Announcement

Collapse
No announcement yet.

Rocky Linux Shares How They May Continue To Obtain The RHEL Source Code

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

  • #31
    Originally posted by timofonic View Post

    There's SUSE...
    where the free developer license?

    * oh yeah zero.

    what is going to happen after sles15-sp5 ?

    opensuse leap 15.6 will match binary with what os version ???
    Last edited by onlyLinuxLuvUBack; 30 June 2023, 02:00 AM.

    Comment


    • #32
      Originally posted by rawr View Post
      Software inherently profits from work done already since copying and replicating is cheap. That's why open source wants to be free, so that everyone can profit of the work already done.
      Of course this is in direct conflict to maximise profit.

      Cooperation and competition are both valid models. Open source chooses cooperation, closed source chooses competition.

      Cooperation favours the community, Competition favours the successful individual and is often opposite to the interests of the community.
      What made Red Hat great is that they found a perfect balance between these two models, giving a lot back to the community and at the same time making a lot of profits (and constantly growing) from the paying customers.

      Unfortunately the IBM acquisition has completely distrupted this balance.




      Comment


      • #33
        Originally posted by Chewi View Post
        I don't like this. It won't end well and will just make a bad situation worse. Instead of trying to find loopholes, they should just respect Red Hat's choice and move on.
        Exactly. If your only reason to exist is to copy and repackage, while the source company doesn't want that, well, that is an awkward position to be. Let's move on to another distro, or offer a further stabilized version of CentOS. If you really need binary equivalence, then pay for RHEL.

        Comment


        • #34
          Originally posted by rawr View Post
          Companies distribute binaries that are integration tested against RHEL. Sometimes there exists code that relies knowingly or unknowingly on buggy behaviour. Ever fixed a bug to find the fix triggers another bug?

          "bug for bug" they replicate the behaviour exactly. You wouldn't fix the bug in the distro, but push it upstream.
          Also often those bugs are just not known.
          I understand your reasoning, but it has largely been irrelevant since containers took over. Every self respecting server program is being distributed as a Docker container nowadays, (and there are various attempts to use containers in Desktop, with varying success).
          Even Red Hat is pushing for containers with various projects like Podman and friends, OpenShift, Quay etc.. which renders the host distro evermore irrelevant.

          Comment


          • #35
            A lot of people commenting that RH has the right to impede other to "copy and repackage" its work seems to forget where this work actually come from: the Fedora community.

            The new RH attitude is going to have a negative impact on that community as nobody want to do something in open source seeing the product of his work being close.

            In turns this will become either an increasing cost for RH or a decrease of quality of their package.

            In both ways, this is the beginning of a death spiral for them.


            Comment


            • #36
              Originally posted by Vermilion View Post

              I understand your reasoning, but it has largely been irrelevant since containers took over. Every self respecting server program is being distributed as a Docker container nowadays, (and there are various attempts to use containers in Desktop, with varying success).
              Even Red Hat is pushing for containers with various projects like Podman and friends, OpenShift, Quay etc.. which renders the host distro evermore irrelevant.
              That's a good point. But I argue against it being irrelevant for now.
              I see three points:
              1. You still need a host system. So also container management needs to be integration-tested.
              2. "Client software" still faces some challenges being containerized and would need to be written and tested with that in mind. I imagine proprietary EDA software like Quartus and Modelsim. Yeah I know flatpak and snap exist but IMO they still need to improve.
              3. What do you run in the container? "Bug for bug" compatible integration-tested software. So probably you would use RHEL within the container. Ideally you run a minimized version of everything, but would take more effort.

              I don't like the term "bug for bug" either, but the idea is reasonable to some extent. They probably should have used something like "behaviourally identical".
              Last edited by rawr; 30 June 2023, 05:05 AM. Reason: Wording, "argue against"

              Comment


              • #37
                Originally posted by cynic View Post
                A lot of people commenting that RH has the right to impede other to "copy and repackage" its work seems to forget where this work actually come from: the Fedora community.

                The new RH attitude is going to have a negative impact on that community as nobody want to do something in open source seeing the product of his work being close.

                In turns this will become either an increasing cost for RH or a decrease of quality of their package.

                In both ways, this is the beginning of a death spiral for them.
                Also had that thought. Also I mean Red Hat exists for a long time now. So I assume they found a model that works economically.
                I wonder if the change is based on Oracle being a (perceived) threat, greed or both. If Oracle destroys Red Hat, they have nothing to leech from.
                Just thinking out loud

                Comment


                • #38
                  Originally posted by mxan View Post

                  It was over the second IBM bought Red Hat. Everyone saw this coming a mile away. Giving Red Hat a second chance (whether by just coughing up the bucks for RHEL, or by using other clones instead of just ditching them entirely and settling for Ubuntu or Debian) was a massive mistake.
                  This has been parroted mindlessly these days. My question is: what does Ubuntu or Debian offer that Red Hat doesn't? CentOS already has 5 years of full support. For the 10 yeas support, Debian's is just not good enough, restricted to a few set of packages, so a no-go. For Ubuntu, you'd have to pay Ubuntu Pro (so, similar to a RHEL license). You may claim Ubuntu is cheaper (it might be, IDK), but seriously, for the free-tier, CentOS Stream is just as good as the other ones if not better.

                  Comment


                  • #39
                    My question is: what does Ubuntu or Debian offer that Red Hat doesn't? CentOS already has 5 years of full support.
                    They didn't end support early for their distros like CentOS did. You can go with CentOS stream but what guarantee you have that they will not end support early? They did it once...
                    I think that the whole problem is not that RHEL did what they did (Close the supply of their packages), I think that the problem is in the way they did it. They should clearly communicate that they are going to cut off 3rd party distros and that they will be closing any loophole.

                    However saying this, RH needs to remember that they are also redistributing OpenSource for money, if all software did what they are doing now they would be screwed as well.

                    Comment


                    • #40
                      Originally posted by cjcox View Post
                      We do use CM (ansible, sure wish Red Hat didn't "own' that) where I work. And while some of it is specific to a Red Hat-like world, we could adjust those few things and probably re-deploy fairly quickly.
                      For the record, the only reason I chose Ansible over its competitors is that it didn't require a daemon running on the machine to be controlled and, given what a bloated hog of a Python mess it is, I'd jump ship to a Rust-based alternative in a heartbeat if someone else would go through the hassle of reimplementing all the idempotent "This is the state I want it to be in. You figure out how to get there without messing up stuff that needs no changes," bits.

                      Comment

                      Working...
                      X