a cutter slowly loses its edge. dimensions shift before anyone hears the change.
run ./quality-observer
quality engineering, softly observed
the process drifts
every process drifts. quality is how we notice.
define entropy
what is entropy?
entropy is a way to talk about disorder. left alone, tidy things become less tidy. clear signals become noisy. sharp edges wear down. instructions get copied, interpreted, skipped, and forgotten.
in everyday quality work, entropy looks like drift, uncertainty, decay, and the slow loss of control. it is not always dramatic. sometimes it arrives as a tiny change that nobody notices until the data starts whispering.
list entropy --quality
entropy in quality
a process rarely falls apart all at once. more often, it wanders.
a measurement system ages, moves, or gets bumped. the numbers still look official.
repetition dulls attention. the eye learns the pattern and skips the surprise.
the work changes, but the document does not. people build from memory instead.
a new lot, material source, or hidden process change adds variation upstream.
temperature, setup, training, fixtures, and timing all tug the process off center.
inspect human-attention
human attention is part of the system.
inspection depends on people, and people are not cameras. we adapt. we predict. we get tired. after seeing the same part hundreds of times, the brain starts saving energy by recognizing the usual shape instead of studying every detail.
this is inspection fatigue and habituation. it does not mean inspectors are careless. it means the quality system must be designed with human limits in mind: rotation, lighting, clear criteria, error-proofing, and checks that help attention return.
load anti-entropy-tools
quality tools pull the process back into view.
audits
compare the real process to the intended process before habits become hidden.
calibration
make sure measurement tools still tell the truth.
spc
use data over time to spot signals, trends, and unusual variation.
control plans
define what matters, how it is checked, and what happens when it moves.
ppap
prove a process can meet requirements before full production begins.
car / scar
document problems, correct causes, and verify the fix stays fixed.
gage r&r
separate real part-to-part variation from noise in the measurement system itself.
process capability (cpk)
ask whether the process, as it actually runs, can even meet the specification.
5 whys
move past the first explanation and keep asking what allowed the failure.
inspect entropy --interactive
drag the entropy scale.
move from low entropy to high entropy and compare real quality examples.
drag the control. the readout updates with the condition, example, detail, and a practical quality response.
low entropyhigh entropy
open quality-archive.log
quality archive
a quiet record of where entropy shows itself before the system fully understands it.
log_001 ::
xbar drift
missed defect after 8 hours of repetition
Data: X-bar moved from 49.8 to 51.2 over 6 hours
Root cause: Temperature change in the machining center
Action: Implemented hourly temperature checks
log_002 ::
ucl watch
when variation became normal
Data: 5 consecutive points approaching UCL
Root cause: Gradual tool wear not detected early
Action: Reduced inspection interval from 2h to 30min
log_003 ::
operator loop
the psychology of inspection fatigue
Data: Defect detection rate dropped 23% after 4 hours
Root cause: Visual fatigue and habituation
Action: Implemented 15-min breaks every 2 hours
log_004 ::
system note
entropy in manufacturing systems
Data: 12% increase in variation after supplier change
Root cause: Unapproved material substitution
Action: Supplier audit and approval process review
cat /notes/on_time.txt
- control is temporary.
- stability is rented.
- order decays not only in systems, but in attention.
- where do we see entropy hiding in systems we trust?
micro spc chart // xbar watch
live sample stream · animated control window
tail shop-floor-notes.log
notes from the shop floor
"the chart did not shout. it leaned."
"a gage can be wrong with perfect confidence."
"the operator knew the fix. the document had not caught up."
"variation is a message. quality is learning how to listen."
whoami
who's watching the process.
built by a quality professional who likes watching systems drift back into control. currently looking for quality coordinator / quality engineering roles in the milwaukee area.