In answer to jxk, yes new aircraft types must be fatigue tested though not now in a water tank. However B737 was never properly fatigue tested relying instead on read over from 707 and 727.
Further to my post #126, I wanted to better explain how the lap joint evolved, by posting some illustrations from NTSB report on Aloha accident. That doesn’t seem possible on pprune so those who are interested should refer to pg 15 and on of the report at: http://www.airdisaster.com/reports/ntsb/AAR89-03.pdf.
Prior to LSN 292 the outer skin at the lap was fatigue critical at top row of rivets (the most heavily stressed row), due to stress concentrations at the countersink and likely failure of the cold bond process. After LSN 292 the outer skin was locally reinforced with a hot bonded doubler. In both cases the next weakest link is fatigue of inner skin at bottom row of rivets. That's what failed on SWA aircraft. And if reports of loose rivets are true then premature failure is not surprising. What is alarming is the much greater difficulty of detecting cracks in inner skin compared with outer. There’s a further hazard in that when the inner skin starts to crack it is initially braced by the outer skin and cracks become excessively long before controlled decompression happens.