Ubuntu is based on Debian. Mint is based on Ubuntu (except for Linux Mint Debian Edition, which is based on Debian.)

When one distro is based on another, in what ways is it different from the parent distro? What has been added or taken out or changed?

EDIT: After reading the answer from Chris Down I wonder if I should have approached this from a different angle, asking, What are the components of a distro other than the kernel and and applications?


The degree to which something is a derivative of another is contextual to the project itself.

As for your specific examples:

  • Mint (aside from the Debian edition) relies heavily on the Ubuntu infrastructure. It not only is based on Ubuntu (in that Mint is essentially Ubuntu with in-house modifications), but it also relies on Ubuntu's package repositories, too. The main difference with Mint is that they have a set of tools designed specifically for Linux Mint. Also, the philosophy of the Mint project is compatible with proprietary software (they put percieved user-friendliness above libre software), whereas Ubuntu discourages it. As an example, Mint comes with Adobe Flash preinstalled, whereas Ubuntu does not -- this extends to many other components.
  • Ubuntu is based on Debian in that not only did it start off building from Debian's codebase, it also utilises a number of tools that were originally designed for Debian (aptitude is a notable example). The ethos too is radically different, Debian's core ethos emphasises stability and predictability, whereas Ubuntu's poses ease-of-use.
  • Can you flesh this out some more? What's the 'Ubuntu infrastructure'? What is 'Debian's codebase'? What do they consist of? What are the pieces of a distro, other than the kernel and the applications? – user11583 Sep 17 '11 at 19:01
  • 1
    I guess the four primary parts of any distribution are its ethos, its package manager, its default environment, and its init sequence. Those are the main things that people mean when they talk about a distribution's infrastructure. Ubuntu, for example, took Debian in a completely new direction (an ethos of usability rather than stability) whilst still using a vast number of its tools (aptitude, dpkg, etc). In this case "based on" mostly means "forked from". – Chris Down Sep 17 '11 at 19:08
  • @Chris Down: I disagree 'forked' is not the correct term. For a perfect example, most packages start from Debian upstream. A fork by definition diverges after its inception, not to take away anything from the many great Ubuntu innovations. I just feel like Debian still carries some weight at the foundation, and the Ubuntu org feels the same way. – J. M. Becker Apr 20 '12 at 4:46

Another line of development: Mandrake (today Mandriva, after merging with the Mandrake offshot Conectiva) was derived from Red Hat, mostly diverged by emphasizing KDE instead of Gnome as a default desktop environment. The Fedora project was born as a collection of packages for Red Hat, and later took over the Fedora distribution when Red Hat Enterprise Linux became a separate distribution. There are several minor distributions that are Fedora-derived. Then there are distributions like CentOS that take the sources for RHEL and recompile them (replacing the Red Hat logos and other branding), and distribute the results without the need for a maintenance contract) and Scientific Linux, who do essentially the same as CentOS but add some packages for use at Fermilab, CERN and others. Then there is Oracle Unbreakable Linux [sic], also a Red Hat clone.

What are the differences? As this is open source, if somebody isn't satisfied with what their favorite distribution is doing, they can fork it and go their own way. Setting up the whole infrastructure (webpage, download sites, build farms) isn't cheap/easy (even though most (if not all) of the e.g. Fedora infrastructure sources are available free, and I suppose Debian's too, this is just a small piece of the job). But much more importantly, getting a bunch of dedicated volunteers to package software, keep up with upstream, fix bugs, answer questions, and so on is a huge undertaking. It is certainly much easier to just handle the part that bothers you most (i.e., set up a collection of alternate/supplemental packages for some distribution) and leave the rest to the parent distribution, or even select packages from said parent without much other added value as the base of what you are trying to do. The differences will depend heavily on what the derivative is trying to accomplish, the resources it has available, and the time elapsed. On the other hand, all distributions are built on the same basic infrastructure of open source software (the Linux kernel (even with Debian's BSD-kernel based line and such), GCC, glibc, X.org, Gnome, KDE, ...), so there is an overarching commonality there.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.