Here's a counterargument: if a tool starts developing userbase, it's likely to have even more users in the future. Up to a point, the more users it has, the more it will have. So a breaking change while the userbase is still relatively small can be justified on the grounds of making things better for a larger number of soon-to-be users.
I agree. People complain that Swift changed too much in the earlier versions but I'm glad it did and I'm kind of sad that it's slowed down now and there are things I wish could be changed but can't. Yes it was a bit of a pain but there were migration tools that helped. And that pain is long since forgotten.
