The overhang quality is on the low side of OK, but that’s without any particular configuration tweaking:
A few strands didn’t quite anchor to the far side of the overhang opening and hang down inside:
Curiously, the slab on the other side transitioned from the usual 0.10 honeycomb fill to completely solid about half way up:
As with all the other objects, there’s no stringing or oozing. Looks good to me!
The slic3r configuration, which doesn’t show the bridge speed of 100 mm/s:
; generated by Slic3r 0.9.8 on 2013-04-08 at 19:55:09 ; layer_height = 0.25 ; perimeters = 1 ; top_solid_layers = 3 ; bottom_solid_layers = 3 ; fill_density = 0.10 ; perimeter_speed = 100 ; infill_speed = 200 ; travel_speed = 500 ; scale = 1 ; nozzle_diameter = 0.35 ; filament_diameter = 1.70 ; extrusion_multiplier = 0.9 ; perimeters extrusion width = 0.40mm ; infill extrusion width = 0.40mm ; first layer extrusion width = 0.39mm
The OpenSCAD source came directly from Starno’s object on Thingiverse.
The Smell of Molten Projects in the Morning 
The solid infil is because you probably had the “Solid Infill Threshold Area” option… it defaults to 70mm2 ..Set it to something small if you want to avoid it switching to solid.
P.s.. Slic3r 0.9.9 is out, which has a LOT of nice options, such as raft (not so relevant these days) but also “Avoid Crossing Perimeters”, “Support material to layer x” etc etc. Well worth switching to, I haven’t found really any places it falls down compared to 0.9.8
Ah-HA! The solid fill starts where the outside area of that solid wall is about 105 mm^2, but the inside area would be around 70 mm^2. Thanks for pointing that out!
Shortly after that, I went directly to the Git head at 0.9.10-dev and immediately turned on the “Avoid crossing perimeters” option, which improved the surface finish. Back in the day, the Skeinforge “comb” option did something similar, but retracing infill threads at full throttle without acceleration control shook the TOM too badly to be practical.
This is working out wonderfully well…