There was very little "Why" in this article. It seems like:
1. Microsoft bad
2. USA bad
3. I had a gut feeling.
I think the Actions support being so lacking is a deal breaker for us. I know we can bring our own. But I really don't want to manage infra here and take on that responsibility.
The "why" was pretty clear in the article: the author thinks GitHub is doing unethical things with the code the author was putting there (specific things, not "Microsoft bad"). Now you and I may not agree with those things being unethical, but the author does, and I'd say that's a pretty good reason to stop using them.
I would have loved to see that talked about more. I disagree that it's pretty clear. I just read it again and the "Why" (which again is in the title) is lacking for me. I didn't find this article particularly useful or informative. I'm glad you found it useful and informative and I think it's okay that we disagree on that.
This article pretty clearly seems like it is not focused on convincing you to switch, it is assuming you already want to get off of it and is explaining how to migrate.
The title says WHY to ditch github. There's maybe 2 sentences where that is discussed. That's really all I was interested in, "Why should I leave github to something with fewer features and less community support". This article did not justify that for me.
I'm glad you found it interesting and a good read!
The fact you do not find the why convincing is not really relevant to the main content of this article. Some people read more than titles and a few sentences before dismissing it because it doesn't agree with their already-held views. Sorry you did not find it sufficiently interesting. Migrating off of github is an extremely common problem. If you see no need to migrate off of github, why would you be interested in reading this in the first place, let alone commenting on it? Why do you feel you need to be convinced to switch off of it?
> Compare this to a Workflow. All that YAML, all that code from someone else, all that additional complexity in spinning up containers or maintaining runners—just to build a static site? Nah, we've been played for absolute fools.
> While there are definitely valid use cases for Actions, I suspect they are fewer and further between than we've led ourselves to believe.
I use Actions to run automated tests and build documentation previews for my open source projects, which I think are very common use cases. I think this could be one of the largest obstacles to a lot of open source projects moving off GitHub, which when ignored like this reduces the credibility of a "How to Ditch GitHub" article. Don't misunderstand - I'm not saying that moving off GitHub can't or shouldn't be done even in these cases, but there's going to be some friction that deters people.
The point about static sites is pretty valid, but I'd point out that the method the article describes is just obfuscated FTP. Unless you really want the domain provided by Codeberg, wouldn't it be better to use a simpler free web host (not a software forge) and just upload your built site onto it?
For a self hosting solution I'd recommend gitea: it's a single go binary that needs almost no installation (only a config file), it's very fast, updates work great (replace previous binary and restart), supports ldap and has all the features of GitHub your love and even various features that free GitHub does provide (ie branch protection for private repos).
> GitHub's use of everyone's code to train its language models
Unless GitHub knowningly trains on code in private repositories (not that I am aware of), it does not matter where you host your code, or whether it's GPL or MIT licensed. Someone is training their models using your code. That's just a matter of life.
Does anyone have any experience using Radicle (radicle.xyz) for a non-toy project? I ran across it in an HN comment the other day and have been playing with it, but am curious to hear from someone who's done more than that.
There was very little "Why" in this article. It seems like: 1. Microsoft bad 2. USA bad 3. I had a gut feeling.
I think the Actions support being so lacking is a deal breaker for us. I know we can bring our own. But I really don't want to manage infra here and take on that responsibility.
The "why" was pretty clear in the article: the author thinks GitHub is doing unethical things with the code the author was putting there (specific things, not "Microsoft bad"). Now you and I may not agree with those things being unethical, but the author does, and I'd say that's a pretty good reason to stop using them.
I would have loved to see that talked about more. I disagree that it's pretty clear. I just read it again and the "Why" (which again is in the title) is lacking for me. I didn't find this article particularly useful or informative. I'm glad you found it useful and informative and I think it's okay that we disagree on that.
This article pretty clearly seems like it is not focused on convincing you to switch, it is assuming you already want to get off of it and is explaining how to migrate.
The title says WHY to ditch github. There's maybe 2 sentences where that is discussed. That's really all I was interested in, "Why should I leave github to something with fewer features and less community support". This article did not justify that for me.
I'm glad you found it interesting and a good read!
The fact you do not find the why convincing is not really relevant to the main content of this article. Some people read more than titles and a few sentences before dismissing it because it doesn't agree with their already-held views. Sorry you did not find it sufficiently interesting. Migrating off of github is an extremely common problem. If you see no need to migrate off of github, why would you be interested in reading this in the first place, let alone commenting on it? Why do you feel you need to be convinced to switch off of it?
This is currently topical: https://www.businessinsider.com/github-ceo-developers-embrac...
Using alternatives - and asking for (or helping build!) the features that make GitHub so good - makes switching easier for those who come after you.
Let’s let those alternatives flourish :-)
> Compare this to a Workflow. All that YAML, all that code from someone else, all that additional complexity in spinning up containers or maintaining runners—just to build a static site? Nah, we've been played for absolute fools.
> While there are definitely valid use cases for Actions, I suspect they are fewer and further between than we've led ourselves to believe.
I use Actions to run automated tests and build documentation previews for my open source projects, which I think are very common use cases. I think this could be one of the largest obstacles to a lot of open source projects moving off GitHub, which when ignored like this reduces the credibility of a "How to Ditch GitHub" article. Don't misunderstand - I'm not saying that moving off GitHub can't or shouldn't be done even in these cases, but there's going to be some friction that deters people.
The point about static sites is pretty valid, but I'd point out that the method the article describes is just obfuscated FTP. Unless you really want the domain provided by Codeberg, wouldn't it be better to use a simpler free web host (not a software forge) and just upload your built site onto it?
For a self hosting solution I'd recommend gitea: it's a single go binary that needs almost no installation (only a config file), it's very fast, updates work great (replace previous binary and restart), supports ldap and has all the features of GitHub your love and even various features that free GitHub does provide (ie branch protection for private repos).
I prefer cgit. It's used by the linux kernel and many others.
Same. gitolite + cgit, stable combo, less things to worry about.
For those who don't know, Forgejo is a fork of Gitea.
> GitHub's use of everyone's code to train its language models
Unless GitHub knowningly trains on code in private repositories (not that I am aware of), it does not matter where you host your code, or whether it's GPL or MIT licensed. Someone is training their models using your code. That's just a matter of life.
> Unless GitHub knowningly trains on code in private repositories (not that I am aware of)
they absolutely have
Proof?
Does anyone have any experience using Radicle (radicle.xyz) for a non-toy project? I ran across it in an HN comment the other day and have been playing with it, but am curious to hear from someone who's done more than that.
+1, interested in mirroring my repos to Radicle and running a node if it’s not too hard