I do wonder what GitHub is going to miss, like Google Code before them and SourceForge before them.
Although as I said to Jan, if I knew, I'd be doing it :-)
Hindsight is definitely 20/20.
Google Code finally introduced Mercurial support and I wonder, if they'd done it sooner, (a) would Mercurial have beaten Git? (b) would Google Code have beaten GitHub?
Or at the very least, would Google Code not be shutting down?
After I kept hearing about GitHub, I made an off-hand remark in one of my talks that "while 100% of the apps in Pinax are on Google Code, it wouldn't surprise me if most are on GitHub within six months".
And sure enough, 100% of them moved to GitHub.
But even in that first year of Pinax, it became clear the open source developers I was collaborating with preferred Git over SVN.
When I started Pinax, all the component apps were on Google Code. 100% of them. They dominated the Django app space.
Google Code came at just the right time when I needed good SVN hosting.
And they had other innovations too. I really liked their tag-based approach to issues. It enabled tremendous flexibility without the configuration nightmare.
The main reason I left SourceForge was I became a huge fan of SVN over CVS and SourceForge was just too late to implement SVN support.
As Baron points out, SourceForge was so much better than self-hosting CVS, which is what I did in the 90s. (That or RCS!)
I was a big user of SourceForce around 15 years ago. In fact, I was an earlier enough user, I was offered a chance to participate in the VA Linux IPO. Boy did I miss out on that one!
People dissing Google Code probably didn't use SourceForge. People who hated SourceForge probably never self-hosted projects. #gratitude
— Baron Schwartz (@xaprb) March 13, 2015