What we really need is for people to propose alternatives to systemD that address its purported use cases while removing some of the needless complexity it introduces, and restoring the primacy of general mechanisms over hacked-together policy. This has happened with some components, e.g. elogind but not really as a comprehensive effort.