Hello, I had brought this issue up in the past, but I believe is neglected in the middle of the storm. I (deadly) request from the board to migrate the project from Subversion to GitHub. Below I explained my personal take on the issue:
*Using Git:* - Developers will not need an internet connection to work on the code base. (My patch that increases Java client's performance by >2x was written on the train.) - Developers will be able to keep separate branches for each feature that they are currently working on. - Instead of a single huge monolithic commit, a feature will be pushed in small, semantically meaningful commits via branch merges. - While working on multiple features sitting on top of each other (e.g., [feature2:] writing a console client, [feature1:] using a JSON parser previously fixed), developer will not need to wait for the owner of the project to apply changes to the master. *Using GitHub:* - GitHub migration will attract and encourage a lot more people from F/OSS community to contribute. (You can just login and click "edit" on a file. Then project gets cloned for you and your changes are sent back as a pull request to the origin.) - We could have easily implemented a CI via Travis+GitHub integration in a matter of minutes. (We have been talking about it for almost 2 weeks.) - Similarly, we could have integrated code coverage <https://coveralls.io/> in a couple of hours. - GitHub provides a really developer-friendly and efficient push-review-modify cycle that is accessible through a web interface. (Right now, we need to attach patches to JIRA issues and wait for somebody to download them, apply them, run them, and comment on the patch.) - Using GitHub pages, we will be able to keep and track the static files (JBake/Jekyll powered content, files under http://devicemap-vm.apache.org/data/latest/ used by DM clients, etc.) in a VCS environment. Best.
