Announcement

Collapse
No announcement yet.

System76 Moves Ahead With Writing Their Own OS Installer

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • #21
    why not adopt/adapt Calamares?
    The idea of Calamares arose from a desire of several independent Linux distributions to come together and work on a shared system installer. Instead of everyone working on their own implementation and forking forks of forks, why not work together on something that can be used by many?


    to expand on this, wouldn't it have been better to join in that effort rather than reinvent? Or even just rewrite calamares in Rust if that's their thing.
    Last edited by speculatrix; 11 September 2017, 06:33 AM.

    Comment


    • #22
      Originally posted by speculatrix View Post
      why not adopt/adapt Calamares?
      The idea of Calamares arose from a desire of several independent Linux distributions to come together and work on a shared system installer. Instead of everyone working on their own implementation and forking forks of forks, why not work together on something that can be used by many?


      to expand on this, wouldn't it have been better to join in that effort rather than reinvent? Or even just rewrite calamares in Rust if that's their thing.
      Calamares is a massive "framework" that combines both front- and backend in an opinionated way and lets developers extend it via plugins.
      Some of the backend steps are closely tied to the UI.

      System76's approach is to decouple backend from frontend development and provide an easily consumable API. Personally, I think the latter approach is superior.

      Comment


      • #23
        Originally posted by unixfan2001 View Post

        Calamares is a massive "framework" that combines both front- and backend in an opinionated way and lets developers extend it via plugins.
        Some of the backend steps are closely tied to the UI.

        System76's approach is to decouple backend from frontend development and provide an easily consumable API. Personally, I think the latter approach is superior.

        thanks

        Comment


        • #24
          Originally posted by bug77 View Post
          I don't think it's NIH as much as them needing to provide timely support for everything they offer. Because we all know that if you can't upstream your patches fast, the work you end up doing maintaining them separately, rivals writing from scratch anyway.
          Just like you, I have no stake in this, but I wish them the best.
          While submitting patches to projects controlled by third parties and waiting for them to be accepted seems like a lot of work...
          Writing your own software 'from scratch' (ok, i'm sure they reuse lots of code) and maintaining that doens't look like a small feat either...

          Comment


          • #25
            At least they wrote it in Rust

            Comment


            • #26
              Originally posted by nomadewolf View Post

              While submitting patches to projects controlled by third parties and waiting for them to be accepted seems like a lot of work...
              Writing your own software 'from scratch' (ok, i'm sure they reuse lots of code) and maintaining that doens't look like a small feat either...
              Submitting patches is easy. The problem is what you do until those patches are merged. Best case scenario, you have to work a few things out and that's it. But more often than not, there's nobody available for a timely review, or you get rejected for whatever reason and the whole process can take months. And this whole time you have to also maintain your work downstream. Which is still manageable for a few changes, but once you reach a dozen or so, things start looking a little less rosy. It's still doable, of course, but it depends on how much resources you can afford to put into it.

              Comment


              • #27
                Or you pay the open source project so that your patches get reviewed in priority.

                Between paying a couple thousands of dollars for that and paying a hundred thousand dollars for a dev, which is smarter?

                Comment

                Working...
                X