As a newcomer, I find the P 39 re-design discussion civil, well-informed, friendly and fun. Here are some thoughts on the subject: possible changes in design that might have improved the P 39 performance as a fighter or attack aircraft. Though perhaps bending the original challenge, it seems interesting to consider changes that could have been made in later years (items 1 and 2 below) as well as in 1938.
1. Revisit the airframe structural design to have it conform to the adequate British (or German) stress limits rather than the more conservative American values. Relaxed stress limits contributed from the start to the fine performance of the F8F Bearcat and substantially improved the already excellent P 51 Mustang. The empty weight of the P51F, for example, was 5635 lbs. compared to 7125 lbs. for the P 51 D, a huge 20.9% improvement. Similar weight saving in the P 39 would improve, in probable order of value, rate-of-climb, range, maneuverability, top speed, landing speed and take-off run and would allow some latitude for weight-increasing modifications. Admittedly there would seem almost no chance of the Air Corps agreeing to this until well into the war.
The extremely rugged (sometimes over-rugged) construction of U.S. fighters seems to have been the common cause of rates-of-climb lower than they should have been.
1.1. As part of the airframe re-think, add armor, but make it part of the load-bearing structure of the airplane as was done with the IL 2 Shturmovik and the Fairchild A 10. The compact grouping of the pilot, cooling system and engine make the P 39 well suited to protection from below by one undersurface sheet of armor. A particularly thick sheet would be a nice feature of a ground attack version of the plane
2. I agree with several of you that, especially for the fighter role, the engine should be changed to a Merlin. There seems little chance, though, that this could be considered before September 1940 when it was agreed that Packard would produce the Merlin in the United States.
3. Increase the size of the leading-edge air intakes, ducts, and, if need be, radiators, to provide satisfactory cooling. The intakes have always LOOKED too small, noticeably smaller in fact than the leading-edge intakes on the radial air-cooled F4U Corsair. By all means keep the intakes on the leading-edge, however, because they are an asset there. As with intakes right at the nose, such as on the P 47 and F6F, they avoid airflow shear or turbulence from upstream surfaces, but do so without increasing the drag-producing cross-sectional area. In may opinion, leading-edge intakes, sometimes used only for the oil-cooler or carburetor, likely contributed to the good performance of aircraft such as the B 17, Mosquito, F4U Corsair, F8F Bearcat, Hawker Tempest and Fury, Yak 3 and 9, and the promising Westland Whirlwind.
3.1. Optimize ducting to fully exploit the 'Meridith effect' (1934), the jet thrust of heated air from the radiator, as was done with the Spitfire and Mustang..
3.2. Remove the carburetor air-intake from behind the canopy and incorporate it into a leading-edge intake. This would slightly reduce drag and accommodate the change to the Merlin which, unlike the Allison, has an updraft carburetor.
4. Remove the nose wheel. The Airacobra has always seemed too small to afford the luxury of a tricycle landing gear and performance in the air should take precedence over landing and ground handling. (The Bf 109 clearly followed this precedence a bit too far.) Removing the nose wheel frees valuable space in the nose (expensively bought by the unorthodox engine location), saves weight, and means less to go wrong or to maintain. As demonstrated by the Airacobra's tail-dragging naval cousin, the Airabonita, the main wheels can be attached to the front spar and stowed on retraction between the front and main spar. With the Airabonita, however, the radiators and ducts were moved backwards, out of the way, to the undersurface of the wing. In my opinion this should not be done. Instead, the wheels should be attached further out on the wing even if this requires reinforcement of the front spar. Alternatively, they might be retracted directly backwards as on the F4U. With the latter option, a notch would be needed in the main spar to accommodate the landing gear strut and the spar would need reinforcement at the notch.
5. Change the armament. The elimination of the nose wheel allows the general sentiment of Piper106, Shortround6, Markus, kreighund and delcyros to be taken further. With no nose wheel, a pure fighter version of the Airacobra would benefit from removing the 37mm cannon and going to 5 or 6 closely grouped guns in the nose, all 50 caliber with a common trajectory. A bomber interceptor might similarly be armed with, say, five identical 20mm cannon. Also, with no 37mm gun and no nose wheel to contend with, what the British claimed was extremely poor armament accessibility could be improved. A ground attack Airacobra, however, should retain the 37mm cannon and combine it with, say, four 20 mm cannon in the nose, even if this means fairing over unavoidable bulges.
6. On a fighter version of the plane, change the seating to give the pilot a much more reclining posture as in the Bf 109. This is effective and important in reducing drainage of blood from the brain in a tight turn or zoom. In the Airacobra, it also makes more space available for fuel,under the pilot. With no the 37 mm cannon, there would be room forward for the pilots feet. On a ground attack P 39, the original seating should be kept.
7. I agree with delcyros that automatic wing slats should be added to improve maneuverability and lower the stalling and landing speeds (as on the Bf 109) and perhaps mitigate, or even eliminate, the tendency of the Airacobra to tumble and go into a flat spin. The disadvantage, as experienced on the Bf 109, is possible disruption in aim when the slats suddenly deploy.
8. I agree that getting rid of the automobile doors is a very sensible idea. This would save weight and would simplify and improve both fuselage and canopy structure. A further minor point, if the new canopy slides, an antenna mast could still be avoided, as on the FW 190, with a pulley/spring arrangement inside the canopy.