Infrastructure · Data Structures
MECE
Mutually Exclusive, Collectively Exhaustive partitioning
Definition
Mechanism
Exhaustive partition. Divide the problem space along one stated criterion into categories that are Mutually Exclusive (no overlap) and Collectively Exhaustive (no gaps). Test: can any item belong to two categories? Is any item uncategorized? Refine until both pass.
Invariants
- Exclusivity: Intersection(Subset_i, Subset_j) == NULL.
- Exhaustiveness: Union(Subsets) == UniversalSet, with any 'Other' subset within `misc_tolerance`.
- The subsets at a level partition along a single stated dimension.
Risks
Failure modes
The Other Bucket Trap
'Misc' exceeds `misc_tolerance`, so the partition covers the space without describing it.
Dimensional Error
partitioning criteria mixed, e.g. Colour beside Origin.
False Exclusivity
edge cases quietly assigned to force a fit.
Relationships
Root pattern: builds on nothing in this library.
Used by
References
Mind
Flag a one-label classification forced onto an object whose relevant properties point to several categories
Accept-set intersection protocol with commit-reveal for interest-based negotiation
Reference
Parameters
max_depthInt, [1, 10]- How deep the tree can go
misc_toleranceFloat, [0.0, 0.1]- Max % of items allowed in 'Other'
Data schema
{
"type": "object",
"required": [
"root_set",
"subsets",
"completeness_proof"
],
"properties": {
"root_set": {
"type": "string",
"description": "The universe being partitioned"
},
"subsets": {
"type": "array",
"description": "The mutually exclusive categories"
},
"completeness_proof": {
"type": "string",
"description": "How exhaustiveness was established, including the disposition of edge cases. Required because an unevidenced exhaustiveness claim is what False Exclusivity below consists of."
},
"criterion": {
"type": "string",
"description": "The single dimension this level partitions along."
}
}
}