I'd guess that this is more due to the popularity of GitHub among potential contributiors to these projects. I've seen it mentioned many times here on HN that a potential contributor may not want to put in the effort to create an account on another VCS provider or send a patch to a mailing list, whereas submitting a pull request on GitHub is a much less involved process for them, given that they likely already have an account and are familiar with the process.

However you could definitely argue that given that the projects in question are alternatives to GitHub, a developer interested in contributing will likely be invested enough to sign up for the alternative service.

"I'd guess that this is more due to the popularity of GitHub among potential contributiors to these projects."

Doesn't that support bmsleight's point? If Github is more popular than NotGithub, even among the developers of NotGithub, that doesn't exactly inspire confidence in NotGithub.

Oh, it definitely could be considered as evidence that Github is a better platform for code hosting than gitea, gogs, etc. However I think that the difference is popularity is far more likely to be caused by network effects around Github as well as Github's comparative age.

There are also other factors impacting a users's decision to choose one hosting platform/software over another unrelated to the functionality of the software, such as code licensing and self hosting capability.

In summary, popularity certainly can be an indicator of whether or not a software or service is worth using, but it's not a direct measure of the quality or robustness or whatever other factors you may care about when making a decision.

No, it is about making a compromise because you live in the real world.

This isn't a social network. There's more to a source code hosting platform than popularity.

For me it's the ability to self host on cheap hardware. For that Gogs gets my support.

Please look at: https://kallitgea-scm.org http://fossil-scm.org Both eat their own dog food.

Since I hadn't heard of it and had to Google: https://kallithea-scm.org/ (typo in the link above)

https://code.rhodecode.com also, hosting RhodeCode sources.

Kallithea is a fork of RhodeCode too:

https://en.wikipedia.org/wiki/RhodeCode#Move_to_open_source

It could be argued that this is smart. Imagine you merge something that, as a simple example, renders pull inaccessible. Your authoritative source for the fix would then be itself inaccessible. People running that code would be in the dark as to what is wrong, or may struggle to report problems (if issues become inaccessible). Manual steps would be involved because their CI/CD might also break.

You'd expect them to never run the latest code in prod to avoid exactly this, but a cherry-picked vulnerability fix is one example of how this could realistically happen.

The implication that these services are an "alternative" to GitHub is only coming from the blog post, none of them try to directly compete with GitHub as a platform and service. Instead they provide software mostly for self-hosting a GitHub-like service.

There's no point in running the service on your own, instead you need to go where the people/developers are and where you can be easily discovered.

I guess Gogs have an excuse, as they had to host the project somewhere, and perhaps they haven't started dog fooding because of migration difficulties, e.g. the issue tracker.

But I wonder what's the excuse for Gitea as they're a pretty recent Gogs fork?

They're both pretty amazing though: Self-contained binaries and less than 5 minutes setup time, so totally suitable for ad-hoc deployment.

Github also does not cost time and money to use for open source projects.

I've used a few GoGS servers, but none of them were public. Hosting a public competitor to Github is a different set of skills to developing a self-hosted VCS server.

He's keeping it real, going with Apache Allura.

Here's an explanation: https://lobste.rs/s/gokjbo/gitea_1_1_0_released#c_dg9pwe

Well, there is a clear rationale behind that. This was brought up in another HN comment thread [1]:

https://github.com/go-gitea/gitea/issues/1029

https://lobste.rs/s/gokjbo/gitea_1_1_0_released#c_dg9pwe

[1] https://news.ycombinator.com/item?id=17245323

https://notabug.org/ does eat its own Gogs fork^W^W dog food [0].

[0] https://notabug.org/hp/gogs/