Last night I received access to the Google Wave Developer Sandbox. My original request for access was made when I had a lot more free time than I do now (I’m measuring free time on a negative scale now: I don’t expect to have any until May). My plan at the time was to integrate Google Wave with bugtracking software (written from scratch). I haven’t actually looked at the wave APIs yet, but I had considered all three methods of extending Wave for use with a bug tracker:
- Embedding: An embedded wave on an external bug tracking webapp would make discussion much simpler and more robust. Most bug trackers don’t even have threaded comments right now. The ability of a wave to allow synchronous or asynchronous discussion would make the discussion system much more useful (most discussion on bugs currently happens “out of band” on mailing lists or in chats). In addition, a “subscription” feature could be as simple as adding yourself to the wave. Finally, the ability to edit wave postings (wiki style) would allow the data in the bug report to be easily organized and classified. People wouldn’t have to read entire lists of comments to get an overview.
- Robots: At the simplest form, adding a robot to an existing wave would create a new bug report based on that wave. It would be nice if you could also have the robot automatically set bug attributes such as project, version, milestone, architecture, etc (whatever the project demands).
- Extensions: An extension to wave could allow bugs to be integrated into other waves, instead of just waves into bugs. Instead of being stuck on a fixed bug report server somewhere, the bugs could be sent to other wave users. That could even include upstream authors. This is the most interesting prospect, because it could lead to a “global bug tracking service” where bugs are not attached to a single reporting tool, but are easily routed to the people most able to fix them.
In Arch Linux, many, but not all, of our bugs are related to specific packages. It would be nice if bug reports could be attached to packages in our web interface. Then it would be easy for users to report bugs on specific packages, and for developers to find all bugs related to their packages.
Our team includes a few dedicated (and extremely under-appreciated) users who classify incoming bugs and assign them to the developers who ought to be responsible for fixing them. I imagine a wave scenario where this can be done as easily as forwarding all new bugs to these “bug wranglers” wave accounts, and allowing them to add other developers to the wave and have them automatically assigned.
Ideally, we would host the wave server ourselves, so as to be in complete control of our information. Of course, if someone else wants to add themselves to a wave from a different service, that should not pose a problem.
Unfortunately, I suspect that when I have time to complete this project, I’ll have been swept up in some new and exciting venture. But it’s certainly an interesting tool to think about.