It seems obvious to me that any voting system that involves ranking would tend to discourage more open ballot access, as people are more likely to honestly and accurately rank 2 or 3 people than they are to rank 10. If it can handle proportionality almost as well as single transferable vote, without additional complexity to the voter, it may actually be better fro ma practical standpoint.
While there are some good arguments for allowing ties in ranked ballots (unforced vs. forced preference), I think there are pretty big problems with forcing tied rankings.
> If it can handle proportionality almost as well as single transferable vote
I don't think it can, but because there is no clear mapping between actual preferences (or preferences that would be provided on an unforced preference ballot) and approval ballot (or any other limited-rank-preference ballot) markings, that's a hard empirical question that's not really analytically addressable.
Approval plus proportional is a bizarre concept; approval, to the extent it makes any sense for public elections (and there are serious problems with it for that purpose) makes sense really only for inherently single-winner elections.
For proportional systems with large bodies elected together, proportionality to first-party votes is probably the most sensible; for bodies small enough that doing that has significant wastage due to the threshold necessary to get one seat, something like STV (possibly modified to drop loser elimination) makes more sense than approval.
(A proportional, STV-like generalization of approval with a form of winner elimination might make sense, but seems intuitively worse than STV itself, even if approval ballots had consistent meanings.)