That's a pretty awesome SVG; the titles are links to distro sites, it's attractive and readable, and SVG makes it easy to Ctrl+F. I am happy to the author of that file. Thanks for sharing it here.
There’s probably no way to easily gather the data, but it’d be very interesting to see this same chart with the number of active contributors (or some proxy like number of committers/patch authors within the previous month) to each project represented by the width of the line. It would give a better idea of which branches are serious and invested in and which are one-man hobby projects, which ones are growing fast and which are dying, etc.
It's further complicated because of upstream development (everyone benefits from Red Hat's kernel updates, CentOS benefits from pretty much everything Red Hat does), as well as dowstream bug collection and triage (Debian benefits from Ubuntu's installed base).
Another interesting parameter to look at would be number of users. Comparing number of active developers (for each project specifically) with number of users would give a pretty good idea of which projects were benefitting from upstream development and which were serving as backbones for downstream user-oriented projects.
It's an impressive effort, although it's missing the first Linux I regularly used: Tom Oehser's tomsrtbt (http://www.toms.net/rb/), a single-floppy release aimed at system rescue.
Highlighting the lineage aspect, however, loses useful context information. For example, Robert Shingledecker was a significant part of the development of Damn Small Linux for five years before he started the Tiny Core Linux project. While there's no commonality in the distribution-specific code that I'm aware of, and hence no reason for a graphic connection on the chart, I think it's safe to say that Tiny Core's design is a continuation of Shingledecker's thinking on what makes a good minimal Linux.
It's interesting to see that most distros are descended from either Debian, Slackware or Red Hat. I guess that trying to improve what already exists is less work than starting one from scratch.
Wikipedia page on SLS states: "Similarly Ian Murdock's frustration with SLS led him to create the Debian project".
Slackware originated from SLS. So strictly speaking SLS and Red Hat appear to be the two originating distro's from others have decended in one way or another.
Brought back some good memories..of visiting distrowatch.. compiling custom kernels..swearing at KDE/GNome for being too slow and using fvwm..moving from xine to mplayer..
I used linux for 5 years and used redhat in 2001, gentoo for a brief time, slackware for 2 years, fedora for a year, back to windows for 1 year and now I'm a happy OS X user for 3+ years.
Not sure if I'm going to get downvoted for my dislike of the fragmentation, but I've been always wondering what if all the efforts could be somehow combined to make a really great one...
There are many, sometimes very different measures of what "a great one" does and would mean for an OS, as I'm sure you know. And therein lies the beauty of having rich diversity to choose from.
I can sympathize with the irritation of seeing duplication of effort. But when that irritation pushes a mind toward the tempting presupposition of some single measure of goodness which could have been a guiding light the whole time, paring away effort duplication and dead-ends... well, the thinking has gone a bit absurd.
There are many regions of possible-OS space, not everyone needs or wants to be in the same one.
