Announcement

Collapse
No announcement yet.

GNU Grep & Sed: Fallout Within The GNU FSF Camp

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

  • #16
    Originally posted by kiwi_kid_aka_bod View Post
    Try taking a look for grep, sed, and all the other GNU command line tools in your init scripts.
    Not going to disagree with your post (you're spot on) but in response to that one sentence: systemd.


    Originally posted by pingufunkybeat
    Remove all GNU software from your computer (not GPL-licensed, not FSF-endorsed, only the "GNU operating system" stuff) and see how far you get with that Seriously, try it, just for a laugh.
    I'd actually be rather surprised if it wasn't possible to boot an entire modern Linux-based OS with zero GNU components these days. The shell environment would likely be rather unpleasant, but then the cutting-edge GNU/Linux shell environment is already unpleasant to 99.999% of human beings, so no big loss there. I bet you can boot up Linux, get into X11, and have a working KDE desktop without any "official" GNU components, right now.

    Obviously, _building_ a modern Linux-based OS without using GNU software is still utterly impossible.

    Comment


    • #17
      Originally posted by elanthis View Post
      Not going to disagree with your post (you're spot on) but in response to that one sentence: systemd.




      I'd actually be rather surprised if it wasn't possible to boot an entire modern Linux-based OS with zero GNU components these days. The shell environment would likely be rather unpleasant, but then the cutting-edge GNU/Linux shell environment is already unpleasant to 99.999% of human beings, so no big loss there. I bet you can boot up Linux, get into X11, and have a working KDE desktop without any "official" GNU components, right now.

      Obviously, _building_ a modern Linux-based OS without using GNU software is still utterly impossible.
      Let's see:
      1. glibc -> uclibc/musl/bionic: I doubt this works as well as I wish it did.
      2. init usually has some shell scripts -> needs sh, coreutils or replacement:
      bash -> dash/mksh/zsh/ash
      coreutils -> busybox
      systemd might avoid this, but I doubt it's to the "boots without a shell installed" point.
      sysvinit is nongnu, though.
      3. X11: not GNU, may need shell; see 2
      4. KDE: similar; while Alpine has Qt, they don't have kde. This may be #1 kicking in.
      Some other desktops may be viable, though.

      Re: building a linux distro:
      I'm fairly sure that binutils and gmake are the only absolutely necessary GNU software.

      Comment


      • #18
        Originally posted by johnc View Post
        GNU seems like such a relic of the past now.
        Except of course it's software is default in pretty much every Linux installation out there, not to mention being supported or coming straight out of the box in other operatings systems like the BSD's, Haiku, Reactos, etc

        Of course your response was not a reflection of reality, only your dislike for GNU, FSF and GPL.

        Comment


        • #19
          Originally posted by pingufunkybeat View Post
          Anyway, as was commented elsewhere, this is a tempest in a teapot. A maintainer stepped down, but made it clear that he still supports the goals of the project. Big deal...
          Yes, but Michael will jump at anything which can further his anti-FSF agenda. And since the real thing wasn't as infective as he'd hoped, he decided to make up a 'FALLOUT within the GNU FSF camp' title while cherrypicking excerpts. What a surprise that the line where the author states that:

          'I do support the ideas behind the FSF as strongly as ever; and I am grateful to the FSF staff for the support I have had since I joined the GNU project in 1999.'
          - didn't make it into Michael's summary.

          Comment


          • #20
            Originally posted by benjamin545 View Post
            FSF blurting out something about a powerVR driver yet has mothing to show for it?
            And who else does? They've drawn attention to PowerVR as something which lacks open source driver support, that doesn't mean they have the resources/information to do it themselves.

            Originally posted by benjamin545 View Post
            how long before a couple fresh out of collage programmers would come up with a more feature ritch replacement written in a more maintainable codebase that was more easily integrated with modern componants.
            Go ahead, what's stopping you or anyone else from doing this, yet somehow this has failed to materialize. Probably because you know, the existing tools work very well and that people see little need of rewriting them in <insert high level language here>.

            Comment


            • #21
              Originally posted by Ibidem View Post
              I'm fairly sure that binutils and gmake are the only absolutely necessary GNU software.
              Bison, Flex, Gtk+, automake, autoconf, bash, coreutils, cpio, ed, ...

              The list is rather big. However, many of those projects were put voluntarily under the GNU umbrella. They were not developed by GNU.

              Comment


              • #22
                Originally posted by Marc Driftmeyer View Post
                In all seriousness, Linux would not be what it is today without billions of dollars from Corporate Development.
                This don't excludes the GPL stuff, but what woud be if the linux kernel would BSD or something alike licensed?

                Comment


                • #23
                  Originally posted by elanthis View Post
                  I'd actually be rather surprised if it wasn't possible to boot an entire modern Linux-based OS with zero GNU components these days.
                  It should be possible (BSDs get most of the way there, after all), but there is a good reason nobody does it

                  The vast majority of Linux distributions are very closely tied with GNU. Remove GNU from Fedora or Ubuntu and they will not boot. Making them boot again would be quite a lot of work.

                  My point was that just because somebody doesn't start bash and go sed-ing and grep-ing their way through source code, doesn't mean that they don't depend on GNU software every day.

                  Comment


                  • #24
                    A GNU fork to a more flexible umbrella ?

                    Comment


                    • #25
                      Originally posted by pingufunkybeat View Post
                      It should be possible (BSDs get most of the way there, after all), but there is a good reason nobody does it

                      The vast majority of Linux distributions are very closely tied with GNU. Remove GNU from Fedora or Ubuntu and they will not boot. Making them boot again would be quite a lot of work.

                      My point was that just because somebody doesn't start bash and go sed-ing and grep-ing their way through source code, doesn't mean that they don't depend on GNU software every day.
                      Wait, hasn't Ubuntu been using dash for many years now?

                      https://wiki.ubuntu.com/DashAsBinSh

                      BSD is almost completely free from GNU. I think gdb and ld are the only GNU components remaining ... and LLVM has infant projects for these two (not sure if BSD will pick them up though).

                      Admittedly, the newer BSD tools have less features and some even had performance problems (e.g. like BSD grep), but they're maturing

                      Comment


                      • #26
                        Originally posted by nslay View Post
                        Wait, hasn't Ubuntu been using dash for many years now?

                        https://wiki.ubuntu.com/DashAsBinSh

                        BSD is almost completely free from GNU. I think gdb and ld are the only GNU components remaining ... and LLVM has infant projects for these two (not sure if BSD will pick them up though).

                        Admittedly, the newer BSD tools have less features and some even had performance problems (e.g. like BSD grep), but they're maturing

                        Comment


                        • #27
                          Originally posted by crazycheese View Post
                          If they really wanted to use BSD code, then so be it. It wouldn't affect the existing BSD projects in any way (i.e. they would still be open source).

                          BSD code will live on to solve real problems in corporate, government, academic and open source environments (and already has). I mean, imagine the applications of LAPACK and BLAS (used by everyone) for example ... now try imagining the same for GSL.

                          Comment


                          • #28
                            Originally posted by kiwi_kid_aka_bod View Post
                            As to the idea of college fresh programmers coming up with something comparable in reasonable time? Hogwash! It would probably take a dozen years while they rewrote the code dozens of times based on increasing experience to fix all their early mistakes.
                            i made a usable, mostly feature complete, unix standard dd in assembly, and i am noob
                            took me... idk 7 weeks avg 1-2 max3 days a week couple hours a day (most of it was reading syscall codes and standards)


                            years ?
                            for a better programer then me ?


                            other GNU programs probably need years, but coreutils/binutils probably dont

                            Comment


                            • #29
                              Originally posted by elanthis View Post
                              ...but then the cutting-edge GNU/Linux shell environment is already unpleasant to 99.999% of human beings, so no big loss there.
                              As another poster pointed out.... completely and utterly false. Even going from GNU grep to Solaris grep is like trading (Transformers) Bumblebee for Triumph Stag, heaven forbid awk or something /slightly/ swiss-army-knife-ish.

                              Here though lies a true statement: Most GUIs that are pleasant to 99.999% of human beings are 99.999% unpleasant to people that actually want to get something real done.

                              Comment


                              • #30
                                Originally posted by nslay View Post
                                If they really wanted to use BSD code, then so be it. It wouldn't affect the existing BSD projects in any way (i.e. they would still be open source).

                                BSD code will live on to solve real problems in corporate, government, academic and open source environments (and already has). I mean, imagine the applications of LAPACK and BLAS (used by everyone) for example ... now try imagining the same for GSL.

                                Comment

                                Working...
                                X