The removed bits discuss habeaus corpus, emoluments, and congressional oversight of the military.
The removed bits discuss habeaus corpus, emoluments, and congressional oversight of the military.
It would be nice if Congress and legislatures actually used git as a matter of law, having names attached to every commit and so on. The way they handle repeals is absurd...
Look at the UK's Hansard[0] as an example. Every word spoken in Parliament leading up to legislation being introduced is tracked and published. Those conversations eventually turn into Bills on the Parliament site[1], and eventually those bills turn into legislation[2]. These websites are all digital versions of the old paper copies which go back centuries.
[0] - https://hansard.parliament.uk
One of it's biggest flaws, actually. Completely understandable, of course, they were working on paper... weren't any reasonable alternatives.
>Look at the UK's Hansard[0] as an example. Every word spoken in Parliament leading up to legislation being introduced is tracked and published. Those conversations eventually turn into Bills on the Parliament site[1], and eventually those bills turn into legislation
I'm ignorant of how things work in the UK's Parliament. But somehow this all seems doubtful. In the US, legislation isn't drafted in such a manner at all.
Laws in countries with codified Napoleonic legal tradition have no problems changing the text of the laws on the book. Yes, tracking the parliamentary debate and proposals would be messy, but tracking the changes in the articles of the law isn’t. This is a simple GitHub version of the German constitution for instance (picking an article that actually has been amended, here the one about the military draft and conscription):
Perhaps add a fixed offset to dates to bring them past the Unix epoch?
You don't provide a single example either. But sure, lets change the entire legislation workflow to conform to Linus's idea of a good product.
Laws consist of strings of characters in a particular sequence, occasionally you don't want to be additive but subtractive, and these sequences are grouped into multiple papers. It's just code. Code on paper, but code nonetheless. They can't easily check what the previous version was, or compare two or more different branches. Lots of strike-throughs and other such crap.
Many of the defects in legislation and the legislative process itself stem from how they try (so desperately) to keep it from being like code. Trying to hide who edited what and when. Fuck, Congress has this big long drawn out process for reconciling two slightly different versions of the same piece of code... with every single act. You've simply never considered this.
>But sure, lets change the entire legislation workflow to
I wasn't changing any workflows. I was just going to dump existing legislation into a repo so I could see what it would look like. Not sure what you're reading into this. You're stomping on an experiment that someone was going to run for free in a way that didn't interfere with you and ridiculing the idea in casual conversation.
>and doesn't even normally use text-files
It's text. On a "file" (which is what we called bundles of associated paper back before computer filesystems, they'd put them in a filing cabinet).
And no one was offering to run this experiment for free that I could see. You were proposing a law. If you are proposing a free experiment, great! more power to you. Now convince the government to actually adopt it.
And yes, the product is text, but the current workflow is not text-file based at all.
Do you know of another version control system that is superior to git?
Is there something special about legislation which makes it incompatible with git, that another version control system would be better at?
There's nothing sacred or special about the sausage-making that is legislation. It is a clunky process done poorly with bad tools.
>But not an argument for git, which even sucks at things like tracking where code is moved to and from.
Which is good, because that's not an issue in legislation. What's important is who authored the code, when it was introduced, what changes have been made to it since, and tracking things like votes/ratification/signing. These are all doable in git. No one much cares if subsection eleventy clause 9b-ii says in subsection eleventy or is moved to umpty-four. I can see how that can be a problem in software, but the processors legislation runs on don't much care.
>And no one was offering to run this experiment for free that I could see.
I don't think dang or anyone else is going back around and ninja-deleting my comments, they are there for everyone to see. Like this one...
>>>I wanted to do the Constitution (and US law in general) in git, but dates before unix epoch weren't possible. I don't think it's since been fixed. I had went so far as to start digging up treaties (about 700 of them, ratified) and draft constitutions (would have been non-master branches of some sort).
Maybe you got my stuff confused with someone else's. The comment tree seems to move around as they're re-ranked.
I'm quite partial to Mercurial myself, and jj seems to be up and coming. But in any event, the idea that even if git were a perfect VC for programming, the idea that it is therefore a good tool for legislation is rather suspect.
And yes, it absolutely does matter when Section 11.2.33 paragraph 6 (in rev 19928), becomes Section 32.1.4 paragraph 8 (in rev 200480). Lawyers and lobbyists get paid good money to track stuff like that.
For one thing, even if one particular person proposes a change, things like cosponsors, exchanged and proxy votes, wide commenting and committee authorship make the workflow, at the very least, not a natural fit for git. Sure, it could probably be shoe-horned in. But a good system to be required by law? Not really.
>>>>I wanted to do the Constitution... > Maybe you got my stuff confused with someone else's.
Speaking of getting one person's stuff confused with someone else's.
Passing new legislation adds, removes or modifies text from an entity's legal documents (US or State code, for example). You see this in bills that say things like "This modifies Section 47, subsection 3, paragraph 2 to say 'people' instead of 'persons'".
So a bill can be thought of as the commit message, and implementation of the bill are the changes to the text. The executive is the "merge approver".
We can likely agree that some of this is an accident, but what we won't agree on is how much given that some of the removal is distinctly useful for this administration. But yeah "haha, the US government has become so incompetent + evil we can't tell what's an accident and what's malicious!"