Comment by kasajian | Hacker News Reader

Comment by kasajian | Hacker News Reader

He has a right to do anything he wants with hit source, but Open Source is nothing more than a license. We had GPL 2 before there was github. All you had to do is either deliver the source with your software, or make the source available upon request, even if it's by snail-mail.

My only exception to this article is the section, "People underestimate what open source costs from the maintainer side. It’s not just writing code. It’s issues, pull requests, discussions, people demanding things from you for free. I got spammed like crazy. And now with AI, spamming a maintainer is so easy: low effort issues, generated pull requests, and every single one of them takes a bit of your time and energy"

This is all self-inflicted. Don't take Issues. Don't take PRs. Make the repo read-only to everyone but you. There's zero effort, unless you're planning on keeping all the source on your computer and never push it anywhere. Well, I guess, in that case you're saving yourself from doing a push.

So if you don't want to do all that stuff, that's fine. You don't have to. But don't pretend the only alternative is to close the source.

I mostly agree, but from experience, when you readonly your repository, you encourage a bunch of forks, then people ping you in those based on git blame, and many other out-of-band contacts. But I think that cat is out of the bag anyway.

A terminal is critical software, you type passwords into it, it sees all the output, you need to trust it, so while I'm sad to find out about this project only today, the closed-source is indeed a deal-breaker for me

> you encourage a bunch of forks, then people ping you in those based on git blame, and many other out-of-band contacts

Not that this really addresses the point that you were making, but at least the git blame part is pretty solvable by just using a throwaway email for the commits. If you're already not caring about getting contributions, you could even rewrite history to amend everything to use the new email retroactively!

Nothing inspires trust like a fake email address that will be ignored, and also make sure that you don’t cryptographically sign your commits with a public key that could doxx you

/s

There is no pretense here that closing the source is the only alternative. It's only the alternative the author has chosen.

For better or worse we live in a society. Even being perceived and having a simulacrum of you in someone's head, much less a collective simulacrum in many people's heads is restrictive. Every facet of yourself you reveal to is another string tied upon you. I'm not here to argue one way or another what you should do, but it is not self inflicted, it is a fact of social existence. Even to have the discipline to ignore an email can be taxing, especially if you view the person on the other end as a human. Just being made aware of a bug, or a different way of structuring things is an infohazard if your goal is to just have fun doing things your way. I think closing the source like this is a fairly drastic course of action, but there's plenty of things I've written that aren't open source for the same reasons, even though I'd like them to be, I'd rather be a hermit about it just because the vibe is better.

Yep. People seem to have forgotten about the cathedral vs bazaar. Somehow in many people's minds open source implies bazaar but both cathedral and bazaar are both free/open source. The author just wants to go to a cathedral style project, which is perfectly understandable. There's no need to make it closed source.

100% agree. Especially since you wouldn't have to start taking his word for the things he explicitly lays out in this article, that previously were checkable at no additional cost to any party

I thought that GitHub supports some sort of locking repo, such as not allowing PRs? As X allows comments from accounts mentioned in tweet.

Disabling PRs and issues (and wiki and a bunch of other stuff) is indeed supported by GitHub AFAIK.

Note, though, that disabling PRs in particular is only a couple months old (and was prompted by the current wave of consternation around AI-assisted contributions). Before that you had to set up a bot that would just auto-close them and it still wasn't obvious in the UI that this was how it worked.

This also is mostly a bandaid, but my recollection is that for years there's been support for picking a different email for notifications for specific repos (and if not, I'm positive there has been a setting per org at least for that). Using a throwaway email for notifications for a repository that you don't care about also seems like a pretty reasonable way to use it as a place to dump code without needing to care about being externally reachable.

Maybe people will see the giant list of issues and PRs and be concerned, but if you put a note at the top of the README indicating that's what's going on, the only people who won't notice (or won't accept it) it are the ones that you're probably trying to avoid with this strategy anyhow, so it's no loss to have them go elsewhere.

You can restrict PRs only to 'collaborators' of your repository, or you can disable them completely.

We also had open source before we had GPL, that is how UNIX got widespread anyway, in source tapes.

Amiga had the Fish disks, PC had all the whateverware variants during its MS-DOS days.

8 bits naturally had the magazine listings to type in.

er, no, it was licensed, but available to educational institutions. I remember having to send Bell Labs our university calendar when i wanted to get a copy of LSI-11 Unix circa 1976. What is true is that Ken Thompson visited a number of places over the years, and maybe from time to time he forgot to pick up his Unix install tapes when he left.

Bell Labs/AT&T claimed exclusive ownership of Unix, and the exclusive right to license it, right up to the BSD case in the early 1990s, where it was determined that AT&T had used Berkeley code, and vice versa, to the point where copyright notices were essentially fiction. Berkeley Software Design had to remove a small number of files from their distribution, and then System V and BSD could go their own separate ways.

Reply on news.ycombinator.com