Chorus and Maestro

Distributed Synchronization

Matt Jones & Jacob Wiseman
Faculty Advisors: Benjamin Pierce & Fernando Pereira
Home | Chorus | Maestro | Challenges | Harmony

Challenges and Final Thoughts:

From the start, our main problem was not understanding the whole scope of the problem we set out to solve. We spent the first semester writing the filesystem for Maestro, while that time could have been better spent.

In the end, the choice to write Maestro from scratch around Harmony was not the best decision. Existing systems, such as Subversion, have much more functionality that Maestro. It would have been better to take a system like Subversion and replace its existing synchronization engine with Harmony. We believe we have demonstrated what funtionality makes it different from existing systems. Its not that we didn't finish it, we just met our original goals. Instead of spending time adding the 'extras' that are working quite well in other systems, we decided to focus more on Chorus because we believe that it represents something both new and useful.

Matt Jones & Jacob Wiseman , 2006