I think one of the (somewhat valid) criticisms of the Linux community is that people spend time creating new distros rather than putting effort into improving what's already available. It's the same what I think every time a new distribution pops up on Distrowatch, describing itself as a "Ubuntu-based distribution aimed for [insert-random-language-here] speakers". Switching languages on Ubuntu is simple enough, do we really need another distribution that offers no more than a language package and a custom theme?

Distributions that are fundamentally different from one another are rare, and usually these fundamental differences reside in the package management system and the readily available software. Wouldn't we be well off also if we had only just enough core distributions (not more than a handful) to cover all the features found in Linux systems today? Where would Linux be if all the effort in creating 600+ distributions were instead put into reporting and fixing bugs as well as adding features in Debian, Slackware, Ubuntu, Arch, the Linux Kernel, etc.. After all, these core distributions are still flexible enough for anybody to tailor to their needs.

-A

+100000

@ahmed

Creating a new sub-distro (depending on complexity) can generate new bug reports if the group creating the sub distro dives deep enough into the inner workings of the main distro. But I would agree, sub-distros seem to be nothing more than a software pack. That's why I wouldn't claim to be a distro; I would be a software pack and a set of reasonable defaults for a specific purpose.

However, there are some cases where I can say a sub distro may be a good idea. SalixOS is one of them. They rip out KDE/QT and turn it into a XFCE/Gnome Lib based small install Slackware. It actually works as a good example of a sub-distro... if you want a gnome lib small isntall Slackware

Gentoo by default does not compile everything anymore, it will install pre-compiled binaries from a large package. However you can make it compile everything and I even hear you can do the same thing on Arch tho I have never dived deep into arch enough to know how to do it.

My situation is a bit wierd. For my netbook I am using a LFS build with some slight modifications, for my NAS box I use Arch and for my gaming computer I use windows 7/ gentoo.

For the most part I have a distro on my nas because I am lazy and don't feel like customizing my LFS to work on it, even tho it probably could since they both use Atom CPU's, and because Arch works so well and is easy to update.

I see your point, but since anyone who knows programming may participate in Linux development, aren't there theoretically enough people to do both--creating "new" distros and improving the existing ones?

Hmm, I just noticed this new "reputation" thing...I'm off to read about it.

Honestly I wouldn't. I find this to be one of the very few of Linux's problems. Over-saturation and no standardization of distributions. Way too many distributions that anyone can ever want and they all do things differently which cause a swell of frustration and poor documentation by developers. There's a reason why the top 10 never move much.

Well, I accept how things are by now. I would like Linux to become popular enough to dislodge Windows's dominance, but not enough important people in the Linux community are seriously interested in that. Most Linux users dislike Windows and Microsoft, but it seems that they would prefer to do their own thing rather than crusade against Windows and Microsoft. There is limited sympathy for how Windows users probably react to the proliferation of Linux distros--"Hundreds of these distro things? How the hell do I know which Linux I want?"

Obviously there are enough people available to do both right now, but if the effort was put into improving language, hardware and multimedia support of mainstream distributions, creating new and redundant ones wouldn't be necessary. These main distributions would offer enough to take up a much larger market segment in computing.

Why have more than one lightweight distribution if (say) Damn Small Linux offered everything you could wish for from a lightweight distro? Same for Firewall distributions, live distros, server/enterprise distros, "beginner distros" and all-rounders like Fedora, Debian and Slackware.

Distrowatch lists 665 distributions in its database, 296 of which are discontinued, and 230 are on the waiting list. That's kind of a mess if you ask me.

-A

I must agree, but since people have the freedom to do whatever they want with the code, it's inevitable that distros will proliferate endlessly, pleasing the ideological advocates of software freedom but frustrating the people who are interested in Linux but "just want an operating system right now, damn it!" (The peanut butter analogy keeps coming to mind.) I sympathize with both sides. The well-meaning programmers who modify popular distros slightly and then share the result as a "new" distro must suppose that since they like what they've made, someone else might. What can "newbies" do but sigh and explore distros?

I think what happens is that by dumb luck or by someone's recommendation, they hear of a certain one, and think "Fine, I'll try that one." The first distro I ever tried, seven years ago, was Mandrake; I had just read about it in some book or on some website.

It reminds me of the book 'After the Software Wars' where Keith Curtis views his opinion on (among others) AI research camps across the world. With each one for various reasons developing stuff independently, each organisation actually reinvents the wheel. It would be so much more productive if they joined the efforts and worked together. After all, they have got more or less the same goal.
Linux distros developers do make their findings available for others (source code), but, however you look at it, creating more and more distros doesn't seem to help the state of linux. Probably 10 or so distros would suffice to satisfy everyone's needs. The rest of them don't seem to bring much to the table (a different theme? different selection of packages installed by default? IMO, that's not a good enough reason to create a separate distro.)
But then again, it's a developer's own business how they utilise their skills and time. If they want to develop the Nth distro based on Ubuntu, which, in turn, is the Nth distro based on Debian, that's fine with me:)

A unique distro doesn't 'compete' with all of the other distros out there. The idea that any distro competes with another seems silly since they don't all target the same groups of people.

I kept seeing mentions of package management and hardware support while one key aspect seems to be missing here: philosophy. Ubuntu tries to make a user-friendly distro that anyone can use. Honestly I can't say much about Debian or Slackware simply due to ignorance, and Gentoo doesn't seem to have enough focus within its dev community to really give any real indication of any direction but at some point each has had its own philosophy. Had it not been for issues at the start we may only have ended up with Debian or Slackware by now and the other would be nothing more than a page in a history book, if it ever existed at all.

The ideas that drive one distro can mean more than anything else (see Ubuntu). Package managers, default installs, and general configuration is simply an extention of the originating ideals. Some ideas, like user-friendlyness, are so subjective and vague that a thousand people could try to implement it and someone would still disagree.

Some issues may arise from a software standpoint (to patch or not to patch, how to handle packages, what optimizations to use, ect) which can have lasting impacts on the end users experiance (distro-specific patch implements a bug, package manager feels kludgy, generic (slow) optimizations) which may drive them to reinvent what they feel is lacking (apt-get v. aptitude, portage v. palidus).

To be quite honest the massive mess of distros can be a Good Thing if everyone is willing to share ideas and work together. Sometimes one concept may not work well somewhere else, or one idea may catch on and improve things for everyone (/usr/local-based package manager and static bins. I can dream, can't I?).