3 ms·

All of this is an argument for version control--which I wholly support, and did from the first message. But not an argument for git, which even sucks at things

All of this is an argument for version control--which I wholly support, and did from the first message. But not an argument for git, which even sucks at things like tracking where code is moved to and from. Yes, yes, you can use "-C -C -C" if you really want to know and have time to wait, but that is terrible compared to systems that track moves as first class happenings. 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.

NoMoreNicksLeft 1y ago
>All of this is an argument for version control-- 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.
- compiler-guy 1y ago
  > Do you know of another version control system that is superior to git? 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.

NoMoreNicksLeft 1y ago

>All of this is an argument for version control-- 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.

compiler-guy 1y ago
> Do you know of another version control system that is superior to git? 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.