Rune Allnor
10-15-2006, 01:11 PM
Hi all.
Imagine some activity that goes on 24 hours per day, 7 days per week
for some weeks or months. Imagine that the staff assigned to work on
this activity is chosen more or less random. Person A can have
a few stray shifts here and there while person B has a string of shifts
during one limited period of the main activity. Basically there is a
certain
turn-around of the staff.
Imagine further that some fault condition develops on the equipment.
Every now and then, a glitch occurs and need correcting. Due to the
staffing situation, no one technican is likely to see the glitch happen
more than one, maybe two times, in his watch, so no single person
is likely to see the glitch pattern developing.
What is needed is a "glitch log" system where each technician
logs all his jobs, however small, such that one can use the logs to
infer that one particular system tends to fail more than it ought to.
While the benefits of a working logging system are obvous, I can
see a number of pitfalls in the implementations, overhelming
everyone involved with paperwork would be the main one. My basic
philosophy is to have the techies do what they already do, with
as little overhead or corrections to their work, as possible.
I believe it would be far smoother to have somebody consend to
do just one tiny thing extra, than reform absolutely everything
in their method.
How would one design a "glitch log" system to be able to catch the
fault patterns, without drowning the technicians in bueraucracy?
Can this be done at all? What overhead is involved in achieving
something like this?
I am sure somebody out there, even outside NASA and FAA, have
done something like this and got it to work.
Rune
Imagine some activity that goes on 24 hours per day, 7 days per week
for some weeks or months. Imagine that the staff assigned to work on
this activity is chosen more or less random. Person A can have
a few stray shifts here and there while person B has a string of shifts
during one limited period of the main activity. Basically there is a
certain
turn-around of the staff.
Imagine further that some fault condition develops on the equipment.
Every now and then, a glitch occurs and need correcting. Due to the
staffing situation, no one technican is likely to see the glitch happen
more than one, maybe two times, in his watch, so no single person
is likely to see the glitch pattern developing.
What is needed is a "glitch log" system where each technician
logs all his jobs, however small, such that one can use the logs to
infer that one particular system tends to fail more than it ought to.
While the benefits of a working logging system are obvous, I can
see a number of pitfalls in the implementations, overhelming
everyone involved with paperwork would be the main one. My basic
philosophy is to have the techies do what they already do, with
as little overhead or corrections to their work, as possible.
I believe it would be far smoother to have somebody consend to
do just one tiny thing extra, than reform absolutely everything
in their method.
How would one design a "glitch log" system to be able to catch the
fault patterns, without drowning the technicians in bueraucracy?
Can this be done at all? What overhead is involved in achieving
something like this?
I am sure somebody out there, even outside NASA and FAA, have
done something like this and got it to work.
Rune