Disclaimer: I'm a GitHub guy but also a big Rubyforge user (3 years and a handful of gems there).
That said, the technical issues raised in this article could be addressed in other ways: RubyGems could be patched to search for dependencies in all of the gem sources and Rubyforge could add a way to remove gems from the index (no more outdated 'mash').
But why? Why not just keep Rubyforge central? It's worked well so far.
A few reasons.
Foremost is this idea of an "official" RubyGem. What does that word mean? It means that if I disappear, so does my project. If I lose interest, my project dies. With larger projects this doesn't happen so often. It would be hard for Rails or MooTools to just vanish. But for smaller projects, which are legion, this is a constant. People lose interest - often right when you become interested. Let's call it Defunkt's Law.
Defeating Defunkt's Law is one of the primary motivations behind GitHub. Let the Internet decide the 'official' version of smaller projects. If you Google 'mash' and my version comes up before mbleigh's, guess what? Mine is the official version. We have all these tools for deciding what's good content on the web, why not apply them to code?
sr's git-wiki is a great example of this. Google "git-wiki" and al3x's version comes up - a fork with distinct features. The market has spoken.
It's true that within GitHub, finding the "blessed" or "official" or valuable fork is not the easiest thing in the world. But we're working towards this goal - which project in this tree is the most interesting? Which is clearly a feature fork? Which are just impulse forks?
While entropy can seem scary, we believe the future of development is distributed.
RubyForge was one of the reasons I got really interested in ruby/rails development. A central location is a must for new ideas to be easily moved around.
What about a means of pulling updates from the GitHub gem server over to RubyForge? Or a hook in GitHub that ties gem publishing into RubyForge? Seems there are a number of ways to solve this problem with just a little collaboration between RubyForge and GitHub, without sacrificing any of the centralization of RubyForge or the distributed coding goodness of GitHub :)
That said, the technical issues raised in this article could be addressed in other ways: RubyGems could be patched to search for dependencies in all of the gem sources and Rubyforge could add a way to remove gems from the index (no more outdated 'mash').
But why? Why not just keep Rubyforge central? It's worked well so far.
A few reasons.
Foremost is this idea of an "official" RubyGem. What does that word mean? It means that if I disappear, so does my project. If I lose interest, my project dies. With larger projects this doesn't happen so often. It would be hard for Rails or MooTools to just vanish. But for smaller projects, which are legion, this is a constant. People lose interest - often right when you become interested. Let's call it Defunkt's Law.
Defeating Defunkt's Law is one of the primary motivations behind GitHub. Let the Internet decide the 'official' version of smaller projects. If you Google 'mash' and my version comes up before mbleigh's, guess what? Mine is the official version. We have all these tools for deciding what's good content on the web, why not apply them to code?
sr's git-wiki is a great example of this. Google "git-wiki" and al3x's version comes up - a fork with distinct features. The market has spoken.
It's true that within GitHub, finding the "blessed" or "official" or valuable fork is not the easiest thing in the world. But we're working towards this goal - which project in this tree is the most interesting? Which is clearly a feature fork? Which are just impulse forks?
While entropy can seem scary, we believe the future of development is distributed.