we need a strategy for the future
You're right! I'd like that strategy not to repeat the mistakes of the past.
For example, I think we should collectively acknowledge that deliberate fragmentation, in the hope that we'll have eventual convergence, has rarely worked for us. Experimenting with multiple competing object systems on the CPAN and hoping eventually something will get into the core has left us with a core that's (apart from mro and parent) unchanged since 5.004 or so for objects with any sizable CPAN dependency chain pulling in multiple competing object systems.