20,000 improvements to established methods

20,000 improvements to established methods. 100 well-known methods × 20 improvement operators × 10 contexts. 6734 are marked feasible and 13266 infeasible by a stated rule matrix — the infeasible ones are kept in the file because a ruled-out improvement is information. No improvement has been implemented or measured.

1. What counts as an improvement

Cheaper, faster, smaller, safer, or more portable — but not a different measurement. Each lead names an operator, a method, and a context, plus the characteristic risk of that operator.

2. Feasible improvements by operator

OperatorCount
cut sample volume 10x620
cut analysis time 10x530
replace a hazardous reagent490
miniaturise to chip scale410
parallelise the runs392
replace operator judgement with a metric360
eliminate a separation step350
add an in-line blank350
make the measurement ratiometric330
add a simultaneous second channel330
replace with a proxy measurement312
reduce the temperature excursion290
convert batch to continuous280
replace a consumable with a reusable one280
multiplex the assay270
add an internal standard270
replace an expensive component250
remove a calibration step210
switch reference to differential210
use a cheaper detector200

3. By method category

CategoryCount
spec1,534
sep1,344
prep1,344
bio1,092
img696
elec400
mech180
therm144

4. By context

ContextCount
industrial QC703
teaching laboratory703
materials production703
research laboratory703
environmental monitoring703
food safety703
clinical laboratory639
forensic analysis639
field and point of use619
remote or deployed619

5. The rule matrix

An operator is infeasible for a category when it contradicts what the method measures — you cannot "eliminate the separation step" from a separation method and still measure the same thing, and a clinical or forensic context forbids a proxy or a cheaper detector where traceability is required.

6. Limits


Incoprea Research · study 9, method improvements · 3 November 2026. Composition and computation. No experiment performed.