OK - I do know a bit about airfoils, but have never understood why a super-sonic wing looks like a trapezoid, while sub-sonic look (somewhat) like their inspiration - the flight feathers of birds. Then there was the BD-10J (Jim, give it up) with it’s weird leading edge.
Can anyone explain this?
Please, let’s keep this away from the newtonian v bernoulli principles, OK
And: NO airplanes on treadmills 
That’s a real wide-ranging question.
All else equal, an airfoil with a leading edge at 90 degrees to the airflow is most efficient at processing airflow into lift. E.g. the Piper Cherokee “hershey bar” wing.
However, as speeds approach the high subsonic to trans-sonic range, the drag gets awful. Adding leading edge sweep pushes the knee in the drag curve upwards, at the expense of reduced lifting efficiency at all speeds. The efficiency reduction is due to spanwise flow.
Meanwhile, as wings get longer (e.g. any airliner), there are structural efficiencies to be gained by tapering the wing. So you get a trailing edge sweep which is less than the leading edge sweep.
Once you get supersonic, there is further advantage in having the leading edge wing sweep exceed the Mach shock angle. Hence the deep deltas of Concorde, the Dassault Mirage series, the F102, F106.
The pure trapezoidal shape of the T-38 & F-104 are not found in more modern designs. That was state of the art thinking in 1960.
Also, understand that different wings are for different purposes. The wing on a Rafale or F-16 is designed for aggressive manuevering, not for cruise. The wing on a 777 is designed for cruise, not for aggressive manuevering.
While the intended speed regime has an impact on the design, the cruise vs manuevering mission has a lot more impact on overall wing shape. Compare, for example, the wing shape of a DC-7 and a P-47 or P-51. Similar speeds, similar tech. Very different shapes.
It’s an error to compare the wing shape of a 747 and an F-16 and conclude the difference is due to their different max speeds. It’s got a lot more to do with their different limit G loads and flight envelope optimization points.
Belated ETA:
As we get farther into building airplanes out of something other than aluminum, we’ll see other different shapes emerging. Right now there are some aerodynamic compromises made to achieve weight & constructability goals. Different materials will still have tradeoffs, but the optimal point will be in an different place in the tradeoff space.
The future materials almost on offer today are not that radically different from a structural POV, so the new shapes won’t be grossly different from what we see today. But they will be obvious to a discerning eye.
Hi all - sorry for the multi-post fest here.
Upon re-reading the OP, I think I totally missed the question. TheOP was probably asking about wings cross-section, whereas I was speaking above about wing planform.
The uber-short version is a supersonic airfoil generates lift via assymetrical shock waves, whereas a subsonic one generates lift via turning the subsonic flow. Two very different lift sources drive two very different lifting shapes.
After that, it all comes down to creating a shape which works tolerably well across the whole speed regime from liftoff / touchdown to Vmo/Mmo.
The shape which works well at 100KEAS/M0.15 is quite different from one which works at 500KEAS/M0.75.