Friday, December 25, 2009

Proposal for a new Linux distribution

I have decided my life isn't interesting enough, so to make it more interesting I've decided to create Yet Another Linux Distribution (YALD). This YALD will have the name ... drum roll please ...

Fairy Linux

Named after the Fairy Penguin of Australia, it is meant to be yet another 'lean' distribution of Linux. Here's what I had in mind as a beginning set of requirements;
  • Fit in, complete, at around 200MB. That's installed.

  • Based on the latest kernel, whatever that happens to be. That would be something in the 2.6 line, and greater than 2.6.32.

  • Based on the latest glibc, whatever that happens to be, and not one of the embedded/limited substitutes found in other small distributions.

  • Based on one, and only one, desktop environment. I'm seriously thinking of KDE 4 (seriously). There is a strong reason for this, primarily that KDE uses Qt, which will make a good foundation for another choice I have in mind.

  • Stripping out as much of the current software as possible, especially the duplicates.

  • A limited CLI userland. I had thought of using Busybox, but after reading about how litigious the current project has turned out to be, I've decided to go with the regular tools. I have half-seriously considered a comment that someone should take the BSD tools and create an alternative to Busybox. That would be a sub-project.

  • Make sure that Ruby and Ruby/Qt are installed. The complaint has been made, more than once, that there is no equivalent to an easily approachable programming tool like Basic/QBasic/Visual Basic. I've used Ruby and Ruby/Qt; the tools and bindings are there to build reasonably sophisticated (and fun) graphical applications without having to pull out the C++ and regular Qt libraries.

  • Target x86 systems to start with, but I would like to eventually migrate to ARM exclusively.

  • Make sure that all necessary drivers are with the distribution for the best user experience possible. For x86 that means nVidia and ATI. I'm well aware of what happened in the case of Kororaa, but the commercial modules will go in and stay in. As for the GNU purists, I don't give a damn what they think or how they feel.

  • Base all of this on an existing distribution. Although I've built, from scratch, embedded Linux systems for experimental purposes, building a full-up and installable distribution is non-trivial. It makes a lot more sense to base off of an existing distribution. The base distro I have in mind is Linux Mint. Linux Mint is just that good.
This is just a starting list; I'm sure I'll come up with more requirements, and then they'll need to be racked and stacked for importance. Only after that will I make the decision to go through with this or not. I need some ROM as to the level of effort and resources required, rather than jumping in feet-first.


  1. -- 200 megs installed? With your laundry list that may be quite tough.
    -- Have you considered 64bit? Both Risc and x86 small chips are steadily moving towards 64bit.
    -- KDE/QT, very cool! However, KDE alone is probably going to make it difficult to fit into 200 megs.
    -- The Ubuntu netbook remix/moblin may be worth looking at.

    Have a good one.

  2. Have you thought about basing it on arch linux? Just did a search on your blog, didn't find anything on arch. I think arch would be a perfect candidate for the base system. I think arch has their modularized version of KDE so one can install just the base without any kde apps and then build up from there. I have no idea how you would make it work on an arm system out of the box as it only supports i686 and x64.

    Some negatives might be:
    - Small userbase (compared to the big distros)
    - Not that popular (still ~10 on distrowatch)
    - Hence there might not be enough info/help on the net

    I do like the idea of a small, lightweight distro with kde. Good luck.

  3. *obligatory 4chan "install gentoo" spam here*

    That probably wouldn't be the best idea though, as you are going for user friendliness and gentoo is just ridiculous to mess with.

    200MB is quite a goal, but even if you have to go to like 400 or more, no big deal.

    I LOVE the fact that you plan on cutting the included software out. Let the user decide what they want. What I hate about ubuntu is that I always end up deleting half of the base software and replacing it.

    You got some really good ideas and I am interested to see how this will go. Keep it up.

  4. I also would love to have my own remaster but I would prefer to base it on Debian; however I cannot find anything that makes for easy remasters like is possible with the various *buntu tools out there. I have no trust in the stability of the *buntus (and hence Mint), but if there was a Debian remaster tool that would be wonderful...

  5. From Debian came Ubuntu, from Ubuntu came Mint and from Mint will cometh Fairy :)

    Waiting for your first release, all the best! I guess based on Ubuntu and Mint tradition the release will be code named something beginning with 'A'.

  6. I'm seriously thinking Arrogant Ass.


All comments are checked. Comment SPAM will be blocked and deleted.

Note: Only a member of this blog may post a comment.