I guess I might be wrong. However, consider: Regarding XBMC / Plex and Emacs / xemacs, instead of a fork it looks like it's more of adding feature X that the o

I guess I might be wrong. However, consider: Regarding XBMC / Plex and Emacs / xemacs, instead of a fork it looks like it's more of adding feature X that the original devs didn't want (not being derogatory here, it's substantial work). I'm sure that they merge in new code continuously and the codebases haven't diverged so much as to be independent. It doesn't look much like a fork in that sense. I'd argue that KHTML is fairly dead. No one really uses Konqueror (I worked for a shop that developed a KDE based distro and even there everyone used Firefox / Chrome), which is probably KHTML's largest user apart from some internal KDE stuff. Hell, even Konqueror has had webkit mode since about a year ago. Qt supports webkit. Tons of other browsers, mobile and otherwise, use webkit. I see the same downward trend for Debian, from being all over the place to being reduced to mostly a server distro. Meanwhile, Ubuntu is coming up with its own server version with pre-made virtualization images and stuff. So, maybe the "weaker" forks just didn't die fast? Of course, this whole idea isn't a law or anything and I can't expect it to apply to everything, but that's the general trend I see anyway. Also, I guess another thing I could claim is selection bias: You've probably never even heard of most of the forks that died. Of course, I don't have the data to prove it either.