To be perfectly honest, I don't see a "good" solution being implemented, at all. Sorry for being so pessimistic... I believe that a physical barrier between flight crew and the passenger cabin is a good idea, but I can't see any aircraft manufacturer (either The Boeing Company or Airbus Industrie, as the two major widebody players) implementing it.
Why do I say this? Firstly, it will cost a truckload to modify/retrofit existing aircraft designs, so I doubt you will see it in existing aircraft lines. Certainly, the cost of doing this is far outweighed by the cost of the accident, but I imagine that a cost/benefit analysis would find it more expensive than it was worth. Secondly, there are weight considerations, and other "tolerance" issues to take into consideration, such as stresses already present on the airframe.
In general, aircraft designs as much as possible are based on the concept of commonality; that is, any new aircraft is similar to existing aircraft. The reason for this is so that air and ground crew need not retrain for each type, a simple familiarisation course would be suitable. In addition to this, it allows ground support to keep less stock on hand, thereby saving cost - for example, most of the internals of the B757 and B767 are similar; ergo a lot of the parts can be "mixed and matched" for these types.
I needn't point out that to retrofit existing aircraft lines (and any new derivatives) with a "firewall" between aircrew and passenger, and an additional pressure door and associated paraphernalia, would be a hugely difficult and expensive task - if existing aircraft designs allowed it at all. New aircraft designs, such as the "Boeing Sonic Cruiser" that has been mentioned once or twice, will probably share some commonality with existing designs - I would predict that a large proportion of the airframe, particularly the centre of the fuselage, will remain the same as it is currently. Aerodynamically, it works.
Aircraft tolerances are tight at the best of times - an example is the speed, at altitude, at which an aircraft must fly (it's so-called Mach number). Often, at altitude, there is a very limited gap between the minimum safe flying speed of the aircraft, and the maximum safe flying speed. Too slow, and the aircraft suffers a stall. Too fast, and the airspeed relative to some "bumps" on the airframe, such as antennas, exceeds the speed of sound, causing excess stress on those components (and the skin of the airframe). Adding extra "bits" to the aircraft (extra doors included) adds extra variables which need to be controlled and measured; this is an expensive process. Once again, it's cheap compared to the cost of lives, but the cost of making the modifications may well exceed the benefit of doing so, on average.
An alternative solution would be to put a better physical boundary between the cockpit and the cabin itself - say a more secure door (bullet-proof? fire-proof?) with stronger interlocks, which mandatorily is locked while the aircraft is in transit, and must be unlocked from the cockpit side. This might work, although there is still margin for error (locks fail, people forget to arm locks, suitably determined criminals will find ways around it, etc). It would take quite a push for this to be implemented, however, as once again this may require extensive modifications which both airlines and manufacturers would be resistant to making.
I don't know what US air law is like, but I know that it is very rare that the transponder needs to be or should be switched off in-flight, here in Australia. Matter of fact, I believe it should never be switched off in-flight. Perhaps some form of interlock on the transponder is required; a code, perhaps, something like the RSA SecurID codes, that changes regularly and can only be supplied by a central authority, should be required to change the transponder to the off state. Any attempt to switch the transponder off, could perhaps lead to automatic (and silent) change to the emergency state. While this would not prevent incidents like this, it would certainly give authorities some time to react, and a clear method for identifying the aircraft.
A physical, complete, barrier is by far the most effective way of preventing access to the cockpit, but the expense will probably rule it out. On the other hand, the "next best" efforts are probably not good enough.
I know this is a hard time for those who live in the US, and I hope this comment hasn't struck a nerve.