Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

Slight tangent: Here's a secret about forking. When you fork a project, you anger the loyalists. A lot of large open source projects fork because of deep disagreements / polarization within the team. For minor stuff, it usually just gets settled internally. If the project does get forked, it usually ends with either the original project dying or the fork dying. I can't recall any projects off the top of my head that were forked and both the fork and original remained successful.


I think the trend is shifting at least when it comes to projects hosted on Github. Forking there is less of a divergent path, but more of a way to tinker and make changes and push them back to the main code source. At least that is what most people do, and I think it is better for it.


Emacs / xemacs.

NetBSD / OpenBSD.

Debian / Ubuntu.

XMBC / Plex.

KHTML / Webkit.

sbcl / cmucl.


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.


You know you're on Hacker News when a discussion about Tolkien turns into a discussion of Emacs forks =)


There are two activities that can change a nerdy fourteen-year old's life: reading The Lord of the Rings and learning emacs. One is a childish fantasy that often engenders a lifelong obsession with its unbelievable inaccessibility, leading to an emotionally stunted, socially crippled adulthood, unable to deal with the real world. The other, of course, involves orcs.

;)


That's why John Galt uses Vim :)


Is Ubuntu a fork of Debian or a polished package based on Debian? I ask because the Ubuntu update schedule is based around the Debian update schedule in such a way that new Ubuntu releases remain synchronized with new Debian releases.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: