I’m still on the fence here. This seems like a positive development as the product was clearly costing Bandlab money to develop and stabilize but not generating any cash in return, and probably not driving a lot of sales for other Bandlab services. This could not go on. IMO, things will be fine if a couple of conditions are met.
[1] An existing Cwb installation should continue to be viable, even if it requires re-installation because of a crash, i.e., the last version should be available as a stand-alone .exe, and no longer need to "phone home" or be authorized by a server.
[2] “New” Sonar should be available on a pay-per-version or per-year model that also remains viable in its present state if a user chooses not to continue to upgrade. (Yes I was one of the fools who trusted Gibson with 100s of dollars for “lifetime updates,” which is why I have some cynicism here. )
If , on the other hand, they go to a model where a subscription is required to keep existing software running, that would be the breaking point for me.