Axiom Foundation launches to publish law as open, verifiable code
axiom.orgIt is going to be a difficult task. I looked through examples, and it seems that what they do is represent conditions set out in the law as a boolean expression that answers the question whether something is allowed or not (X is allowed if Y, Z and not W), and display it as a graph/AST. The problem though that the inputs to this expression are answers to non-trivial questions, like "Excudable Educational Assistance" or "Has Felony Drug Conviction" without specifying what exactly is considered a "drug conviction".
People might not know that, since it’s mostly in French, but France has somthing similar as open source for their tax system:
Nice, the french also are at the frontier of law-computing i'd say see Catala Lang [1]
I find that easier to read than giant graph of node thingies with lines (cue ancient debates about UML...).
Have wondered whether attorneys would find it easier to work in a programming language than english, once they'd learned the former? Do any have strong opinions on this?
Not a lawyer, but have friends who are, and the answer is absolutely not lol.
A lot of legal language looks like code, sure. It can be formulaic and full of terms of art ("actual malice" has nothing to do with colloquial malice, for example) but law just isn't code.
Laws themselves are basically human drama run through a compression algorithm. They get written by all kinds of people at all different times for all different reasons. Case law is also law, and that's hundreds of years of people fighting in court lol.
When you hire a lawyer, you're asking them to take all of that and apply it to your hyperspecific set of facts, in such a way that you'd probably win if someone wanted to fight you in court about it.
Also, the stakes in law are high. You usually get one (1) shot at litigating something and that's it. No do overs.
This is not a "source" for a tax system, but a model written after a law, which is the actual source of truth, and not this model.
Also see the german parliament’s laws repository on Github: https://github.com/bundestag/gesetze
This is just a dump of plain text files which are nearly useless without markup and adaptation. For example, when a law refers to some article of another law, there should be a hyperlink, but there are no hyperlinks in the repository. Or when a law contains a term, defined elsewhere, you need a hyperlink to the definition of that term. Or when a law says things like "the rules for operating X are set by an Y Ministry", the word "rules" must have a hyperlink to the latest document defining those rules.
Some paragraphs or articles might be temporary amended/repelled by a Constitutional court, or have some related case law, so there must be links or notes.
Just to give you an idea, here is an example of a "proper" commercial law search system with notes and hyperlinks (in Russian, but Firefox translates it without issues): [1]. I noticed that the German system has no markup, so the German site feels like working with code in a notepad, without coloring and "go to definition" after using an IDE. Or like trying to read law from a gazette/journal of legislative acts.
[1] https://www.consultant.ru/document/cons_doc_LAW_37119/823a1b...
Interesting but there are many laws worldwide that are ambiguous or subjective, hence the reason why lawyers exist.
A big reason that political discussions here get so silly is because so many people here believe that the world can/should work like code, and that the right algorithm/data structures could solve everything. The ambiguity and subjectivity only there because we haven't put programmers in charge of law!
The fact that everyone is jumping through so many silly hoops now that they're writing code non-deterministically should serve as a hint, but it doesn't seem to be.
> should serve as a hint, but it doesn't seem to be
Beliefs, unlike theories, are not subject to disproof. There's a well-understood human tendency of people to react to dissonance by denying reality and intensifying their beliefs. This is a large part of why rational science was viewed with such hostility by the church: adopting the scientific method as a belief is a memetic immunization of sorts against belief cults, which depend greatly on self-reinforcing processes triggered by doubt. Teaching people to doubt belief without intensifying it is the ultimate heresy.
The U.S. in specific does an atrocious job of teaching critical questioning of beliefs, for various reasons that are subject to their own independent arguments. The U.S. in specific also churns out a great many techies who believe that the law is code, and many of those launch into an Icarus parabola under the unyielding sunlight of judicial discretion. One might evaluate a theory of causation between those two factors by studying whether Stanford explicitly teaches how to apply the Scientific Method to beliefs to Business and Finance graduates, and then contrasting the crash and burn rate of graduates among a representative sample of colleges divided along that line.
Very convenient for selective law enforcement.
There's also many countries with no clear laws for things as specific as say recording in public or open pit mining so local lawyers have to string together a bunch of tangentially related laws to make a point. And then each of these laws can be seen as subjective.
It's so simple. Only a fool would pick the cup given to him. Now with stderr.
> Encoded > One source of truth
Oh dear. The one source of truth is the statute.
The act, really; the codified statutory law is an artifact built on top of the various acts passed into law.
Good point. Thanks.
Great! Now I just have to mechanically convert it to TLA+ and the verification results are hopefully exploitable business ideas.