How do those similar projects actually handle ambiguously written laws? All that I see in Catala paper is basically "let this whole thing bounce back and forth between lawyer and programmer until it becomes unambiguous."
Computational Law derives its power from its emphasis on deductive reasoning. As such, it simply cannot be applied in cases requiring analogical or inductive reasoning. Fortunately, it is sometimes the case that there are enough judicial rulings that the net result is, in effect, a set of categorical constraints even where the original wording of the regulations is not definitive. And, in such cases, Computational Law can be applied to the combination of regulatory and judicial law.
I guess this is why the French programmers are concentrating on the tax code which might be more precisely written!
I agree. There are a couple interesting platform pieces before things start getting really interesting:
1) Getting US laws encoded in GitHub
2) Building a better proof of identity/residence verification provider that feels more like Authy and less like shitty SSN/identity questions, which provides a semi-anonymous user ID and geo range
3) Building a distributed consensus layer on top of that which makes it easy to conduct and tally anonymous polls across geo ranges for identity/geo verified users
Wouldn't it be great to have laws managed in a git repo.