TFS for Java - bad idea?

Question

We\'re considering TFS for our .NET based projects and as a task management platform. Some teams develop exclusively in Java and they\'re quite happy with SVN (Subclipse).

Answer 1

After a year of working on a Java/JVM project using TFS, I would like to dissuade anyone from doing this. While TFS may be considered top-of-the-line for .NET developers, you won't find any Java Developers with any experience with it. There is the plug-in for Eclipse and a port to IntelliJ, but I've had terrible luck with both, though I'm guessing it's mainly because TFS does not work like any other VCS I've used.

On our team, we've estimated a 10-15% overhead due to TFS and complications caused by it. Days of work lost because TFS decided to overwrite files, days of troubleshooting issues caused by incomplete TFS Updates. We have done a branch in 6 months because the entire team lost two days the last time we did. It's common to hear the phrase "I just updated with your latest changes, can you come check to make sure nothing disappeared in the merge?". Instead of using Jira, we're stuck using the terrible issue-tracking in TFS, causing more yet more issues.

Several of the developers on the team have taken to either using git, either standalone or the git-tfs bridge. Others just copy the source tree prior to any 'risky' activities, like updating or checking in.

Either way, I wouldn't recommend it for a team that does not have experience with it...