The feedback from the SoCG rejections was actually quite detailed and helpful, much better than I've been used to seeing from theory conferences. And I've been on too many program committees myself (past and future) to resent it when some committee members say something about a paper that, taken out of context, seems nonsensical. But here's my unhelpful summary of what they wrote anyway, just because I feel obstreperous today.

xyz Graphs:
not geometric enough; graph drawing is off-topic for SoCG.
Self-overlapping curves:
answers too many of the questions it opens.
Quadrilateral mesh partitioning:
it could have been a good applied paper, if it had more theory, and it's too long, but should have been longer. (Note: all my submissions fit strictly within the ten page limit.)
Road networks:
would have been better to have left out the motivation.
Straight skeletons:
should have submitted in the form of a 30-page journal version rather than a 10-page conference abstract.