<snip>
The ALCOA factory where we did the Old Way/New Way trial, as described on
the web site, had two large overhead cranes in the casting pit building.
These cranes run on rails that go from one end of this huge bulding to the
other and they carry the crucibles of molten aluminium. The control panels
inside each crane were identical in appearance but not in function. In one
crane the top left button turned the crane to the right; in the other crane
the same button turned that crane to the left. Crane operators, when
transitioning from one crane to the other, would experience interference
from their prior learning (the other crane) and this resulted in several
collisions between cranes. The company was unable to cure the problem with
re-training so they ripped out both control panels, installed panels
identical in appearance AND function and then re-trained all the crane
operators to use this new control panel, all at considerable time and
expense. Old Way/New Way would have fixed this problem by using control
panels that were coloured differently and then re-programming operators
with specific set of control behaviours for each panel/colour, enabling
them to transition readily from one to the other, and back again. The
panel's colour would serve as a cue for the "correct" set of control
responses for that panel.
<snip>
I do not believe that training should be a crutch for bad design as has been
identified in this case. Replacing the panels is the correct solution, in
this case. Anything else is second best.