Conversation

Notices

1. Alexandre Oliva (moved to @lxo@gnusocial.jp) (lxo@gnusocial.net)'s status on Monday, 06-Dec-2021 13:10:14 EST Alexandre Oliva (moved to @lxo@gnusocial.jp)
   stability is a very overloaded term. it's often used to mean either "doesn't crash much", "doesn't change often", "doesn't change much at a time", and combinations thereof. there's a contradiction among these terms already: changing a little at a time implies changing often (or lagging farther and farther behind). not crashing much is a matter of QA, but also of how much customization there was: the more one departs from the shipped bits and configurations, the more likely one is to hit a case that QA didn't cover (the combinations of possibilities grow exponentially). rolling releases take smaller bumps that may break stuff, compared with distros with identifiable releases, but some such releases are sometimes little more than global rebuilds with a different toolchain, which still has some potential to break stuff. as usual, issues are nuanced, rather than binary, and require deeper analysis rather than shallow hard rules
   1. lnxw48a1 (lnxw48a1@nu.federati.net)'s status on Monday, 06-Dec-2021 13:47:54 EST lnxw48a1

      in reply to
      @lxo Stability as not crashing may sometimes come in the form of restricting the set of supported hardware and software variations to a relatively small number of known good combinations.