Trigger functionality in direction = up vs direction = down

[andrewpbray]

Conversation continued from #47 . . .

But if we have, say:

1. enter red sticky
2. discuss red
3. discuss red
4. discuss red
5. enter blue sticky
   What happens if someone scrolls down past blue sticky, then back up? If you scroll back up through 4/3/2, aren't you still looking at blue until 1 enters again? Are we backtracking to find the last focus-on, or do users just need to repeat focus-on for 2/3/4? (I think the latter is a good solution, I just want to make sure I understand)

Hmm, great question. You're right about the expected behavior. If 1 and 5 are the triggers, on direction=up, there will be a mismatch between narrative blocks and sticky.

I think it's a question about what the best defaults are. Different options would include:

1. remove all .cr-active on exit. (requiring the user to re-add the focus for blocks where they're still talking about the same sticky).
2. use what we have but by default use the filter to propagate the focus meta down blank narrative blocks. this would lead to cleaner source docs, which I value.

I think most authors will compose their stories top to bottom and most readers will read them that way, which I guess is why I favor a more minimal source that privileges that direction. But we could built in exceptions that allows for directional symmetry. You can imagine that we:

A. introduce a whole slate of what are essentially on-exit behaviors that could be added to trigger like 1. This actually starts dovetailing with the use-case presented by the Permutation Test Demo #46.

B. introduce a .cr-clear class that can be added to triggers like 2 that, on entrance, clear all cr-active (and override the behavior of 2 until it hits another trigger)

Thoughts?

[andrewpbray]

Follow-up based on zoom call: we're going to go ahead with 2!

[andrewpbray]

Note to future self: once this is complete, consider adding in a space block feature.