I can say, for the moment, other than having to kill Sonar after every use via Task Manager, it's been pretty solid. It's not that Sonar is unstable or full of bugs, but that there are probably limitations and mishaps with API's offered by Microsoft and other software integrated into Windows. With that said, still, there are probably some elements of testing left out due to time or budget constraint, and some ways of finding bugs that could stand improvement.
The issue I still have is that reporting potential bugs needs work. If a user experiences an issue, merely reporting it in a forum or dialogue, even at the basic level should be enough. The reason I say this is that, as these things accumulate, it may produce trends. For example, ANY issue reported for graphics or the GUI should be accumulated as a GUI/video issue. It shouldn't matter (yet) exactly what the issue is, but that there are "x" number of video/GUI issues. If this seems to be a problem area,
then more detail should be requested from the user by Cakewalk to see if there's similarities and potential bugs. That's more proactive.
I still don't think the Bug Reporter tool is effective, because 1) there's no incentive to report bugs, and 2) it's not always possible to reproduce with specific steps. I came from an industry where my job was reporting bugs. A bug may be found in a result, not always a specific procedure. I've proven that theory many times with Adtran engineers. Intermittent bugs are hard to find, and the tools offered by CW don't fully encompass that variable, IMO.
Yes, there's multiple ways to report issues, but all of them have a challenge that prevents efficiency to some degree.
Otherwise, I work just fine with Sonar, although I still have some major gripes and issues. If X2a were the end, I could still use it for years to come. With the coming of X3, I think there was no better way with the current Cakewalk strategy to fix issues and implement new features. I just wish the two were disparate and not a package deal.