IMHO, a bug tracker is essential in any project larger than a few thousands lines of code (so, every project!), and with a lifetime longer than a few months. When these two measures are exceeded, it is virtually impossible for standard human beings to keep track of issues to be resolved...
And of course I also agree that the bug tracker itself is not enough: there must be a "QA team" that triages the bugs, makes sure they are valid, prioritizes them... in short, makes sure that the developers (the ones that can actually fix the bugs, and which are always a scarce resource) don't waste time navigating in the bug database, and focus on fixing them.
But there's something that nobody is pointing out: regression tests. In my experience, they are the key factor that helps in being sure that as the project grows, the number of bugs stays under control and does not grow exponentially. And having a build farm that continuously checks out the latest source, builds it and tests it on all the supported platforms, and reports the results in the most accessible way.
And, possibly, having the policy of refusing contributions that do not include automated regression tests. I know this sounds a bit draconian, but it is the only way to be sure that the tests have good coverage...
Copyright © 2017, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds