Categories

When Distros Go South

While reading my email and catching up with Google+ posts this morning, I happened upon an article by Jim Lynch. I enjoy Jim’s writing and this morning’s offering didn’t disappoint. The opening of the article announced the contents succinctly:

“In today’s open source roundup: Is the systemd controversy a battle of the old versus the young? Plus: Which Linux distros do you consider innovative? And a journalist spends a week using Ubuntu’s Unity desktop.”

Which Linux distros do you consider innovative?

A lot of them actually.

oops, wrong distroI can easily name five distros that have brought some massive changes to the way we interact with our computers…and not all of them on the good side of my list. A more pertinent question might be, which Linux distros are in it for the long haul?

One of the biggest advantages, and an often-perceived disadvantage, is the overwhelming number of choices in the Linuxsphere. I could give a good argument either way if I were pushed into it, but what it comes down to is stability.

I don’t mean system stability, that is a given with Linux, but developer stability. Can I depend on this distro to be around in two years…in five years? How is the project funded? Does this innovative project have a large development community that can step in should the lead developer become ill or takes a lengthy sabbatical?

Yeah, the latter has already happened. The community stepped in and kept the distro alive. That transition was turbulent at times, but the point remains, that distro is still available today.

Reglue deployed two different distros on many of our kid’s machines during a two year period. Both of these “one man show” distros went the way of pay phones and VHS tapes. These were both distros ranking in the top fifteen on DistroWatch. Both of them had a large user base. The surprising thing is that both collapsed with little or no forewarning, leaving many who use Linux to look for a backup solution.

Fortunately, we have other distros that we can respin to meet our needs. But to the point, scrambling to get these systems in the field updated with a supported distribution was a huge tax on resources and workloads. It wasn’t anyone’s “fault.” Life happens; distros die.

I did find amusement in the mention that NixOS should be considered as “innovative.” No, I’m not poking fun at it. NixOS has a lot of new things worthy of using, but my laugh-out-loud moment came when I clicked to read the distro’s “about” page:

Declarative system configuration model

In NixOS, the entire operating system — the kernel, applications, system packages, configuration files, and so on — is built by the Nix package manager from a description in a purely functional build language. The fact that it’s purely functional essentially means that building a new configuration cannot overwrite previous configurations. Most of the other features follow from this.

You configure a NixOS system by writing a specification of the functionality that you want on your machine in /etc/nixos/configuration.nix. For instance, here is a minimal configuration of a machine running an SSH daemon:

{
boot.loader.grub.device = “/dev/sda”;

fileSystems.”/”.device = “/dev/sda1″;

services.sshd.enable = true;
}

After changing /etc/nixos/configuration.nix, you realise the configuration by running this command:

$ nixos-rebuild switch

This command does everything necessary to make the configuration happen, including downloading and compiling OpenSSH, generating the configuration files for the SSH server, and so on.

Now I can’t think of anything more enticing to a new Linux user than the above instructions. That will drive new users to NixOS in droves. They will be racing to argue the merits of emac or vim. They will enter the fray to push forward support for their preferred Python IDE.

OK, so sometimes I take my satire a bit too far…

My sincere wishes are for NixOS to be on the survivor list a year from now. Fact is, I am downloading it now to try out this evening. The biggest problem NixOS faces is the limited audience it attracts. I’ve not found even one new Linux user since 2005 that asked me, “Yeah, but can I compile my own apps?”

Many new Linux users are still struggling with the concept of the pager in the bottom panel, mirroring the number of virtual desktops they can use. They still think that the live DVD or flash drive is black magic. You can “innovate” all you want, just don’t drop the velvet rope when new users try to get in.

Innovation without a focus on attracting new users is a project just waiting for the chance to die.

The following two tabs change content below.
Ken Starks writes and publishes The Blog of Helios, a finalist in our Best FOSS or Linux Blog competition. In addition, he's the person behind the Reglue project that refurbishes older computers and gives them to disadvantaged school kids in the Austin, Texas area.

6 comments to When Distros Go South

  • I can see your point that users totally new to Linux probably will prefer a slick UI and not mess around with the internals.

    Still, a distro like Gentoo, where everything is compiled from source by the user, is alive and well. It may be a niche distro, but if you visit forums.gentoo.org, you’ll find that there’s a lot of old-timers like me who have been using Gentoo for 10+ years, as well as a constant trickle of new users.

    I take it that Nixos as well is not a distro that primarily want to appeal to the regular Joe user. Rather, it’s intended for people who don’t mind getting their hands dirty with fdisk and compiling their own kernels.

    Luckily, the Linux world has space for both noobs and hard-core users. There’s no need to make fun of either of them.

  • Duncan

    One great thing about Linux distros is the relatively low threshold to creating your own. It’s a very small step from “Hey, with Linux I can configure it my way” to “Well, now that I have it working the way I want, I might as well put it out there for anyone else interested in it working that way too.”

    This is why my answer to the “too many distros or not?” question is always “It’s just about right.” Because with Linux it IS possible to “do it your way”, and once it’s “done your way”, there’s virtually zero bump to simply putting it out there for anyone else interested (and with a bit of luck to get some help with it as a result), literally, anyone who wants it done differently and has the time and a bit of skill to do it CAN do it, so the level of supply and demand by definition are a near perfect match at a near-zero (beyond just doing it for one’s self, which might or might not be trivial depending on how much you want to change from a near-match distro) cost.

    So there can never be too many or too few distros, because literally anyone that wants to change a few config files from their base distro can do it at basically zero additional cost, and they must have a reason for doing it or they’d not bother.

    OTOH, because that means it’s trivial creating a distro and a lot of folks will do it, it’s equally trivial for such a one-off distro with just one or two people actually doing it to die, and they often do. But that’s OK too, altho it can mean a bit of temporary pain for users of that distro until they get properly switched to another. Tho of course if there’s enough users with high enough skill level, they can simply carry on with it themselves.

    Duncan

  • So, Ken, apparently innovation doesn’t count unless it’s making the UI friendlier for John Q. Newbie?

  • Onan the Barbarian

    Ken,
    I guess you’ve evaluated mainstream distros that don’t risk dying, e.g., Ubuntu, Fedora, Debian…
    Can you elaborate on why they aren’t suited to your needs?

  • “So, Ken, apparently innovation doesn’t count unless it’s making the UI friendlier for John Q. Newbie?”

    As short-sighted as you maybe believe me to be, in my world, the answer is yes. But I said, “in my world”. We install anywhere from 90 to 250 Linux-powered computers a year to school kids. Finger on the pulse and all of that. I did not really give the Gentoo and Slackware distros enough consideration as they do have a solid user base. But in comparison to the user base in the “user friendly” distros, I would hazard a guess that users of friendly distros greatly outnumber the others. Your point is taken…I did not clarify my position as well as I should but I still stand my ground on the risks of small development distros having a much shorter lifespan, regardless of how user friendly they are. And that after all, was the point I am making.