Ladies and gentlemen, before we return this thread on its tracks, I'd like to give my comments on some accidents mentioned here.
Regarding the Roselawn ATR-72 crash: there were two accident reports, one by NTSB and another by its French counterpart, BEA. They were pretty much the same in their factual information and analysis sections, but they differed wildly when it came to probable causes. NTSB blamed the airplane designers and inadeqate regulatory oversight during ATR certification. BEA accused the flightcrew of being negligent for continuing the flight in icing conditions when there were escape routes available. So who got it right? No one - both were seriously wrong.
NTSB claimed that ATR knew that their airplanes' ailerons are prone to hinge moment reversal, especially with ice contamination, and should have been redesigned. Problem is that aileron moment reversal on uncontaminated wing occurs at 26° AoA - deep into stall so it was considered to be a non-issue. Another problem is that atypical icing over Roselawn was caused by supercooled large droplets (SLD). At the time of crash, they were largely unknown fenomenon outside some narrow meteorological circles and even worse, they cannot be created by water spray tankers during icing certification. So ATR and DGAC sprayed and sprayed and sprayed ATRs from tankers and never had problems with ailerons. Also SLD tend not to create large ice deposits on leading edges, they flow chordwise and then freeze. Lack of ice on leading edges can lead one to conclude he's flying through light icing, while ice builds up out of sight.
However, it's absolutely wrong to blame our deceased coleagues for improper operation of their ATR. Sad fact is that they were faced with unknown phenomenon and so became test pilots without ever realizing it. While their actions preceeding loss of control were somewhat unusual, they were far from unreasonable. They did turn on de-icing level 3 (i.e. boots) but were never concerned about icing because there was not much ice on the airplane. Also there was argument that 'since it brought down the airplane, icing was severe'. Well gents, if ATR had powered ailerons, or if they were better designed, if there were brazilia or SF 340 instead of ATR, pax and crew would land happily at KORD and SLDs would be still be unheard of. Biggger de-icing boots, ice evidence probe, increased speeds for tkof/ldg, stick shaker that fires earlier with anti-ice turned on and whole bunch of procedures for dealing with icing are all legacy of Roselawn. It's very improper to chastize the crew for not adhering to procedures that were developed as result of their mishap.
Back to airbus. Halbsheim flypast. Airplane was too low. Then it was too slow. On the top of it, idle thrust comes into equation (comanded, not as result of failure). I apologise for not having the nerve to explain why flying low and slow can be health damaging with any airplane.
Strasbourg crash was caused by misseting 3300 fpm instead of 3.3 degree flight path on FCU (that's autopilot control panel in airbusspeak). FCU has been redesigned since to prevent similar occurences, and i don't know whether it was designed by real pilot, but by my definition, real pilot always check if the autopilot does what it's commanded. With the speed they had, they shouldn't have descended with more than 1000 fpm and they hit 3300. Also PFD have shown unusal nose down attitude, yet they missed all the clues and there was no GPWS to save them.
Belle Harbor crash wasn't caused by excessive rudder deflection. Yes, rudder has some design issues but what tore off the tail was rapid rudder reversals. For the time being, they seem to be commanded and not caused by control unit failure. For those unimformed: you never, ever cycle any of your flight controls rapidly during flight - dangers involved are: PIO, control failure or even airframe failure. So this crash wasn't really type specific.
Toulouse A330 crah is here: I'd just like to add that there are claims that crew went out flying after long day in the office, so fatigue could also be the issue.
Regarding the Gibson dive: crew was blamed for trying to fly with flaps extended and slats retracred, to improve cruise performance. Couple of years later test flight in this configuration was atttempted to prove the theory but it only disproved it, as so rigged 727 was unable to climb to high twenties, let alone cruise above FL300. This claim comes from Stanley Stewart, and while I couldn't verify it, I wouldn't mark mr. Stewart as unreliable source.
To sum it up: FBW or no FBW it's still subject to aerodinamic forces and gravity like any other airplane. I'll just repeat what wileydog3 said: "do the same thing with the 'bus as with any other airplane.. play to the strengths, respect the weaknesses".