Two quick thoughts... 1. By not automatically hoisting that information there is a requirement for the developer to know those details for all transitive depen

Two quick thoughts... 1. By not automatically hoisting that information there is a requirement for the developer to know those details for all transitive dependencies and handle those issues themselves. A task that was automated under dep (and in other languages) is now a manual task in vgo. I do not look forward to doing that when I import Kubernetes into a project as a dependency. 2. Writing a solver that can work over a rather large codebase (e.g., kubernetes) in a timely manner (as fast as those for other languages) can and has been done. Why are we trying to avoid satisfaction problems, whom do they benefit, and why?