Update
Main points of this update:
- Improved parsing capability (EBNF support) allows TML to more easily specify context free grammars.
- First order logic solving.
EBNF --> Grammar --> FOL
From observing Github the latest significant improvement made to TML is the First Order Logic Solving capability. In the Github we see EBNF which is a meta syntax notation useful for defining languages. Tau requires TML (Tau Meta Language) to define languages and for the language translation features. Whilst TML already supported parsing and context free grammars the new EBNF notation allows for more ease in specifying context free grammars of Tau to solve first order logic formulas and this should in my opinion accelerate the development process significantly. The remaining bottlenecks are primarily located in the performance of TML.
At this point from what I can see on Github TML has everything needed to build an alpha of Agoras. Of course if second order logic is implemented then this opens a new window of capabilities enabling solving second order logic formulas. To be honest I do not really understand how Ohad intents to implement second order logic because the process is extremely intricate and esoteric. I will say that if it can be implemented then it's another feature which accelerates development significantly.
TML uses what is called a Grammar to define new languages. TML itself is used in the performance test as defined by TML.g. In essence a Grammar defines the syntax of the language whilst EBNF is the notation used to define syntax. My understanding for how this process works is the definition for a language is our Grammar TML.g which gets translated into a parse tree (first order logic parse tree). The self parsing test was passed. In essence they upgraded TML with EBNF support. Note, due to current bottlenecks in performance TML.g is not yet being used in practice and is only being used in test scenarios at this time.
So by having a solver it will allow the TML programmers to essentially speak in first order logic and let the machine automatically solve it. A similar upgrade will come if and or when Ohad implements second order logic. This is good news because soon TML will begin looking more and more like the more familiar programming languages we are used to. This will allow for Tau to be written in TML itself. It's looking good in my opinion for Agoras Live and the Alpha to be implemented by the end of this year.
Reference
https://github.com/IDNI/TML/commit/61dd76581cc9b40e18a2e702f25e1dd14c2be629