Not quite. The windows were square, true - but the corners were rounded, specifically to avoid cracking at the corners. Look at pictures, & the rounding is obvious. The cracks didn't begin directly at the corners of the windows, but around rivet holes near the corners, from a combination of the extra localised stress from the overall (albeit with rounded corners) squareness of the windows & the fine cracks created by punch riveting. They then spread along stress lines in line with the straight bases of the windows. There were also stress points caused by the sharp edges of countersunk bolt holes, helping crack propagation.
The square windows might have been fatal eventually, but it was the combination of them with rivet holes with fine cracks caused by punch riveting (a production line decision contrary to what was specified by the designers) which caused those two fuselages to fail so dramatically & so soon.
The official verdict was failures caused by a combination of design flaws & manufacturing faults.
Note that the fuselage which was tested to destruction, revealing the causes of the failures, underwent 3000 pressurisation cycles before failing, & at significantly higher pressures than encountered in service. The crashes were at lower numbers: one had fewer than 1300 cycles. The difference may be due to slight manufacturing differences. Punching rivets through aluminium produces rather variable results. The holes are rough edged, but unpredictably so.
The investigators concluded there was a failure risk any time after 1000 cycles, & a maximum of 9000.
Note that other jet airliners under development were modified on the drawing board as a result of the Comet crash investigation. They were 'basic' design flaws that everyone else was also making. The state of the art was being pushed to its limits & beyond. A lot was learned about fatigue & stress, which might have been learned from Boeing or Douglas crashes if De Havilland hadn't got there first.
