Closed TeamSPoon closed 4 years ago
The ability to be decidable and self-define at the same time, is more accurately described as "rare" rather "unique". Indeed Datalog got such abilities, and a paper by Imhof shows two different proofs for that (for both Datalog and TML, namely LFP and PFP).
I'm not aware of an implementation of a Datalog self interpreter, and I don't have an algorithm to implement TML self-interpreter, at least not a "full" self-interpreter. But, we don't need to write a self interpreter at all, and the theoretical result of the theoretical ability to self interpret, tells us that we can add a quoting operator to the langugae without breaking the logic. And indeed a quoting/unquoting operator is planned to be supported in the future.
Also so what you are saying is currently TML/Datalog
The ability to be decidable and self-define at the same time, is more accurately described as "rare" rather "unique". Indeed Datalog got such abilities, and a paper by Imhof shows two different proofs for that (for both Datalog and TML, namely LFP and PFP)
1) are provable to do self-define.
yet
the theoretical result of the theoretical ability to self interpret, tells us that we can add a quoting operator to the langugae without breaking the logic. And indeed a quoting/unquoting operator is planned to be supported in the future.
2) would need to use the quoting operator.
Btw, I know how to do this as well but I know also it practical limits and know that It can only self-interpret a frozen copy that is unable to do anything other than what running it would have already done. IOW, the best it can do is re-interpret the image of what the interpreter had to do that day. To do what you say basically you are claiming TML violates Turing.
So my synopsis is that TML currently can not do Let alone ever be capable of being a decidable language that holds unique characteristics to interpret and redefine itself and STILL interpret user TML.
Do you agree?
closed due to irrelevance and nonsense
TML will be a decidable language that holds unique characteristics to interpret and redefine itself.
We know that Datalog can theoretically be coded in datalog just no one has tried yet.
Anyways thus TML can theoretically be coded in TML.
This issue is a "feature reminder" that it would be nice to have a test example to verify when this begins to happen.