Tau ($AGRS) Development Update | March 2022 (Ep. 46)
- Subscribe to the Tau Youtube channel for the opportunity to win some Tau merchandise.
Development Team Update
- Juan continues to integrate Umar’s work on the type system and bit universe. He’s started a significant refactoring of the entire TML code base with some help from Ohad. He’s also revisiting the safety checks, the guards, the program transformations, strings, and grammar processing.
- Tomas has finished his Earley parser work and made it Unicode capable. He’s also putting the final touches on translating the parsed forest from the early parser into a raw program so we can execute.
- Umar has embarked on another challenging task. This time binary optimization for the force generation in case of highly ambiguous grammars. It is a very tough problem, but with help from Ohad, he’s making some progress on that.
- David has finished the fact generation last month. But he made some additional changes to it this month. He’s also currently working on the TML parser and removing the possibility of negation of facts. He’s also working on grouping the different optimizations for TML programs into groups similar to the GCC grouping for optimizations. He’s also very active on the academic panel side, a math professor and a Ph.D. himself; he’s making many contributions to the theoretical front.
- Luca is on vacation this week. He’s been assigned primarily some work on the theoretical side, where he is helping show how decidability of theories can be done using an encoding into Monadic second-order logic over finite binary trees.
- Andrei and I have been testing and fixing some Agoras Live bugs. We interviewed several testers, and we finally settled on one tester. They have begun testing. He’s also made a significant upgrade to agora’s live logging procedure by implementing the wallet connect API, a standard API that allows you to use any wallet that supports the wallet connects API. We have many wallets that can be used with Agora’s Live, which is a significant improvement.
- Working on writing some texts for Tau’s new website on the topic of Trust and Identity and drawing information from the whitepaper and a conversation with Karim for the part of the research he’s carrying out regarding Monadic Second-order Logic.
- Prepared three sessions on this topic, and I’ve delivered the sessions also online, and I am now preparing the following sessions on this topic.
- Continued careful testing of Agoras Live platform, improving its UI along the way and fixing some minor issues. Found a tester who has started work, showing how the wallet connect works. (https://youtu.be/1QGLztXgNRc?t=333)
- Studied how to use the decision procedure for MSO of the infinite binary tree for deciding other theories. There’s the possibility we’ll use this decision procedure for our theories, and so we want to be ready and know how other theories embed within this logic. Each theory needs different methods to encode its structures within the infinite binary tree, and so we want to understand those methods. Will present them to the whole team once he’s ready with the findings.
Professor Umar Mohammad:
- Revisited the forest generation problem. The goal is to improve the degeneration of the forest by employing some better techniques. One of them is to do the binarization of the right-hand side of the production, and the other is to do a left and right optimization there. Doing so should reduce the combinatorial explosion that will happen otherwise. This is a challenging problem, so it’s taking its due time.
David Castro Esteban:
- Made minor changes in the facts of optimization and merged the pull request.
- Changed the current TML parser to ban several features we don’t want.
- Regarding optimizing TML programs, refactored the current source code to isolate all the code related to optimizations.
- Joined the academic panel, and focused mostly on Boolean algebras.
- Templated the early parser class so it can use any chart type allowing to finalize the Unicode support of the new parser.
- Finalizing the conversion of the parse forest into the raw program structure so we can execute TML code that comes from the early parser.
- Working on a large-scale refactoring of the TML codebase. Many different features developed in the last year that needed some maintenance and consolidation; in particular, this type of system upgrade that Umar already started and the bit universe transformation. Looking forward to enabling the performance gains through the several transformations that are already available in the TML design.
- Participating as usual in the research aspects of Tau and following Luis’s lessons in Second-order Logic, which is still a feature that we are analyzing regarding how to develop in the best way in TML.
- Continued work on taking the logic we developed into being a software specification language. One of the directions was to study infinite systems of Boolean variants. Had a serious breakthrough on this front, giving answers to the questions we were looking for.