We tried out 1.x because we found ourselves publishing non-breaking feature changes in patch releases.
That said:
1.0 indicates stability and a lot of our APIs change a lot. We're already going to have to do a deno_ast 2.0 release.
People become more concerned about breaking changes and having to bump a major. There's something psychological to it. Breaking changes for our development workflow is better in many cases.
Probably publishing non-breaking feature changes in patch releases with 0.x is ok. We should just do that.
We tried out 1.x because we found ourselves publishing non-breaking feature changes in patch releases.
That said: