you would need to be able to enter them somehow (which goes back to the above discussion on step sequencing) or analyze the note sequence and decide on when to have a rest based on some criteria. you could do something like this:
say, you record note A for 2 second, 2 second pause, note B for 3 seconds, 1 second pause, note C for 6 seconds, 2 second pause:
AA__BBB_CCCCCC__
now you clock it externally. we could infer some timing information from the first note. so any pause that is equal or longer than note A will become a rest:
A_BC_
this could be one way, but i’m not sure it will work well for all scenarios.
edit: thinking about it more, it might actually work reasonably well this way. so then the length of the first note would be used to determine ties/rests. linearization would work similar way (or maybe have the new option in addition to how it works right now). i’ll give it a try.