I think that's called Quality Assurance.

You should never break something that is already working. Any new functionality shouldn't negatively affect the existing code. New code should work.

The beta testers, I am sure, do a good job. I don't know if they are directed to test certain functions only, or if they have leeway to look for broken code. Do they ever test for problems that existed in the previous release?

PGM doesn't seem to have an effective process for delivering a clean product, which is why we have so many fixes in each build. You have to accept with a product like BIAB/RealBand there will be some untested combinations of features. Much of that must be due to the legacy code. If it wasn't developed or previously modified in a controlled way, new changes will eventually clash with old code.

PGM's approach seems to be discount the new release (e.g. 2021) cost and let the users sort out the bugs. If you don't like this, you can always buy the BIAB at full cost in June when most of the bugs will have been fixed.

The other thing is that most of the errors seem to be related, not surprisingly, to the new code, or to the non-traditional BIAB functionality as a backing track generator. The basic program seems to be fine since it really hasn't changed since day 1.

I still think that PGM should be more forthcoming with their release approach. We should expect a working, essentially bug free product with each release. Today, we often have to do our own regression testing to ensure that product still works for us.

BIAB for Windows Version 2019 (643)
Windows 10 (1909)