From what I have learned by working with Phoenix LiveView by now is that you do not want deeply nested live-views. One Route LiveView to take care of params, two more layers is ok, but it becomes extremely hard to follow if there are more layers.

I think I have still not 100% figured out what components are supposed to be used with, but I can tell you that usually nesting components is a problem.
We are using components to preload more efficiently, and to take care of events, but every parent using the component has to subscribe to a higher level event… Lets just say, we have not figured it out yet.

If you have a lot of LVs, let them communicate via Phoenix Pub. Elixir is extremely good in handling different processes, as long as they communicate with each other - I actually started always returning {:noreply, socket} in handle_event/2 and let my handle_info/2 handle it, as long as I know that there is an event triggered. And LiveView diff has become so good and fast… Honestly, I have no idea how they do it, but you can literally just send thousands of updates in 5 seconds, and your user does not even see anything.

Let me know what your current set-up looks like, LiveView is still a very new technology, and I want to hear other people’s thoughts on how to do it!

Yes, same here. I think it really helps to remember that components are in the same process. Nested components are communicating with a parent (always the parent liveview), so at what point should all the components have their own liveview?
Component events in our current model usually trigger a global event, which the parent should subscribe to anyways, and not change their own state at all, because the parent will take care.

I don’t know how you guys are using LV and components, but I would like to learn!