<< Duvie, I'll take your comment personally, then point out a big difference between the life death stats for the ECS board and the Shuttle board. You will note that the ECS distribution contains a statisticlly signifigant pile of deathes between 2 weeks and 2 months, even some after 2 months, While the shuttle shows NO deaths in this time period. >>
In the actual AK31A survey, there was no category for deaths after 2 months. When I started the survey, I thought the board had been out for only two months (in fact, I've seen no indication it was widely available before October; I suspect some folks may have mistakenly piled in their AK31 (rev 2) stats).
The K7S5A has been around quite a bit longer than the AK31A, so there is a possibility that some of the differences between the two are a result of "statistical aging."
That said, I think there were some problems in the K7S5A that caused deaths after a few weeks, possibly due to something as simple as battery drainage (especially for those boards that came with JP4 on the "CMOS clear" setting).
Please folks, let's not read too much into this. Bad batches of mboards do go out, and some folks can get really bad statistical clusters -- just due to random bad luck. Perhaps a more important survey would be: find the 4 major vendors of these boards, and ask how many boards purchased from each vendor were still alive, or were now dead.