• 1 Post
  • 8 Comments
Joined 9 months ago
cake
Cake day: March 12th, 2024

help-circle





  • The article is pretty bad.

    It argues in bad faith against GitHub, conflates addressing the reader and their own community (suddenly “we …”), fails to see how despite not being FOSS you can be pro FOSS, names the vendor lock-in but fails to present advantages and disadvantages… I think the tone is pretty bad as well.

    For the most part, I dislike when projects and people self-host. It’s a barrier to me to read and participate. Different interface and UX, no account, different needs for registration, Terms of Service and Privacy Policy, unclear long-term stability.

    I like Forgejo and Codeberg. It has a good and well-known user interface, is fast to use, FOSS, and has a centralized platform., and is working on federation which could ease pain points of distributed and split hosting.

    I find SourceHut UI unstructured; very confusing.

    When GitLab came up, there was a time when I used it for my new projects, and moved some onto there, but eventually moved back to GitHub. Hosting it myself at work; self-hosting it is huge, heavy, bloated.

    The main reasons I still use GitHub are that it is free, feature-rich, fast, familiar, and one platform. I much prefer a low barrier to entry and uniform between projects as long as GitHub acts well enough, even if it is not FOSS itself.

    I would hate to see people follow this article and further spread out FOSS, increasing barriers to entry. I have left exploring projects and contributing because of that barrier on multiple occasions and projects.

    I’m hopeful for Codeberg and Forgejo. Codeberg can serve as a centralized platform already. Should Federation land, it can serve as a base for self-hosted instances, reducing many pain points of a heterogeneous and self-hosted-distributed field.


    Side story: When SourceForge became shit, I created and executed an issue ticket migration for a significant FOSS project. Thankfully we can change platforms like that when you’re not fully locked in but have accessible or natively distributable data.