Yaron Schneider

> What I'm trying to consider is a world where pluggable components definitions are reusable across components I would say this is out of scope, due to: 1. It isn't...

> Oh wait, just to make something clear, when I say "pluggable components definitions are reusable across components" it literally means this ["Reference fields in CRDs referencing other Kubernetes objects...

The only thing I'm favoring less is the `ComponentRegistry` design. I can get behind a `PluggableComponent` CRD with runtime validation of duplicate names. However, I'm proposing an alternative design that,...

> I don't have any strong opinion other than making things clear and easy to use -- in that regard I think that using only the `Component` CRD with some...

Since users don't know about or interact with our registry in any way since this is an implementation details at the lowest level, we should keep this out of scope...

Can we make pluggable components consistent with how we discover the binding direction today?

> In addition to that: I'm going to remove the pluggable components feature flag as we are relying on socket creation to work. So I think it should be sufficient....

Does this still happen if you restart Dapr after Redis restarts?

@thinkgandhi Dapr state stores require the implementation to honor the Update and Delete operations - how do you see that play out with Blockchain?

@juazasan is this ready now? Can we move this out of draft?