Monthly Development Update - January 2021
Transcript of the Tau-Chain & Agoras Monthly Development Update – January 2021
We are very excited to welcome new team member, Luca. who is a student of mathematics with Prof. Benzmüller. You’ll see his background is very impressive. This month we continued work on the core TML product and the Agoras Live product which is being rebranded as we all know. On the TMlL side Murisi kept working on the performance of the eval predicate and is running benchmarks on eval and has prototyped a program transformation / optimisation that will help with running eval. It also helps with CQC optimisations and he is also working on a parallel version. Lots of focus going on making eval faster. Juan continued his work on ERC-20 token with the help of Ohad and Fola for the transition to ERC-20. WOrking with the exchange that will first list us. Working on some tools for that. He has also been working on the payment system for Agoras live. Switching it from the Omni layer to ERC-20 layer so trying to replicate the feature of micro payments we had with Omni layer. THomas continued his work on both TML and the TML IDE [ https://tml.idni.org ]. On the TML side he finished the transformation of nested programs into one merged program and he added flow control, just like any other language IF, THEN, ELSE and WHY statements. He also on the IDE side made the executions of TML run in the background thread so that they don’t block the UI thread. He converted the IDE page to a progressive webapp to give it a more native app feel. He is also working on the documentation for TML and reworking the introduction to TML. All in preparation for a !.0 Release of TML hopefully for Q1 of this year. On the Agoras Live side Andrei is doing thorough testing and fixing minor issues as he goes and other optimisations. He also implemented the feature called “recurring group calls” for future utilisation. We are redesigning the discovery algorithms for students and teachers to find one another. Mo’az is still refining as part of the rebranding effort for agoras live.We transitioned from using tools called sketch and zeppelin to using one tool called figma which integrates the functionality of both. He has redesigned the front page and redesigned smaller pages to accompany it. Andrei is still implementing new designs. We are meeting on a regular basis to push the launch Agoras live as soon as possible.
I'm very happy to join the project this week. I’ve been in talks with Ohad regarding Prof. Benzmüller. I’ve joined to work on the second order quantifier boolean side of the TML language THe logical foundations of TML in the higher order domain. Regarding my background, I'm doing my masters at the moment. I’ve done a bachelor thesis together with prof Benzmüller and dana scott who is a famous logician from the US and publishing a thesis with great success. Besides this I’ve been working on VR development during the past 2 years. I’m looking forward to getting involved further with my mathematical skills.
Last month I’ve been working on performance and the general direction of these improvements has been reduced in table sizes as I noticed that large tables sizes were causing a little bit of a slow down in interpretation of quoted programs so I continued my work on finding ways to avoid computations and a way of storing results in tables. I’ve also been working on a different algorithm for domain generation. Before I was generating domains for quoted programs by just enumerating each element of the domain but now I’m using mathematical formulas in order to generate domains. This uses modular arithmetic remainder and quotients in order to include lists in numbers. I’ve also worked on binary body transformation which essentially converts old TML rules into rules that only have one conjunct. So if a rule has five conjuncts before it is reduced into multiple rules with the single conjunct. THis has been done before but my work has been to generalise that in order to support tml rules with negation which is more complicated. We are still prototyping the functionality and we hope to find a faster way of faster transformations than the one we currently have now. I was benchmarking performance, looking at the intersection of secrecy conjunctive query containment and CQNC which is containment of negation. I was looking at the effects of those optimations in eval and trying to see if one could help the other. The results haven't been so good but I hope to find ways in order to make CQC more applicable to eval in the coming months. ANother line of investigation has been the parellisation of eval. We know that TML programs when they are encoded in a certain format are parellelisable and can be run on multiple cores so I was looking at alternative implementations of eval which are in a format which support this kind of parellisation. I’ve also been playing around a bit with the passive generator and seeing how we can improve its performance. With eval there's been a lot of improvements in the speed over the past month by over 10x or 20x. I hope to continue this progress in the performance.
I have finished the transformation of nested programs including conditional statements IF and WHILE. There is more information and examples in the readme of the TML repository and also you can try this transformation and conditional programs online [ https://tml.idni.org ]. I’ve moved the TML execution into a web working so it doesn’t block the main thread with the UI when running TML programs. I’ve fixed the UI errors and made TML execution more stable. I’ve converted the page into progressive web app so it can work offline and installable on some devices. Now I'm working on the updated introduction to TML which will replace the old intro and I’ve started to examine options to store files right in the browser or elsewhere in the cloud.
This month we decided to use figma for all our design work enabling designers to work with developers in a more streamlined manner. I’ve moved the first iteration of Agoras Live and the design system to the Figma platform. We’ve made some revisions to the design of the front page, how we display profiles and some additional features. I’m implementing the team's feedback and doing further research on user behavior which will shape further design improvements.
I’ve been fixing small issues on Agoras live, making sure everything is working before handing over to a professional tester. I’ve added image compression. [ https://youtu.be/2g67sVlxzSg?t=950 ]. I’ve moved to the current design however the priority currently is to have everything working. Every teacher can now set their status to busy and/or create a public lecture for everyone to join for a payment or for free. When you select a public lecture you want to set it to everyday or every week and duration.I have a list of fixes and tasks here and I am somewhere in the middle and of course I’m trying to move as fast as I can.
I’ve been continuing my work on the Tau presentation. We are approaching the final version so mostly fine tuning rather than changing the narrative. It’s mostly text driven now as previously it was image driven. We redefined the tau supporter program and is close to being launched. WE have recorded an explainer video. Once it is edited and updated we want to roll out the supporter program. Fola and I have done some research on companies that provide fiat onramp for crypto, in our case agoras. We’ve settled on a good solution being the company https://transak.com who connects to uniswap. This is to be implemented on Agoras Live so users can use their credit or debit card to purchase Agoras. I’ve curated a list of crypto focused youtubers and started outreach towards them. Requested either a review of the project or an interview where we are able to explain the project in more detail. I’ve experimented with 4chan. Besides this I have been doing community support. This month the community member of the month is Felix for contributing a lot to discussions on telegram!
My work this month has been focused on the payment channel system for agoras live but now based on the ethereum blockchain. Submitting payment is in a peer to peer manner supported by the Agoras token so this is in the early prototype stage but already working locally. I’m looking forward to wrapping up, possibly during february, and implementing the payment system in the Agoras live platform. Then getting back to TML development in the short term.
Fola: I’ve been working on the token swap to ERC-20. Despite our efforts the exchanges do the swap on the same day The exchanges we are working with will be doing the swap across different days. The date of the swap will likely be next week or the following week. Beyond that we are working on marketing. I’ve approached crypto and more traditional marketing companies to assist us in this push, alongside the launch of agoras live, the swap and TML. All the great things we’ll be doing over the next couple of months. We are speaking with designers now regarding the rebrand and are close on the decision to rebrand to the new name of Agoras Live. We will push the name out soon. It’s very close.
I’ve been involved with matters relating to the upcoming new exchange listing and Agoras swap to ERC-20. I’ve continued discussions with Prof. Benzmüller about knowledge representation languages that are suitable for Tau. It is a very long and deep discussion which is ongoing. There are many considerations which don’t typically arise in other systems because of the complexity of the laws of changing the laws and of course the decidability issues. On the front of Second Order Logic, I have discovered some new methods that are related to permutation elimination. The ability to eliminate muted arguments in relations as well as taking some ideas from SMT solvers. In particular, Acumen encoding. Which brings me to think the solution for Second Order Logic which has taken about a year in research, is close. I think relatively soon we will have some initial solution on the table.
Q&A:Q: Will we be able to use TML in vscode/eclipse/jupyter or will we have to wait for a dedicated environment?
Tomas: You can write TML code in any editor. It’s just a text file. So far there is no special network for TML in those name editors but it can certainly be added.
Q: How has TML been developed so far? Is it “production ready” at every step of the way, or is “quality” only to be added at the end?
Karim: So TML is essentially a runtime environment for the language. You can think of it as a compiler. Our approach essentially is that we only implement a feature if we are 100% sure we have a mathematically correct algorithm to do it. We call that correct by design philosophy to development and as such it’s not an application where it has a lot of dangling features that need to be taken care of. It’s much more amenable to regression testing also. We do a lot of that so the answer to your question is that for sure the quality is built into the product from the beginning feature by feature and we’re certainly not waiting until the end to fix all the bugs. That’s not to say bugs can’t happen. There have been bugs and they have been fixed.
Q: Ohad is BTC and Agoras your only crypto holdings?
Ohad: I’m not a crypto investor at all in fact I’m not interested. I don’t follow markets, I don’t have time so I’m just not into it.
Q: How will “risk free” interest without inflation work. WIll all locked Agoras get a slice of the hedging action on the network, weighted according to how long it's locked?
Ohad: The mechanism of interest relies on the derivatives market that we’ll develop over agoras. A curtain combination of derivatives can give risk free interest as was shown long ago by Black and Scholes themselves. It’s not so simple to explain but I warm refer you to the part of the whitepaper that explains this - [ https://www.idni.org/whitepaper_community_draft.pdf - Derivatives and risk free interest p.38 ]
Q: Will the Agoras tokens be necessary to maintain the security of the network through the execution of validators in a system like proof of stake?
Kilian: Whether or not Agoras or Tau will end up running over proof of stake will be decided on by the users. The consensus can change over time according to whatever the users desire it to be. It’s not fixed yet. It will be decided on user consensus when Tau is launched.
Q: Tau pretends to be a decentralised intelligent blockchain that updates itself and learns from humans, that is, a cybernetic intelligent agent like the one featured in the movie “transcendent”. Is this correct?
Kilian: It’s definitely correct that Tau is able to update itself and learns from humans through our communication paradigm we have developed. “Human - Machine - Human communication”. Users talk to the machine in formal languages so the machine can understand what the users have to say. So over time the machine aggregates a knowledge base consisting of all the formalised knowledge that has been aggregated over the system. The machine can then, for example, engage in discussions for you and comment upon matters or discussions without you having to be there. You can see Tau as a system that over time becomes more intelligent as more knowledge is being formalised over the network.
Q: Will developers be able to use Tau for their decentralised application and teach it tasks to make it smarter for various different purposes?
Ohad: Yes, Tau presents, not only a paradigm for collaborative decision making but also for collaborative software development. First and foremost, developing itself but also other applications. This development paradigm is knowledge oriented and consensus oriented. Knowledge oriented in the sense that people simply add more and more opinions to the system and what the software should do emerges from what is implied from the knowledge that people formalise. Consensus orientated takes the part that everyone agrees on and that becomes the software. So, yes, people can develop new software over Tau which gets smarter with time because people add more and more knowledge and the software changes accordingly.
Q: Since Ethereum is the global network par excellence for smart contracts and dapps due to the network effect. As the first project of this type, could Tau be the global artificial intelligence network on which all new consulting infrastructure is developed?
Kilian: Yes, definitely, as we have discussed before as more knowledge is being aggregated over the platform that knowledge can be used by companies, by individuals, to consult other projects or participate in discussions with that knowledge. The network, on its own, can participate in discussions and consult companies on very specific matters.