Google moves WebRTC code samples to GitHub

What are the benefits of being on Subversion and Google Code at this point for a project like this? Or is it just that moving over is hard?

We updated the submitted title, which referred to source code instead of code samples.

My guess: it's sorta in the same fate as Reader is right now, just not being killed as quickly. There aren't many people maintaining it. And they need to upgrade to a newer backend at some point, but no one is around to rewrite it.

It's also important to note that their support for SVN, HG, and Git is all an in-house implementation to work on top of BigTable.

I'm not really sure why they let it stagnate in the first place, but someone made the call that it wasn't worth investing in anymore. Remember that Google Code really came as an answer to SourceForge which was going downhill. But then we had GitHub and BitBucket appear, which have really taken over.

For pure source repos, Google has been using Gitiles[0] for most of their public projects like Android and Chromium[1].

[0] https://code.google.com/p/gitiles/

[1] https://chromium.googlesource.com/chromium/src.git

This project was previously at googlechrome.github.io/webrtc

It may have also existed at Google Code, I don't know. But this was the repo I was directed to after Google I/O last year.

"Seems like it wouldn't require many more resources"

Actually maintaining is costly. 1. The system around it is evolving, even no features are added, necessary upgrade is necessary to talk to the backend, and sometime it is nontrivial. And since svn is now not popular, should they add git support to the dated infrastructure? 2. Routine SRE works should be done, deployment to new clusters, oncall alerts. 3. Bug fixes. 4. How can the maintainers be rewarded? They are working on a project in maintain-only state, who wants to do that?

Basically, for hopeless projects, better to turn them down than make them alive in half-life.