Turbine D, it was actually a little worse than that. The early 767s with JT9D-7R4 and CF6-80A engines had "supervisory" electronic engine controls - a fully capable hydromechanical control with a digital electronic EEC that would trim the fuel flow from the hydro control (basically a mid step between pure hydro and FADEC controls). If there was a malfunction or problem with the supervisory control, the procedure was to simply switch the EECs off and use the purely hydro control. At the time, the EEC switches were located on the aisle stand - right above the fuel RUN/CUTOFF switches.
There had been a previous event on a JT9D powered 767 where a pilot had, shortly after takeoff, reached down for the EEC ON/OFF switches, but instead set both the RUN/CUTOFF switches to CUTOFF - he quickly realized his mistake and turned the fuel back on - one engine restarted, the other went into a non-recoverable stall. They basically circled back around and landed.
When we (Boeing) investigated, we figured the pilots were idiots and it would never happen again
. Then it happened again (I'm thinking it was ~12 months later) with the CF6-80A event you referenced.
After that it was quickly determined we had a human factors issue, and the EEC switches were moved to the overhead (with the 'half a broomstick' switch guards ACMS referenced as an interim fix until the EEC switches could be moved).
BTW Al Murdoch - dual engine start on the 747-400 was a function of autostart, (which was an extra cost option - most Rolls and nearly all GE customers but only ~ half of Pratt purchased autostart). The APU could handle dual engine start for all three engine options on the 747-400, at least near sea level.