before but now it's what i said i will pick the bad system and just start here so that it does not ask me okay so i go to my user now i am live i have my beat and i have my wall where i can see the the bits that other users have posted um i forgot to write a post from the other user but we can see the posts of a user another user i follow um i can also post some bits in here now it is really simple you can add any any text i mean at the moment it will appear with the mask also and sign the message okay um that's it we can see it in order uh the most recent post it will appear the first one and that's it for the demo you can see it's in development and [Music] well there are many things to do still but we wanted to share anywhere anyway the the presentation because it was it is a nice thing to show i come back to your presentation and just to finish i we want to build a digital social social network just our physical society is uh imperfect but without a magnifying glass on our imperfections um please uh we are on the project as we said it's in development we want anyone who wants to help developing it's invited to we anyone who wants to test the user experience now it's not much to do but in the future will be more and also anyone who wants to share the social network with the community everything will be great you can visit us on this github link where you can see much more about the project and some and how to run the the the example also uh we have to say that for development we used uh the v factory just if you want to to test and well we want in a future we will also add some next steps that we want to add at the project so that other users can contribute in some concrete aspects just to help maybe that anyone can [Music] say anything um well i think that's all i don't know any questions yes there are thank you um a question from elle definitely not i'd fmb uh with the question i see this not been tested yet i mean it is it running on test net per chance um yes it's it's in desnet uh and we well the b note is with the b factory and it's all in test by now yeah uh thank you and a um another question was are there any ideas that we can add as contacts to test [Music] i'm not sure how to answer these questions as in could others test with you um well i the the way to test it would be to to reapply all the all the same development environments and with the big with the be factory and also with the clone and with metamask test accounts so that you can repeat all the the same more or less the same as i did in the in the presentation okay thank you question from tomasz what else besides the censorship persistent features of swarm attracted you to work with swarm uh well um i would say also the the simplicity in we saw that it was quite accessible um with all the [Music] the bjs api to contribute uh what to use it um and we thought that it would be a good alternative and a question from gregor can other projects use the same protocol as in would facebook's twitters and tick tocks be interoperable [Music] well um [Music] it will be fully open so that anyone can fork it in case uh it is interesting um the code is open so but but by now it it is not interoperable with the others maybe in the future it would be interesting to take information from one from one to another yeah and uh aside from contributors on the github uh what would you be looking forward to right now for the project well um with well the any any importations would would help the development we had the swarm ground which has been the first stone in the in a way but it would be great to to have other contributors also any other help that or any other uh i would say [Music] person who shares our values and have new ideas to the to build a better project [Music] will be welcome as we say we love india but we do not own it so anything we invite you to make the project yours also and make it bigger just because it can be a good project for the for all right thank you thank you no further questions oh um so that's the get up as a repository is what you got open now right open social network yeah so for people to find more information then i have church allen who is in the call there we have trump robert you would be presenting on i believe i would be talking about the project galileo which is about uh maps openstreetmaps so i will try to share the screen just a second okay um i guess the whole screen is okay and then the present mode on the slides uh we can see from your background that you love maps so yes i'm all about math um and with corona you know uh staying at home all the time but i can just turn around and see the whole world behind me okay so uh the galileo project um it's uh it was actually presented already at the s'more swarm one event uh it was called fair maps back then um i will share the link to the recording later on so it's it shows the recording shows uh um how the ui ui works and also uh goes through um all the scripts for transforming data and things like that uh although very quickly um but it might be helpful uh if you want more information so the project galileo is uh i would call it it's uh two things so it's a technology proof of concept but it's also a sort of community proof of concept and the technology part is about swarm and ferrous as the layer on top of it so the maps are sourced from openstreetmaps you use some scripts you can use some scripts that are provided on the github page to transform them to proper structure and then import them into the key value store in ferro s that is that that is using swarm as the underlying storage and galileo also implemented a front-end for using ferros to display the maps you can zoom you can scroll through the maps you can add add markers to the maps which are then sourced saved to ferro s and swarm um and um uh then uh because uh uh we have this i'll go to that a bit later so this idea of community providing the maps uh there's also this uh idea of a registry uh pod that contains links to to other maps that's that has been used here there's just a simple uh proof of concept implementation of this so um because the uh the architecture of uh uh ferrous is such that um you have these spots which are like um drives on your on your hard disk and they can be made public so if you expose this spot containing the part can be containing files which don't have like a theoretical limit on their size in this case uh it can be used for key value store or a document store in this case we use the key value store and so if you expose this imagine you expose one of these pods to the public and it contains linked links to where the actual maps are located then this spot would be like a registry pod and maybe now i touch upon the community proof of concept aspect of this this project so we believe that maps uh are something that should be maintained by the community or could be maintained by the community uh so in the openstreetmap case this is the idea so the community updates the maps for galileo we imagine that there could be like country maintainers that could make sure that the map is up to date regularly and publish these maps to their pods such a maintainer so how would they get compensated for this so theoretically they should receive more traffic because they're like the originator if they have their own node right they're the originator of this uh new data that of course uh the freshest maps um and um so also uh so we want to build a community around this so we have in mind some kind of uh community building exercise i will not go into more detail now because i just want to maybe touch upon this that this is something that we have in mind um but then these country maps could be aggregated sort of like in a registry pod so for example fairly to society would have a pod where you could the registry pod that you could apply to so for example you would say i would say maybe i have a country map or slovenia i'm this maintainer could you please add me and a user can uh could pick maybe we could pick several such registry pods or just maybe the best one on his user interface and immediately have all the maps uh available so this is something that we're experimenting with but of course it's maybe not even the maps that are the most valuable it would be later on would be some kind of layers so uh on top of mats maps that add value so i don't know maybe um you have a list of restaurants in your country that you maintain uh sort of like your own michelin list and you would like to share that and you would be the maintainer of that uh but uh yeah the maps some maps are you know the first use case they were covering but uh one could imagine many different uh other types of data that could be maintained and shared in this way so ideas here always welcome you can try out galileo but uh please note that uh currently we're upgrading pharaoh has to work with the latest version of b so you'll have to the instructions in the github say which version is the last one that will work out of the box so in the coming weeks also the newest version should work but uh here is the github link uh it includes the github repo includes the instructions on how to test a map with the local ferrous and how to create and share your own maps so we would also um appreciate feedback on this this process how it works out for you um but uh more will be coming up actually in the uh in the coming weeks so it'd be because we will build upon this idea uh at the upcoming events so um this is like maybe a heads up on this and to contact us you can use the email for little society at itswon.org because this is a fair to society project um well okay this is a quote from galileo i thought maybe for for friday it would be appropriate but basically i i wrap up with this um robert and i would give the mic back to you thank you thank you richard um i it looks like my question would be how are you going to ensure the quality [Music] of the map updates you mentioned there community will be responsible is there any coordination this is actually part of the experiment imagine that uh some maps will be bad so if you if you use this simple case where ferrari society curates dislike manually then i guess you would get complaints about the map and you would remove that provider from from the map but you could like automate the process uh further on with some kind of curated registries you know you you complain you vote you lose tokens if you get removed from as one of the providers so you can you know this can be explored in in various ways but basically currently it's not we're not giving any guarantees right it's but some kind of reputation based system would have to be put in place interesting interesting okay uh we have a question let's see is there any provision for timestamping when the various pods were last updated from the openstreetmap's source data so i guess not at this stage so if you update your data i guess well since ferros supports this updating uh it the new map should be available uh but if you want to like advertise there is a new map this system is not in place per se right thank you that concludes the questions um thanks again and i'd like to now give the room at least the floor to um one of our participants who couldn't be here but he pre-recorded his demonstration uh his presentation and his demo which is igor shadurin with the project zumarok hey my name is igor chadurin i'm a participant of the throne ground program the project i'm going to talk about is called zoomerock which is a platform for viewing and publishing short videos once upon a time when social networks were just starting to appear i really enjoyed creating communities small games and applications most of them were not in great demand but some of them became quite successful the success was manifested in hundreds of thousands of active users and subscribers if you found a way to hack growth then sooner or later the administrator administrators of social networks would come and regulate the permitted ways of growth i don't see a big problem in the regulation of social platforms but as a result of such a regulation owners often lose their communities due to retroactive rules this situation was repeated with me from time to time on different platforms i started seeing similar situations more often among other content creators they lost all their articles photos videos and subscribers in one click of the finger of the social media administration and then one day i asked myself why not create a platform where all your content and subscribers are controlled controlled only by you so that no administration can simply take the results of your work in theory this could be solved using distributed technologies but quite recently a tuning point came when all the technical possibilities for this appeared it became cheap to store user database in a distributed registry due to a second level networks any here content such as photos or videos can now be stored distributed in swarm zoomi rock is an example of social network built on top of fair os swarm and the ethereum compatible xdi blockchain upon registration the user receives a mnemonic phrase which is the only way to manage user data xdi will store basic information about the user and his subscribers ferrous stores old and loadable content photos videos and previews with the help of the sharing methods implemented in fair os you can very easily get a feed of the content of all the people you follow it's no secret that distributed networks are not always quickly synchronized with each other especially when it comes to big data for this within the framework of the zoomiroc project a caching server was implemented that allows an ordinary user to feel as to feel as comfortable as in web 2 although in fact to operate with data in that free this implementation is interesting in that it can be run on any vps server or local machine via docker and the user will securely access the content of the entire network so if a user doesn't trust a public node he can always switch to a private one in addition washington publishing short videos is not all that the created backend is capable of on top of swarm and x die any developer can create their own mobile or web application and connect it to the existing network the entire user base and their content will instantly be available in the new application now the mobile application is available by invitation perhaps in the most moderated app store the apple store this is done in order to immediately start building a bridge between sensor store apps and decent decentralized data it is necessary to learn how to match each other and it's not so easy but if if it succeeds the swarm community will be able to get a huge number of new users thank you for the attention for any questions please contact me by email bye thank you thank you eager and uh we're starting to see a pattern here how uh projects like uh the open social network and zoom rock uh come to swarm eventually because of the limitations by the existing centralized solutions so nice to see for now we go to the last presentation by the project and then we'll have some announcements um for the last presentation i'd like to invite jeremy bott who's working on the project digital life jeremy are you [Music] alive at the moment i am alive yes i'm doing some last-minute tweaks freaks jeremy recently started experimenting with swarm um after experiments with other let's say decentralized storage networks and uh we're very keen to hear jeremy what you've been up to and what the digital life project entails in the meantime i'd like to announce again just like in the previous days of the summer school at the end of this session you can send your um your ethereum address to the host of this webinar which is the black swarm logo to receive the swarm summer school nft to be eligible for receiving the swarms on the school nft at the end of the program we're seeing a screen which is being shared there we go i'm getting a flickering here there we go okay um so yeah um i'm here to talk about the digital light project um i think this is maybe a little bit different than the other talks and mostly talk about the project itself and um very much looking forward to getting feedback and hearing potentials on what others might see in this project before you continue i'd like to point out that you're in your screen share the overlay of uh zoom is blocking part of your presentation oh okay deal with this i'm just going to hide the thumbnails is that okay this yeah this is a lot better yep okay cool uh look at this there you go um so yeah i'm here to talk to you about the digital light project where we uh scan living animals and then we we share the data so uh there we go so first of all i guess a little bit of context is uh to give you an idea where i'm coming from um uh my name is jeremy bott and uh yeah i've been working with computers for for quite a while and specifically uh i enjoy moving analog into digital for quality and sharing abilities and you know obviously the ability to edit and uh hopefully the ability to secure it and store an archive um this is me in like 1995 um where we were moving um i was working at a local radio station and we're moving from reel to reel to uh digital editing workstations and this is me uh demoing the stuff and loving what i'm doing here um and uh you know to give you a deeper picture is uh you know i i've always been interested in uh medical um and uh in particular uh um you know just anatomy and physiology and uh my plan was originally to go on to uh to be um a doctor uh but then my mother shared with me that uh you know jared you know um it's uh you know it's good there's gonna be a lot of school involved and for me that was it was a big turn off because you know i was going to be learning from books that were published that were getting quite old and um and then you know that i realized i was going to be learning ways that other people had seen the world as opposed to having direct access to that data myself and exploring and figuring out the world for myself um so eventually instead i got into just computers in general specifically computer support so i got a microsoft certification and with that microsoft certification means that i became basically a product specialist and when i realized that my entire career was dependent on a you know proprietary closed source piece of software uh well i let it i let it expire um and then i shifted into animation um this computer graphics was a way of exploring reality visually um without barriers and without uh you know kind of a there's no no predefined path necessarily so i've worked on film and tv and video games and then some science work as well and currently right now i actually have one foot in the science and one foot in entertainment um so to kind of converge these paths now is if we we talk about recording life in general um 30 000 years ago in the chavo caves and france um you know these these cave paintings were put up and it really shows you how how long we've been trying to capture life and and store it and archive it and then in the late 1800s someone was recording motion and they're doing it in a very interesting way and funny enough this was actually uh charles darwin and he was using this very interesting way of recording motion through a piece of glass and then kind of tracing this plant as it grew to determine you know the paths in which these plants grew but that was pretty interesting and then right around the same time you had edward muybridge which i think most people if you study motion at all i've seen these this is pretty much the first time we've recorded actual motion um and this was kind of the result so i think this was one of the first times yeah we had like any kind of a video this was i think 1877. so now let's move to 1957 and this was the first digitally recorded digitally recorded animal and it was uh the child of a uh uh basically the inventor of the first digital scanner and this was his kid um and now here we are in 2017 and this is when i connected with the digital light project and here is a toki gekko using photogrammetry processing um so basically a series of cameras and taking picture and then this ended up as like four and a half million polygons and a pretty large texture as well and you know why do we do this well it's easy to explain by mission statement which we stick to pretty religiously first and foremost is open access science you know allowing the world to explore uh whatever you know whoever wherever and uh you know education in terms of being immersive in this you know having direct access is i mean you can see there's a very strong similarity to the science and education i think ideally these these two components would be kind of merged and then of course conservation and conservation is kind of a weird thing because i think we think of conservation we think you know kind of archival maybe you know when we're talking about digital um but really with the digital the goal is to to bring awareness and uh and use that to spark conservation efforts around the world and then of course you know by sharing these so these are some examples of the animals that we've digitized and then re-topologized so that we could store them in an efficient manner online which is uh you know get into i guess at the end um so then how do we do this process so yeah first we plan um we go to the uh the enclosure this is at the san diego zoo this is uh duncan urshik a biologist at the university of massachusetts and then on the right that's actually him at the perth zoo in australia on the left he's setting up to take a picture a scan of a polar bear and then on the right they're setting up to take photographs of a rhino so the actual recording process begins and here we are on location you can see in the upper left in the background you can see the cameras on poles and then he's taking some handheld shots simultaneously to capture you know a finer detail from close up and on the right that's a shark that's basically pulled out of the water and put on the deck of the ship for about i think about 45 seconds and that hose in its mouth is actually flushing salt water through its gills so that it can still breathe and then here's the the scanning that happens in the lab so test it have different setups a lot of them are pretty crazy especially the one on the left um and then the one in the middle is the what's what we call the beast cam rig which is for most small subjects that's what we use it's what allows the high fidelity to occur and there's a reason for that and it's because as much as we try to get the animals to sit still they just won't sit still and if you have a camera handheld and you're taking let's say 20 photographs of an animal then uh you know the the computer won't enable it won't be able to uh pinpoint the locations in 3d space so by having numerous cameras simultaneously firing um it's basically doing a laser cam a laser scan from all angles if you could picture something like that and then of course reconstruction traditionally this is how you would reconstruct something in 3d which was by hand you would line up these isometric views and start pulling polygons and that kind of works for mechanical stuff but when you get into organic stuff you are kind of fooled by perspective and this is a very common thing so a lot of the animals you see for example in film and tv are usually captured from photographic reference and then you can see you know the distorted angles can distort the subjects and give you inaccurate results plus the subjects and film and tv are usually constructed from numerous animals so you actually have like a male and a female and a juvenile and an older rhino for example all merged into one actually i had a friend that worked on which was jumanji and i asked him what kind of rhino that was in jumanji and he started laughing he's like well i think it's an indian rhino mixed with a southern rhino and juvenile you know basically grabbing the attributes of uh numerous rhinos to construct a full one so the process that we use is you can use on those previous pictures you have numerous cameras set up around the subject all those cameras fire and then if you look through each one of those photographs basically you say okay so where's the eyeball and then and then you'd say okay now where is the front right foot and what it does it basically triangulates the positions it's the easiest way to explain it and then you triangulate point by point by point by point and then you end up with a mesh but unfortunately it's usually not that simple the the typical result looks more like this sometimes with after much tweaking then and you can get something that's a little more accurate and flushed out and uh yeah here's some more tests um and then this is one of the better tests that came out for the rhino and this is a view from below so when you turn those points into a mesh and then you project the the texture onto the mesh and this is what you get is uh from down below you can see where the cameras weren't able to get underneath there's a lot of missing data so in cases like that we rely on other data to kind of fill in that missing data whether it's working with the zoologists who have a camera a stick and they'll wave it underneath the rhino and capture video in reference like that or in this case there's uh ohio university has a ct scan of a rhino skull and this was really helpful to reconstruct the front of the rhino's mouth and make sure there's a lot of make sure we maintained accuracy there and then in other cases we might have to do two sets so in this case this was a sea turtle that was uh stranded and been bitten by a tiger shark and you can see on the right side of its shell that there's a big bite mark in it and then of course it's missing its front right flipper it was rescued by this marine center in florida and then uh while duncan was down there he did a scan of the the turtle before the a couple months before they released it so that meant scanning it on the front scanning it on the back kind of using this tool to trim away that access scan data and then you know we had a top and a bottom and then you just kind of flip it over and put it on top and then rig it which allows us to animate it which basically then i can slip it into position and then from there they're merged and there's a lot of cleanup obviously needs to be due a lot of cleanup needs to be done and this gives you kind of an example of like some of the missing data again that happens so that we have to you know reconnect with the biologists and receive that missing data um and it's kind of a fun thing too because you know when you're looking at stuff you're not really sure what you're looking at and you know just kind of give you a puzzle here uh and one of the things that's a challenge this is actually the underside of a rhino's neck and the lower part is the shoulders and the upper part is you know leading into the jaw and you know you're looking at something like this and you're trying to patch this data into the stuff that's missing on the actual rhino it's a it's kind of tricky and then of course verification of that data um so that basically means taking our scan model and lining it up with the original photographs and seeing you know checking the silhouette and see where there's issues and you can see like in between the legs obviously there's a lot of data in there that's been added by some of the processing so we have to fix that and there's numerous methods we have to finding what pieces are accurate and which pieces aren't and then of course working with the biologist to verify the data as well and of course rigging it because one of the things that we do with our project we feel it's important to capture motion and this is something that i think we'll be spending quite a bit of time on the next decades for that matter is because capturing a static object is relatively simple but capturing an object and motioning is very tricky so currently the way we do it is we study the animal's anatomy and then we rebuild that anatomy using what we call a rig in 3d animation music and film and tv obviously by studying uh the anatomy then we can determine where the pivot points would be and then we attach the geometry to those pivot points and then that red hot red that you're seeing is basically saying you know well it's showing me what's attached to the bone that controls the shell of the animal and then as i pick through the bones you that heat map would change depending on what bone is picked uh and this is uh what the uh agama lizard looks like and then yeah then we animate it and uh it's kind of funny some of our partners actually use motion capture to capture the motion of these animals and this is a whole new challenge as well because you can't use a lot of standard setups that were designed for bipedal you know human beings so to do mocap on the lizards and snake is a quite a challenge of itself and i recommend checking out savasa's work it's pretty it's impressive and then another option is using machine learning so basically using video and the video analyzes the motion over time and then is able to project that motion back onto the 3d model that we created and even create a skeleton and some other preliminary tests that we had and here's a that's an example of what happens when you have to animate it by hand so on on the left hand side of the screen is animation curves uh controlling uh the euler coordinates uh the rotation of the head for example and of course the most important part of this process is sharing it so currently we use a site called sketchfab and sketchfab is a commercial website um and it's it's a it's a great site it's really cool you can basically you can view these animals in full 3d i would actually recommend checking it out so yeah i'll give you uh can give you a url at the end but yeah you're allowed to see these animals in full 3d interact with them rotate them around and you can even view them in vr which is pretty slick one of the goals too is to make sure that these animals are accessible for as long as possible because some of these animals already starting to disappear and you know relying on centralized storage obviously is not ideal for a long term when you start talking about you know 20 40 50 60 years and for an example you know one of my concerns is uh acquisitions so you know i kept saying to my partners one day sketchfab's going to be acquired by somebody it's just a matter of time and coincidentally i think about four weeks ago now they were actually acquired by epic mega games the makers of the unreal game engine so now all this stuff is kind of in their ecosystem but so far nothing's changed but again there's an inevitability to that and that's a major concern in terms of conservation and accessibility uh it's worth mentioning that all the software that i use now i've basically converted all my processes over to all open source softwares and these are some of the softwares that we use in order to do the scanning and cleaning up of the models so it's kind of a nice part of it too which means that you know when i teach biologists how to do some of this process you know they have full access to these tools and i think learning about biology and one of the best methods is actually to create one of these models because you become very intimately knowledgeable of the animals when you actually have to create them from these photographs or scan data and in the future looking at color accuracy is going to be a big challenge um scanning more flora scanning more deep sea creatures which actually i'm working on now i'm scanning what's called a snot palace which is a very funky looking thing which is almost impossible to scan because it's uh if you basically took it out from the depths it would just be a pile of jelly kinda like a jellyfish using procedural method methods for texturing so we can get more finer detail which would mean kind of studying these fractal-like patterns and the surfaces of these animals and how it relates to the topology and then reconstructing that through these procedural node networks um and uh yeah and in some cases maybe some things can't actually be meshed right or animated in which case we would just upload the point cloud data this is a tree stump that's at stanley park in vancouver bc as an example and this is actually kind of a fun one to to view online and you can download it as well if you're curious and then one of the bigger challenges i have right now is i'm working on a polar bear which fur is a whole another topic i guess so what's what's next archival i always uh cringe whenever i hear the statement of you know we're going to put these models on the cloud and to me i think calling our current methods of storage the cloud is very deceitful i always think of a cloud as you know decentralized you know no single point of failure you can't grab a cloud you can't move it it's just there and it's ever present and finding a place in the cloud for our models i think is uh is quite important and thanks to a lot of crypto technologies that are available now i actually see that as a reality work to be honest before i wasn't sure if it would ever be possible and that's about it and you know there's always a big thanks to the blender and group and community it's a incredible tool for animation and it's just gotten better in the community especially is absolutely phenomenal for those of you that are interested in 3d graphics or even 2d editing i would highly highly recommend checking it out the group that i'm representing here is the digital life group which is at university of massachusetts uh that is duncan irshik and johnson's another 3d artist partner as well as robert and then myself jeremy abot and then if you want to see the the 3d models and interact with them again right now they're currently available on sketchfab.com digital life 3d and uh yeah there's my contact information there's a contact page on double double07.

com where you can ping me on twitter and say hello and that's it thanks for your time thank you jeremy i'm amazed by the work you do and the way we can still even the future uh these experience these animals even though we're potentially [Applause] a threat to their habitat um before i go into one project there was a question from the audience but uh how many 3d models of animals that have become extinct uh did you collect let's see the project started probably about i guess about six years ago i think so far we're fortunate that i don't think any of the animals have become extinct some of the animals have passed so and that's a big thing too in terms of capturing individuals because i think we we often think of uh you know a southern white rhino as a tribe of southern white rhinos right but the individuals obviously when they pass that's something that can't be replaced as well but the northern white rhinos as everyone is aware of like they've disappeared we don't have a scan of a northern white rhino but we do have some of the southern white rhinos we do have scans of some basically the seven species of sea turtles around the world which um a lot of those are in in bad shape um some in some cases you know maybe there's like 30 000 left on the planet um and you know the slight temperature change or who knows what could happen right those things could disappear at the blink of an eye so but uh yeah i would say fortunately none of the stuff we've scanned has gone extinct yet unfortunately yes uh how do you see uh uh uh remark in the chat now this this could be great if combined with the galileo project for a metaverse of nature yeah yeah yeah yeah totally i would i actually opened that uh project up just during the talk to check it out but yeah that's that's something actually we think quite a bit about um you know this concept of nfts for example it's something i i'm working on in terms of a proposal for various projects actually um but using uh nfts as a way of you know geo-tagging these assets and and then also connecting those assets to individuals that created them because i think that's a there's a science is the process and i think you know putting animals online and saying yes this is a real animal i don't think it's good enough we have to associate that animal with the creators whether it's the digital artists or the scientists or partners or even the funding like you know where you know where did this what it what you know what created this asset who who paid for it who worked on the model and i think from there you know using uh nfts and wallet addresses or just identifiers attached to these as a way of giving it validity and then yeah being able to then pop it on a map even even even more amazing and be able to fly around the world and see the difference and the flora and fauna across the world and then and then who worked on it who was responsible and i think through the reputation too and through crypto technology in general like funding becomes a lot more possible as well and uh something i've actually thought about which i i'd love to ponder more with those that are interested but um with science i feel that there is some topics that are becoming quite touchy and i think so a lot of scientists maybe have opinions or research that they would like to share but how does one share anonymously yet have a reputation and i see that as something that's quite possible now where you know you can actually build a reputation with an id as opposed to having you know your home address available and i think that gives voice to those that are afraid of speaking up and have important things to share definitely is the uh uh these models could they also be used for exploring for instance evolution at one point uh yeah absolutely um and i guess i just realized my camera was on as you can see i'm in a fishbowl too yeah and i think that's a big thing it's like the more the more accessible this data is the more we can use things like machine learning to kind of see as these animals evolve that's a that's a massive thing i would say too um on that note to use machine learning obviously you would need a massive data set so you know there's only a team of three of us currently doing the scanning so having a goal of scanning you know all of life on earth is quite a quite a large thing but using systems that are more scalable that are able to connect scientists with artists think is a huge a huge bonus and then we can you know start to make a much larger library of these scans and then that'll give us enough data to actually you know feed into machine learning to start to analyze and see how these things are evolving about that data set uh how big is your data set right now just curious oh gosh okay so yeah i mean that's a really really good question and in terms of the animals themselves i mean we have i think we have like let's say 60 animals scanned right now um we can you know considering there's like four of us we're quite happy and the quality keeps going up but anyway um this the the the final three topologize models so these models have basically been you know we use techniques that you use in a video game to make them light so that you can animate them in real time right because you can't animate something that's you know six million polygons at least not in real time so i would say each of these models ranges anywhere between 50 to 100 megabytes in size so you know there you go like you know five five to ten gigabytes of information for the final models that we build right but if you know a lot of this stuff comes from photographs that are you know 5k or whatever larger right they're massive um and to again science is really about the process and the source so if you start to look at the actual source data so for example maybe these algorithms to do photogrammetry from the original photographs gets better right it's important to be able to go back to that original data and use that to make more accurate models if you consider all of that data you're talking i mean you know obviously terabytes in size right i think last time i archived some of this stuff where maybe we're about five or six terabytes of data and that's for 50 models well there's more actually there's there's there's a ton more models that we haven't uh put online but uh yeah right now i think we're looking at about five terabytes of data that i usually back up which again not massive but when you start looking at all of life on earth holy cow yeah and um with regards to swarm uh you started recently experimenting with it a bit uh but i understood on what aspect would you uh uh would assistance be of help or or uh what are you looking for right now to be able to test it out properly yeah yeah um uh let's see um so there there you know there's numerous uh projects we've looked at obviously you know our weave was one that kind of grabbed my attention well i looked at all the solutions that were out there so file coin and see it um and uh you know i felt like those systems were quite complex and then plus they weren't permanent storage so are we grabbed my attention because it was a way of permanently storing it but in terms of a working model to collaborate with people around the world that really isn't a viable option right because i mean i don't have a desire to store a bunch of data that we're just going to be used as an intermediary in the process of collaborating with other people and then i came across a swarm via you um and uh it brought my attention because it felt like it seems like swarm is more integrated it's not just this layer that's put on top of the storage system it's not permanent but then that's kind of a nice part too because we're not wastefully storing a bunch of data again the intermediary pieces and then of course there's that incentive mechanism again it's it's it's built in um so yeah there's uh in terms of uh help i would say with everything there's a lot there's a lot there's a lot for me to learn um i'm just kind of still going through the process of uh setting up a note myself and then being able to use it so i don't really have too much to say yet okay i'll be sure to check out our community on discord and it's quite helpful yeah and of course enormous uh amounts of documentation which is available yeah there was there's uh there are a couple questions that i had in terms of uh um how you know the link to ethereum and how that all works and uh when i joined the discord uh two days ago and my questions were answered immediately so it was pretty awesome it was exactly everything i was asking and more so i'm looking forward to diving deeper just have to make the time move away from the animals for a minute and then focus on the storage but great to hear and uh i hope to see uh uh curious to see how the project evolves and how as one can be uh hopefully of assistance in maintaining the the well the experience and the the of these not yet extinct luckily animals right yeah yeah the i guess the the piece that'll be next would be uh for me is uh you know once i've got a node set up is then i'm sure there's presentations online of others that have done it basic you know linking a website that directly references swarm and the data that's stored there that'll be interesting to see and see what the throughput is the models again they're not massive but they are large you know a 30 megabyte file is not a not something to scoff at i mean google won't even let you send that as an attachment so that's something i guess so yeah be curious on performance and stuff like that as well uh perhaps we have an uh an option uh where you could find potential collaborators uh on this but there's someone else who could talk more about this which is not me thank you jeremy but that's angela let's see where angela fits to angela are you online yes i am angel you have an announcement to make i believe i do so i am working as a project manager with the swarm team and first of all let me say that everyone had really great projects and demos we want to keep building on that momentum so we are going to be hosting a hackathon at the end of october focusing on data so more details definitely to come on all of that but in the interim if you guys are interested in participating send us a message to summerschool2021 at 8thswarm.org and those messages can be filtered over to me and i look forward to connecting with everybody thanks robert thanks angela so if you've done uh during this week if you've uh have validated some of your ideas or if you have uh got inspiration from the speakers or already had some and you want to test out your ideas and uh show them to the world and this hackathon could be of interest so if you have uh it sounds interesting please feel free to mail us we'll keep you updated for now that would be via summerschool2021 at ethworm.org of course there will be its own uh there will be an announcement page with more information coming soon and with that i would like to thank the last presenter which was angela and all presenters today uh old projects for taking the time to demonstrate their their use case their their products uh and how they're building and why they're building on swarm um with that i'd like to thank also all participants who didn't present but who were here with us during the week and next to that the team which has been working behind the scenes to facilitate this all again if you would like to apply for the um swarm nft for the summer school 2021 participation uh please send your address to the swarm host which is the black logo of this webinar and with that i'd like to close off this this last day i congratulate you all for uh graduating this summer school of 2021. thank you