220
Youtube link: https://www.youtube.com/watch?v=2MTqsHcTzQc
Project website: www.tellor.io
https://twitter.com/WeAreTellor?s=20
Discord: https://discord.com/invite/n7drGjh
https://www.reddit.com/r/TellorOfficial/
This call has a form of a status meeting. All developers reported what they are up to and what they plan to continue working on. Main topics are the Algorand and Harmony integration, some game theory discussion and discussion on disputes.
Whole call
Nick: All right. Welcome viewers to the first recorded Tellor dev call, I guess in this format, but we're just going to give kind of updates on what we're all working on.
For me, besides that we were just talking about formalizing any changes to the next Tellor and talking about what we might do. That was more of a fun conversation but this week I'm reaching out, I got to reach out to TracerDAO a crown MYU. I know you got some data specs PRs open; I owe you some comments and stuff I'll definitely get to that and then it should be kind of a nice slow week for all of us, I think. Not really a whole lot going on so we're just kind of continuing to finalize some of the things and push people out the door so you guys can let me know if you need help. Also, at 2PM today we're talking to, we have an internal call with a guy from YEARN if you guys want to show up. He's just going to talk to us about his experience going to a DAO so if you want to learn about being in a DAO that'll be...
Owen: Who is it from YEARN?
Nick: I don't know. I just chatted with him for like an hour at Denver and it was a super great conversation just about, he was very knowledgeable about the whole thing and he helps kind of younger employees who you know, if whenever like you fire everyone on the team, you start an LLC and then pay yourself and buy your own health insurance and it's how do you do those things. It's kind of a US guy who went through that whole process of leaving his traditional job and joining a DAO where he gets paid and YEARN tokens occasionally, so it's pretty cool. Anyway, all right Josh, you were here first. Nice headband by the way.
Josh: It's actually a t-shirt sleeve that ripped off and then. Anyways yeah, I'm wanting to still set up that Trueblocks node. I wanted to set it up over the weekend, although I guess it's a good thing just because I guess wanted to get some clarification on how robust we want this thing. There's like, after the initial setup those choices of being able to do like a shiffrin, a command which is essentially just kind of like getting up and running as quickly as possible, but it's definitely not decentralized. I just wanted to get you guys engaged on how important that is, with this particular node setup.
Nick: It's a front end to show people that it's live, like the television. If it goes down for an hour like once a week, it's not the end of the world.
Josh: Okay, but I'll just do the more out of the box set up then, because the shifter scrape command apparently is, you get to set up everything yourself but it's like two to three days just to index the entire blockchain stuff like that and I was going to run that over the weekend but boulder had another fire here and we had to evacuate everything so that screwed up my weekend. Real quick but good news is, we're all safe and it was you know we all kind of overreacted due to the Marshall fires early in the year but I don't know better safe than sorry kind of thing. We're getting good at the whole evacuating thing though so that's fun.
Nick: All right so let me know whenever you got trueblocks up, we can talk about how to host that or did you ever get on aws or not yet.
Josh: Not on the Tellor account, not yet.
Brenda: With a shot you should be able to just reset your password were you able to Lauren?
Nick: But yeah, so we'll set that up and then we can host it and figure out how and then I guess you're working on setting up the front end too or that's pretty much done because of the boiler.
Josh: That's pretty much done, it's just a matter of connecting the two.
Nick: Yeah, just let me know whenever you get that or if you get bored because you kick it off, a three-day process of indexing. We'll look at that let me know and we can come up with something else for you to do.
Josh: Okay, sounds good.
AK: Hi, good morning. I'm working on the Algorand contract basically. We started scripting and testing this week.
Nick: So, the G oracle stuff's basically done after the data spec, right?
AK: The G oracle that's that was my question to you I don't know what else besides that specs we need to do because it's pretty, I mean, they have their own SDK already. I don't know it's a lot to include, to tell you or particularly I don't know if you want to do that. Yeah, the data specs are done. It just needs to be I guess read, understood by you and then if anything I can fix.
Nick: Okay, cool. Yeah, no, that should be good. We can and then I'll reach out to bizdev and we can get some tweets or something from him. So, the G oracle thing we can push down. So, Spud is ready. But thanks, yeah and then we'll maybe put you on something else here soon. You know, I think the next big phase I guess for Telliot's, as we move into Owen and Lauren talking, I want to make sure that, I know, over time we've really been trying to add this disputer piece but then how do we slow down and make sure that what we're working with is quality code, that we really like and it's well tested. So, you know I want to make sure that you guys feel like we're giving you guys enough time to do that. So, we'll probably give a crumb over to you guys, as Algorand sort of wraps up here soon and make sure that everybody really likes it so it should be good. Thanks. Quentin, how’s life?
Quentin: Yeah, life's good. So, I've been looking to Reach, more research. I put the research on the google drive, so you can check that out if you want and I went through some tutorials to, you know, get familiar with Reach language, but I noticed Reach is only compatible for three chains; Ethereum, Algorand and Conflux.
Nick: As of now, yeah.
Quentin: Yeah, so is that a problem or?
Nick: It's good, yeah, didn't, they're working on Cosmos at least. That's at least, that's what I was, I sat next to their CEO at dinner for two hours so he seemed confident of it so that should be good. Yeah, and then I guess just keep pinging Cosmos see whatever they're like.
Quentin: Yeah, I'll do that.
Nick: Awesome, Tally?
Tally: I'm prepping the checklist to get organized for when we go through scripting and function tests and finishing function tests on Algorand v2 is the goal for this week. Yeah, and then I just kind of concerned about the contract before I start doing that and so we're going to meet after this meeting to talk about that. Thanks, I think that's yeah.
Nick: Cool, yeah and then if you just want to stay on afterwards, I'll answer your questions on, anyone, if you want to stay on afterwards too, we'll talk Algorand for a few. So, that's good thanks. Lauren?
Lauren: I got into the aws thing. Josh, I just had to reset the password, the account was already set up with my Tellor email, so it should probably be the same situation for you. So, I'm going to get the disputer monitor up on that and just let me know what phone numbers we want it to send and what stuff we want it to be like. Which chains we can have that up and going and then Ican talk to you later today on whatever we want to do next. Then once that's all set up.
Nick: Okay, yeah. Maybe you can help Josh out too if he has questions on hosting that database. Basically, he's just hosted it. What kind of database is Trueblox?
Josh: I don't know yet. Oh, as far as like on my end, as far as being able to access all that? Would just be PostgreSQL. Yeah, still need to hook all that up, and just for clarification again, sorry to interrupt really quick, but you said something about if I'm taking like two to three days to set this up you have more work. That's like the scrape command, so just get it up and running really quick or do the scrape command which one do you want?
Nick: Oh, you mean because like this I guess we can just get it up and running. Because we do. You're questioning whether or not we should validate the actual chain. Yeah, no it's I think we can trust it. We don't, might be other issues but. That was, no, so that was like you know. So, for those of you that don't know what Josh is doing, running Truebox is like a node on Ethereum. So, then you'll have this node but it's indexed differently, so it's index, I think his main thing is like indexes by address and then all the events are indexed as well so you can access them in a database really nicely. Whereas like a normal node is an index like that. Yeah, that's super good if you guys eventually, I think we should probably do that like with each employee, we should have them, you have to go set up an Ethereum node at least once in your life. Yeah, it'll be good yeah, so maybe at some point we're going to kick you guys and make you guys go do that. It's always a fun learning experience. All right, Lauren, anything else? Okay, cool, Owen?
Owen: I have to get back to Morphwere, I replied to some feedback on the query type and whatnot. Just need to go over that with the Sean and then also Vlad from Diva protocol wants other data, so I'll do that at some point. I need to look at the PRs and data specs because I noticed that the query data is different for a few of them, like the format, they're not using the latest, so I need to do that. And then testing as well as figuring out like if GitHub actions are running this week.
Brenda: They were not running for me yesterday still.
Owen: They weren't yesterday? That's pretty crazy but you know just whatever. I guess maybe I'll look back into that running get of actions locally thing, but maybe like rent aws instance to like run it on that's like faster I don't know.
Brenda: Well, you also have your own so if you log in with your Tellor email you should be able to reset your password so you can start using your Tellor one.
Owen: And one thing I hear, hold on Anthony is not here, so I'll give his update. We're pretty much up and running in Harmony test net. What we need to do before we go to main-net maybe this week or next is actually test the multisig for voting. So, that's what we should get done this week.
Nick: You and I can do that. Okay. I want to put some on our docs whether we do or don't right now, we don't enforce padding in for the bytes values that we put on chain. So, like I noticed writing some tests for it was a sample using Tellor I was doing like if you put on the reporter will submit like bytes for 100. That's only a few characters long but like a uint256 for 100 would be really long running with 18 decimals, like whenever you're supposed to put it on with 18 decimals and even do you pad it out to 32 bytes every time? Do we enforce that? Let's say you wanted a like the old query values you know, what have six decimals I think the legacy ones and people are putting them on without padding it out to 32 bytes which is fine and right now we're sort of not enforcing that they do but the reason we need to note it is, in your solidity file when you're parsing from bytes to uint, if you're doing like that really gas efficient method and people don't pad it out it breaks, so you have to use like you know the official converter way of actually going through it so we just need to note that at some point. So, if people are like why is this breaking and it's if you're using certain methods of converting from bytes to uint into solidity, it breaks everything if it's not actually 32 bytes long. So, if you guys ever run into that let's, I ran into that, I was confused but.
Tim: Yeah so, we should just clearly define, specify what data we're expecting and how it should be encoded.
Nick: Yeah, maybe we should tell people you have to cut it out. So, like the bytes to uint and parser reviews on change seems to work either way but like let's say you had somebody wanted something on chain with zero decimals, would you and they wanted just like the number one put on chain, should you pad that out with zeros at the beginning? Should you have to or is 0x1 a valid value it's the same thing but if you're using some crazy gas efficient assembly code like it'll bust if you try and convert 0x1 into your uint. Yeah, so I think just making sure that people know that that's a potential way to break it. Alright, that's where we're at.
Tim: On Friday I started working on the Diva payment contract and so one thing that we're facing is, sort of Diva prediction markets, they each have like collateral that could be really any erc20 token and we're taking, like I think it's 0.5% is allocated to the oracle and that could be any amount of money so that could be ten thousand dollars or that could be five dollars um so we're only we're only trying to give them a reasonable amount so it doesn't break our game theory and create like a dispute work right. But so, to know to know how much the collateral was worth and how much to pay their oracles, but we already have that reward seeing that from divas or when we're funding, I could just fund them to report for Diva and then report the price of the collateral token so that's the solution I'm looking at.
Nick: So, that would actually be really good, so the, as Tim was explaining, you have this issue with Tellor in that let's say let's say somebody came and tipped a million dollars for the next person who updates the Bitcoin price. The problem is Tim would quickly go and update the Bitcoin price but let's say you know you have to wait for dispute so it's five minutes old. Well, I want that reward so even though Tim had a valid answer I'm just going to dispute him anyway because the million dollars is more than the dispute fee so I'm going to dispute him anyway and submit it for myself and basically then it would and then Josh is going to see mine and he's going to dispute me and it's going to lead to this dispute war of valid values of you know like that will eventually all be settled out but just because somebody put in a giant tip and basically the dispute keeps going up and so the dispute fee would eventually get to be really high or I guess it just goes to the stake amount so then it would just continually you there would be some weird game theory about who has deeper pockets. People just have to buy a TRB and can take on exchanges continuously to be able to do so, right. People would have to keep buying game theory and then whoever has the deepest pockets wins a million dollars. So, it's a thing we don't want to happen just because, so we want to make sure that if Diva does blow up in the case and that point 5% of the contracts is 10k dollars or 20k dollars like which could happen. We want to make sure that it doesn't lead to these dispute wars going on to where people are like casually disputing each other to try and win it.
Tim: I wonder if the reward should still go to that first submitter if the dispute said they submitted a good value and that would fix it.
Nick: What do you mean?
Tim: Right now, if you get disputed you don't get the auto pay reward, no matter what, no matter if the dispute.
Nick: So, you don't get the auto pay or you do?
Tim: No matter if you get disputed, then you definitely don't, you're out of the running for it even if the vote says you were good and maybe that should be, you should still get the reward if the dispute vote says.
Nick: So, we should say you wait for the vote but we kind of remove that from the contract so I don't know how we would even do that. You would have to check is this, a disputed value or something. It would be messy.
Tim: It would be messy. It could fix that very specific case but maybe that doesn't really matter anyways.
Nick: Yeah, I mean it would just lead to a crap ton of buying of TRB you know and then it would be weird, because it would be a weird game theory, because I would actually be, even though I knew I'm getting disputed I'd actually be fine submitting it because I know I'm going to get that dispute fee paid right back to me all right after a vote so like everyone would be erasing dispute and it would be like a who runs out of money first sort of game. We have that problem too, if somebody was, let's say you were trying to break the network by mining every value or by disputing every value on the Tellor, if you're just going to try and break Tellor by disputing every value it actually leads to a faster Tellor because if you see somebody trying to dispute every single value you want to report, because you get that dispute fee, which is then equal to the stake, so every time you submit something you basically in two days after the vote ends are going to get a stake back. So, every single person would be submitting as many times as they possibly could so it would be like even harder to break the network. Which is a good quality to have. Either way, same with the million-dollar tip. It doesn't really break Tellor because eventually you're going to get a good value because everybody's just submitting good values. It would be interesting; it would be cool to see. Maybe somebody we can have some deep pockets we can play out the game theory. You just have to sort of wait for a little bit of time for people to run out of money and then Tellor to sift through the voting. Okay Brenda, do we lose everyone yet?
Tim: Actually, kind of along these lines was I was kind of thinking about when we were deploying that Polygon governance and you did say wait, we should think about this dispute fee rising and rising, and there was an attack vector there so we eliminated at the stakeholders’ stake amount. I was wondering, what if instead of just having a multiplier of two we could just lower that and let it keep rising, but then it just won't rise as much or as quickly? Yeah, I wonder if we want that dispute fee to continue rising but we just multiply by like 1.1 or...
Nick: I think like we figured out; it was the reason that I did it that way was like if it equals the stake amount then it's the same fee as just mining every value in that block, so you know like what's the cost to mine every value in the block and right now you could do that and it's just the stake amount. So, what's the cost to dispute every one there? It's just the stake amount. So, if you would raise the stake amount past, what if you would raise the dispute fee past what the stake amount is? Then it would just push people to mine every value. So, you know there's like obviously like there's like some gas limitations as far as, well maybe people would rather probably rather do the dispute because it's like one less function of having to stake and then mine so like you would save some gas but theoretically it's like you know on Polygon or something that's going to be exactly the same. You know if the dispute for you was a thousand times more than the stake fee, you're just going to mine every value and block it that way since we only allow one mined value per block. That was at least the game theory I came up with in my head but happy to critiques on that. Then it was like there was like no reason to push it up higher you know the fear for me would be like if the dispute fee gets too high then it almost, we run out of funds as a team you know or like the good people run out of funds and then you sort of get into a bind which I saw as being more of a risk in terms of that happening but maybe I'm wrong I don't know. But ideally, we should try and get some good borrowing markets on TRB because then it really helps the game theory a lot. If you're able to borrow TRB for like really cheap, even not really cheap, if you can borrow TRB for 10%, do his dispute and pay it back in two days because you know it was a valid dispute it's a great deal for you. I don't know who wants to make a TRB borrowing market but what would be, I didn't do a full action I don't know that would work. That would be nice though. Brenda, anything your end? Alright. Well thank you all for listening, this was this was a fun time. Yeah, just let us know if you guys need anything, we'll be here all week.
