HI version is available. Content is displayed in original English for accuracy.
Advertisement
Advertisement
⚡ Community Insights
Discussion Sentiment
66% Positive
Analyzed from 3245 words in the discussion.
Trending Topics
#law#laws#legal#read#https#different#more#text#system#code
Discussion (60 Comments)Read Original on HackerNews
https://homepage.univie.ac.at/st.mueller/kafka_english.html
An article on the story: https://courses.cit.cornell.edu/hd11/BeforeTheLaw.html
https://en.wikipedia.org/wiki/Gordian_Knot
So the gatekeeper is the system keeping us from Justice - mostly money, but also other less tangible barriers. In theory, everyone gets a lawyer, in practice some people can afford expensive ones.
Then the end twist got me confused.
Something about how we want to understand The Law, capital letters, but then there's only systems we make ourselves and understand ourselves would feel properly Kafka, I suppose. But you think that would be mapped to journeying towards some kind of Law?
I think it is very deliberately written to be impossible to "understand". If you think you have found its clear and unambiguous meaning, you're wrong.
https://courses.cit.cornell.edu/hd11/BeforeTheLaw.html
We've built the Open Law Platform that governments can use to do this with their official laws. We just launched with the Maryland Secretary of State last month at regs.maryland.gov. You can see the "source code" at https://github.com/maryland-dsd/law-xml, and the "compiled" code at any point in time since they started using our system at https://github.com/maryland-dsd/law-xml-codified and https://github.com/maryland-dsd/law-html. All cryptographically signed at https://github.com/maryland-dsd/law.
It struck me a while ago that almost no present law systems have a regularization or refactoring process. It is an entirely grown object that adds epicycles and it reminds me of the way our household personal agent would record memory until we gave it a self-reflection activity. It seems like a large number of states enacted sunset legislation in the 70s and 80s and many of them got rid of it shortly thereafter. I wonder what troubles they encountered.
One amusing (to me) observation using the model of edits in the OP is that if you say "Section 171 Part 3 is amended to now add the text 'never' at this location" and then later have "Section 171 Part 3 is amended to now add the text 'almost' right before 'never'" you need to have each such edit also be followed by a revalidation of the age of the entire section or you get amusing results due to the text changes.
A fun read for a later afternoon, perhaps.
This is probably bad for some reasons, but I have wondered about a system where the legislature does have to restate the entire legal corpus every session. Like, everything resets at the end of the year and whatever didn’t get passed in the last 12 months is no longer law. Paired with some kind of rule that says you can’t just vote to pass “everything that already exists”.
In my fantasy, there would be less weird cruft in the law, and less bikeshedding about stuff that really should be done.
When you take a measure and tie a control function to it, you make it a target. When you do that, Goodhart’s law applies, and when the specific connection is as simple as “maintaining a criminal law requires charging someone under it with at least X frequency”, the failure mode is obvious.
Now add targeted annotations and other commentary done by specialists into the picture and the real fun begins.
https://igsg-marker.gitlab.io/
(Even if it's only to argue that they aren't all that different in practice.)
https://www.theverge.com/podcast/917029/software-brain-ai-ba...
(fwiw I didn't think that piece hit the nail on the head even though it was generally swinging a hammer in the direction of a nail)
the prophecies of what the courts will do in fact, and nothing more profound, are what i mean by the law.
The reminder that there’s structure and content and they’re different is a good one. Even a small legal document has this microstructure of references, inside, outside explicitly and outside by inference.
If it's your ideas, at least share the bullet points of the prompt it was generated from.
I’ve cadged some rules somewhere for de-LLMing the text; if I can find them I’ll post a follow up link. They don’t help with structure though - and I think this is arguably the most important bit - taking the reader through a logical progression. I didn’t read carefully enough to opine on your structure, but again, thanks for doing this!
IMHO the biggest mistake. It should be like that.
Because right now for mere mortal it's impossible to find out if some law or paragraph is still in effect.
Where and how? Pretty sure our Federal and State Governments here in Australia publish the current in-force law (original Act amended with all amendments) - I assumed that was normal, do they not do that elsewhere?
See for example the Telecommunications Act (1997) [1] but current version amended to October 2025 (the most recent amendment). Then you click ‘All Versions’ and you can read any of the 116 or so versions back to the original in 1997.
1. https://www.legislation.gov.au/C2004A05145/latest/text 2. https://www.legislation.gov.au/C2004A05145/latest/versions
Let’s say you have tax law and the particular law is thousands of paragraphs long. You have multiple changes to the law going on at the same time (totally unrelated stuff in different parts of the law). It’s going to be really hard now because each of the 15 changes need to update its own texts as soon as any other of the proposals get changed and it’s going to affect exactly which order they are voted on etc.
Much easier to just say propasal A changes paragaraph 33 and proposal B changes paragraph 475
Also, not sure what makes it so impossible (debates on whether a given law is in effect seem pretty rare, though it does exist), but that may depend on where you come from and the applicable legal system.
Here is an example: https://jogkodex.hu/jsz/kftv_2013_87_torveny_5995373?r=25
You can check the state of this law at any point in time and even highlight the differences.
I assumed it is so in every civil law country and the law is only unknowable in common law countries where basically every previous court decision is part of the law in some way and they contradict and it becomes a skill and game of twisting words and digging up obscure precedents.
IMO a good middle ground could be attained by everyone having some understanding of the legal system. we could use school for that. i mean, we cover calculus and ancient history, it's not like covering law to some extent would be harder
The approach I take is that every law should expire after a standard, unchangeable time - probably several terms of Congress, say 6 years to account for one full Senate turnover.
Congress can just repass verbatim old laws if they wish - its already written and can be a simple, fast vote. Or we can have debate over outdated provisions like we should have.
I don't (personally) agree with this. Laws should be seen as applying in cases where parties and actions have certain qualities. A retroactive law does not state, "This actually applied to past events and entities." It states, "This applies to entities with the quality of having done an action or met some quality in the past."
I'm not familiar with the EU law system the article is based on. How would it handle a case where a person was found in violation of a retroactive law, and their past violating action was done along with another action that is considered illegal when done during a crime? For example, if somebody wrote that they never used illegal drugs on a government form, and a drug they had used is later retroactively declared illegal, can they now be prosecuted for having "lied" on the form?
Laws come in multiple flavor.
One flavor is for stating something already very consensual in the concerned community. That gives credits to law corpus as a source of justice, so people can somehow be accustomed that law is there to foster social justice through clarified terms visible by all. As opposed to unpredictable judgment of any random citizen along their mindset of the moment.
Then they are laws to throw made up arcana on how laws work. Citizen are still supposed to believe that laws in the end are balanced systems taking many things into account, so mere citizen won't have time, interest, skill and talent to call law by themselves. While they need specialists as middle man, they can still be pretended to be grounded on accumulated wisdom of common sense and consensual underlying basis.
There are also laws, there are their for serving a specific small class in the population, and favor their interest agaisnt all others. As they are well enshrined in the larger boilerplate corpus, there is no way to distinguish them from the others. Well, sometime of course they can be downright clear about establishing de facto inequalities to serve these interests. But they can just as well be thrown in a more diffuse way. Not all law maker are equally subtle of course.
This is not a closed set, many other flavors can be identified, like, "put my name of the glorious me in the title of this law everyone and its dog will have to deal with on daily basis".
Ah, and I almost forgot the flavor that make the most sense to talk about here. Make sure you have enough laws on every matter and the rest to be able to jugulate any layman who is annoying for the dominant class. Capital sentence is not indispensable, it's enough to have enough laws so basically anyone is de facto unable to not be guilty of many things, or at least can be prosecuted until they bind to the "morale order", or suicide.
Over time (~15 years in law and public policy in US and EU) I've come to view that kind of project as something worthwhile, and interesting, but not something that will meaningfully change how laws are created and understood. In the US, as a mixed but substantially common law system, the text of the US Code, which in the LawVM model is the tree, can remain fixed, but the interpretation of that text will still vary over time. (Same syntax, different semantics, which feels like an AI-ism as I write it.)
A couple examples:
I have worked in competition policy. Section 2 of the Sherman Antitrust Act has read the same since 1890 but it's operated across structuralist and consumer-welfare regimes, and the outcomes in Standard Oil, AT&T, MSFT, and more recent cases have all been shaped by those regimes.
US constitutional law may be an even better example of the issue: Article I, sec. 8, cl. 3 (the commerce clause) has not been amended since 1789, but it has expanded and contracted and been reinterpreted over time (most recently with NFIB v Sebelius).
Another example of the problem with a LawVM like system making "the law in force [...] knowable" is US overturning of Chevron deference on 28 June 2024.
Basically, the day before, courts were expected to defer to expert agencies, constrained by the Administrative Procedures Act, in their interpretation of legislation. The day after, courts were expected to make an independent assessment. All sorts of legislative texts were unchanged, but their meaning shifted dramatically: Clean Air Act, Communications Act, etc.
There are a bunch of other issues in figuring out "what the law is", like agency rulemakings in the CFR, and then instruments which sit below those, including sub-regulatory guidance, prosecutorial discretion, etc. Things which fall into this bucket:
- DOJ pot enforcement
- SEC no-action letters (letters which say in effect "we won't try to enforce if you do X")
Altogether, my thinking of this shifted to believing that the real thing that needs to be modeled is not the US Code as a tree serialization format, but the interpretative algorithm (the things lawyers and courts do) that translates that serialized tree into decisions. That interpretive algorithm has a lot of inputs, one which is the serialized tree, but also a bunch of other stuff.
As a final note, I will add that the ambiguity in legislative text is often a feature, not a bug. I have worked on legislative text that was intentionally crafted to provide different plausible interpretations to different coalition members both in the immediate context and in the long term.
In summary: projects like this are good and admirable; they're likely to have more direct utility in jurisdictions on one extreme of the civil law spectrum (i.e. not the US); and in all jurisdictions there likely needs to be an interpretive mechanism that sits above the representation of legal text which has more inputs if the goal is to model what the law is at any given time.
Patent texts read as prose, but are actually precisely structured legal documents. The latest developments in this domain involve LLMs to create and modify patent documents, but even though the legal profession seems to have fallen all in on it, it's essentially rather fragile and error-prone.
I've gone the deterministic direction, which has opened up some very cool, previously unexplored, possibilities!
at that point why not just use something precise like a programming language? have there been efforts in that direction? genuine questions
The name is "law as code".
(My impression: various approaches; mostly academic; some small companies in the space; judging from a loose assessment wrt my career choices as a freelancer: no real business opportunity yet)
A few months ago, for the first time in my life, I had to write a patent document. It was very complicated – too complicated. Noting the structure, I searched for tools, but found only LLMs. So I wrote my own tool.
The amusing thing is, LLMs prefer the DSL-structured document!
A patent document can be represented in a graph. That opens it up to various transformations, refactoring, and validation – all mathematically rigorous! This is far more reliable than asking an LLM to check a document.
Using git enables not only regular diffs, but also structural diffs, which compare legal elements rather than just lines.
The LSP (yes, that too!) makes drafting much easier, with autocomplete and validation as I type.
I plan to open-source the DSL, and the tool that processes its files and outputs jurisdiction-aware, nicely formatted documents...