Value of supporting RBAC and Webhook generation?

Looking at https://github.com/kubernetes-sigs/controller-tools/pull/1339, I am becoming a bit sceptical at at the value of our support for generating RBAC and webhook configurations from go code comments. Using go code comments as a source makes sense to me in the context of DeepCopies and CRDs, as in those cases the comments applies to go types and their fields, but that is not the case for either rbac or webhooks.

Supporting rbac and webhook generation from go code comments means that we have to invent a DSL that is valid as a code comment and supports any possible feature of the two. I don't fully understand the value of that over just directly writing or in the case of kubebuilder generating the RBAC/Webhook manifests.

Thoughts @sbueringer @JoelSpeed @camilamacedo86 ?

Start by reading linked pull request #1339 and the discussion on this issue about comment-based RBAC and webhook generation. Compare the value and scope of that approach with directly writing or generating manifests, then clarify whether support should be retained, changed, or removed; done means the project has a settled direction.

Written by the indexing model from the issue text.