posted by Jordan Spencer Cunningham on Mon 11th Jan 2010 15:57 UTC
Arch Linux Team, 2/6Obviously there are a lot of Linux distros out there, and it's hard to differentiate some of them from one another. Arch is "lightweight and flexible," but it's not the only one. Do you consider this proliferation of distros a strength or weakness of the Linux movement? Is it an unproductive duplication of effort, or an environment of teeming fecundity from which springs unanticipated innovation, or somewhere in between? Thomas Bächler: It's about diversity and choice. In the Windows and Mac OS world, you have a monoculture. The Linux world comes in many flavors. No matter what the banners say, none of the distros out there is like Arch, which is the reason it still exists. Allan McRae: I do not care what other people do with their time, much like I doubt they care what I do with mine. So I have no issues if someone rolls yet another distro to scratch an itch they have. You cannot tell people what to do with their time. Aaron Griffin: In the spirit of what the others said: this isn't a competition. We don't make Arch so that we can win the most users, or get piles of cash. We make Arch because this is the OS we want to use. Others might not agree and so will use other distros or even make their own. Tobias Kieslich: The development in Linux land is evolutionary. People make efforts to improve things. Some projects succeed while some projects' ideas survive and get incorporated into other projects be those Linux distributions or actual software projects. Yes, it creates overhead and the wheel gets reinvented every once in a while. However, it ensures that the best ideas stay and they usually don't get turned down before they get at least tried. This makes it more interesting and arguably better-tested than development that follows a given agenda. "Lightweight and flexible" can mean more than one thing. For one user it means that it runs on older hardware sufficiently. For me it means hackable. Take packages, recompile them, try stuff utilizing the package manager because it is so easy. Embrace diversity-- I consider it a strength. Ronald van Haren: I guess every distribution has a purpose; some are to have an average desktop system for the average user, others have a more specific purpose. Probably most distributions are started because the ones starting it see a need for it for themselves. It may just happen that some others like it as well. Pierre Schmitz: While we might share some single ideas with other distributions, I don't know any other system with the same design principles and goals as Arch. But go ahead and name one. In general, I don't see the problem of duplicated work even if projects are very similar. In the end there is some kind of evolution among distributions. But there is a good chance that even if a project dies, some of their ideas might be adopted by others. I believe we all benefit from people starting forks or working on crazy ideas. Dan McGee: It is definitely a strength and a weakness. The innovation aspect is exciting to see- without the variety of distros we have today, we wouldn't see Linux running on everything from my lowly wireless router to big academic computing clusters. However, I think you do get some duplication of effort that is unfortunate. Some distos heavily patch their packages without trying to get these changes upstream where they will benefit everyone; I think Arch excels here by not dragging around a lot of patches and getting them upstream if at all possible. Hugo Doria: I think that a lot of diversity can hinder the growth of Linux. On the other hand, limit the number of distributions is not the solution. I think the key is in the balance. I am not against the creation of new distributions if they really have a reason to be born. Roman Kyrylych: It annoys me to see yet another Ubuntu clone on Distrowatch every week, with a different set of the included software and different default wallpaper, while distributions that offer something new are rare. From my point of view, there is a duplication of work in many areas of FOSS, but you cannot command developers to do this and don't do that. Everyone is free to do whatever (s)he wants with his/her time. The teams behind some (mostly early) Linux distros parlayed their projects into lucrative professional services businesses. Do you think that door has closed? Will there ever be another big "Linux company" to emerge? Why/why not/how? Dieter Plaetinck: New ones can still emerge, though it seems hard to compete in the world of enterprise Linux distros. Most of it seems just politics. There are a lot more smaller companies who just offer Linux/opensource services, and that's a more vivid market imho, though all of this doesn't have to do anything with Arch. Allan McRae: My magic 8-ball says "maybe". Aaron Griffin: I think it's completely possible. But the days of "software as goods" are past us. A sucessful business model in the current ecosystem provides a service above and beyond what the software provides. That said, I don't know if selling services for just an OS is feasible anymore. That section of the market is cornered by the existing companies. Tobias Kieslich: That's hard to predict and is probably not a function of new Linux distros coming out. It rather depends on the requirements of the market. See a demand for something that can be addressed by a specialized Linux distro, build it and you have a product. Sure it's likely that these spots will be occupied by incumbent, but I would rule out that a new company rises. Ronald van Haren: I suppose it can happen. It happened once when Canonical stepped in, it can happen again. Who knows. Dan McGee: I don't think you will see too many more big (e.g. Red Hat type) companies emerge. I do see plenty of room for smaller, more specialized companies that develop a business model around a free software base, however. What decisions have you made in your development of philosophy that have gone against conventional wisdom? Have those decisions proven to be important to make Arch a superior distro? llan McRae: Rolling release == good. Simple build system == good. Aaron Griffin: Make simple choices to cover the common case. Let the edge case users do a little extra work to get what they want (for example, rebuilding a package via PKGBUILD to add more complex options). This would be one of the implementations of the KISS principle Pierre Schmitz: We don't provide any tools to configure your system or making an update between major versions of packages smooth. Our users are responsible to update their configuration files and make sure their scripts still work with the latest versions of PHP, for example. This philosophy makes our lives as packagers a lot easier and makes it possible to provide the most recent package versions with a relatively small team. Dan McGee: I didn't make the decision, but when Judd started Arch and decided to write his own package manager and tools on the basis that what was out there sucked, it was a rather unconventional step. The standard was deb and rpm; we are one of very few distros not using one of these two formats. Hugo Doria: Let users tailor the system the way they really want. For real.
Table of contents