Things I've learned from Swift Linux, possible new directions

[Jason Hsu]

NOTE: If you think you can do a better job of running Swift Linux than I can, please contact me.  I know that many of you who keep up with this list can run circles around me blindfolded.

I started Swift Linux two years ago.  It hasn't taken the world by storm, and I'm surprised it's still ranked 99 on Distrowatch given that I haven't made a new release since last spring.  If I knew just how much work it would be, I never would have started.  However, I do not regret what I've done, as I learned so much that no book, course of study, or Hello World exercise could have taught me about programming or software development.

I've learned a lot from basing Swift Linux on antiX Linux and then on Linux Mint Debian Edition, but I now better understand why most distros are based on a parent distro that offers a version WITHOUT a pre-installed DE.  As I found with basing Swift Linux on Linux Mint Debian Edition, it's EXTREMELY difficult to remove GNOME due to numerous dependencies.  Most of my counterparts at other distros prefer to start off with NO DE and avoid the bloat and the potential for bugs.  antiX Linux is the only distro I can think of that replaces the DE from the parent distro (KDE from MEPIS Linux).  I suspect that KDE is MUCH easier to remove than GNOME, and anticapitalista not only has more know-how than I have but also special know-how that most other distro leaders lack.  Thus, on the Debian side of the distroverse, there are numerous distros based on Debian itself or on Ubuntu (which offers an alternate installation ISO, which allows a minimal installation), but very few based on other distros.  The only other distro I can think of that is based on Linux Mint is the Commodore OS, which retains GNOME.  antiX Linux is the only distro based on MEPIS.  You don't see distros based on CrunchBang, PureOS, ZorinOS, Pinguy, etc.

I've learned the hard way that having to remove the DE from your parent distro is probably not the way to go.  That said, I had much less know-how when I first started Swift Linux, and it was easier to just outsource so many important things (like the installer, remastering script, etc.)  to the parent distro rather than create something new from scratch.

I've identified a few possible paths for the future of Swift Linux:
1.  Switch to Debian Stable as a base, LXDE as the DE, and Remastersys for the remastering script: Swift Linux is intended to just work (and work well) on 10-year-old computers.  Keeping up with Debian Stable is easier on the development side and the user side than keeping up with Debian Testing.  Debian Stable is also lighter than Debian Testing, and this is important for keeping the ISO small enough to fit onto a CD for 5 more years.  LXDE has more support and help available than IceWM.  SliTaz Linux (which is even lighter than Puppy Linux and antiX Linux) proves that LXDE can be very lightweight indeed.  I've heard great things about Remastersys as a remastering script.
2.  Find a new lead developer for Swift Linux: As I stated before, I'm not the most qualified person to run Swift Linux.  If you think you can do a better job (because you can start with minimal Debian/Gentoo/Arch/Slackware/etc. and create a polished and user-friendly setup comparable to Linux Mint), please contact me.
3.  End Swift Linux and join the team of another distro: The goals of Swift Linux are to combine the lightweight operation of antiX/Puppy (works well on 10-year-old computers and will fit onto a CD for years to come), the superior hardware/driver support of Puppy/Mint, and the superior repository of antiX/Mint.  If there is another distro underway trying to do this, I will consider ending Swift Linux and joining the team of this other distro.

-- 
Jason Hsu <jhsu802701 at jasonhsu.com>