Boundary-Value Test Suite¶

Test or assessment — instantiates Dense-Subset Coverage Design

Adds explicit anchors at edges and transition points where nearby cases may behave differently.

Version: v1 · 2026-08-24 · History
Mechanism #: 922
Type: Test or Assessment
Form family: Experiment, Test & Rehearsal
Solution family: Optimization & Search
Problem family: Correctness, Conformance & Formal Validity Failure
Problem subfamily: Coverage, Partition & Set Accounting
Origin domain: Computer Science & Software Engineering
Also from: Statistics & Experimental Design
Instantiates: Dense-Subset Coverage Design

A boundary-value test suite deliberately thickens the anchor set exactly where behavior can flip: at thresholds, cutoffs, regime changes, and the edges of the valid range. Its defining assumption is the opposite of the interior's: everywhere except at a boundary, nearby cases behave alike, so one anchor covers a wide neighborhood; but straddling a boundary, two cases a hair apart can land on opposite sides of a decision, a formula, or a physical regime — so the tolerance there must collapse to almost nothing. The suite is the mechanism that spends its density budget on those thin, high-leverage seams rather than spreading it evenly, on the theory that a cover which looks dense across the interior can still be blind precisely where a small difference produces a large one.

Example¶

A team ships a change to the pricing engine of an auto-insurance quote system. Premiums depend on driver age through a schedule with hard steps — a surcharge below 21, a discount that switches on at 25, a senior adjustment at 70 — plus a floor and a ceiling on the final quote. A uniform test set of "typical" ages (30, 40, 50) passes cleanly and proves nothing, because every one of them sits comfortably in a flat interior. The boundary-value suite instead pins anchors on both sides of every seam: ages 20 and 21, 24 and 25, 69 and 70, plus a case engineered to land one cent under the quote floor and one cent over the ceiling. Running it, the suite catches that the discount actually triggers at 26 instead of 25 — an off-by-one that no interior case could ever have surfaced, because interior cases are exactly the ones the boundary doesn't touch.

How it works¶

Enumerate the seams. Walk the specification for every threshold, cutoff, cap, floor, and regime change — the places where the mapping from input to outcome is not locally flat.
Anchor on both sides, tight. For each seam, place a case just inside and just outside it, separated by the smallest meaningful increment. The coverage tolerance here is not a comfortable radius but a razor: adjacent cases must be distinguishable.
Add the corners. Where several boundaries meet — the low end of one range at the high end of another — place the combination anchor, since seams interact and the corner is where two off-by-ones can hide each other.
Leave the interior to others. The suite intentionally under-covers the flat middle; that space is cheap and someone else's job.

Tuning parameters¶

Edge increment — how far "just inside" and "just outside" sit from the seam. The smallest representable step is the strictest test; a looser gap tests the neighborhood but can miss a hairline error.
Seam inventory depth — how exhaustively thresholds are hunted, including implicit ones (integer overflow, empty inputs, unit changes) the spec never names. More depth catches more, at rising authoring cost.
Corner combination order — whether to test seams one at a time, in pairs, or in full combination. Higher order finds interaction bugs but the case count explodes.
On-boundary convention — whether the boundary value itself (exactly 25) is treated as inside, outside, or its own third anchor, resolving the ambiguity most defects live in.
Interior sparsity — how few interior anchors to keep as a sanity backstop, trading suite size against a thin safety margin in the flat regions.

When it helps, and when it misleads¶

Its strength is leverage: a handful of well-placed edge anchors catch the defects that a far larger interior sample never touches, because errors cluster at seams. It is the direct antidote to the archetype's boundary-sparse failure, where a dense-looking interior hides fragile edges. The technique has a long pedigree in software as boundary value analysis^[n1], and the same logic transfers to physical envelopes, eligibility rules, and regime transitions.

It misleads when it is mistaken for a whole cover. A boundary suite says nothing about the interior it deliberately skipped, so a green run proves the seams are handled, not that the middle is — treating it as complete coverage is the classic misuse. It also depends entirely on the seam inventory: a threshold nobody wrote down gets no anchor, and undiscovered boundaries are exactly where the surprises live. And it presumes the interior really is flat; where behavior is bumpy between seams, edge testing alone will miss it. The guarding discipline is to pair the suite with a mechanism that covers the interior on purpose, and to keep hunting for unstated boundaries rather than trusting the spec's list.

How it implements the components¶

boundary_and_transition_anchor_set — this is its whole reason to exist: the densified set of anchors sitting astride every threshold, edge, and regime change.
coverage_tolerance — it applies a locally tightened tolerance at seams (adjacent cases must be one step apart) while leaving the interior loose, realizing the archetype's region-varying tolerance.
dense_reference_subset — the edge anchors are contributed cases in the overall reference set, densest exactly where consequences turn sharp.

It does not define the distance metric the seams are measured in (that is Epsilon-Net or Covering Grid) or keep the seam inventory current as the specification drifts (that is Adaptive Refinement Loop).

Instantiates: Dense-Subset Coverage Design — it contributes the boundary-dense portion of the anchor set the interior mechanisms leave thin.
Consumes: Space-Filling Design or Epsilon-Net or Covering Grid covers the interior; the boundary suite assumes that flat middle is handled elsewhere and spends its budget on the seams.
Sibling mechanisms: Epsilon-Net or Covering Grid · Space-Filling Design · Coverage Heatmap · Nearest-Neighbor Assignment Rule · Adaptive Refinement Loop · Sensor or Service Radius Map · Anchor Case Library

Editorial Notes¶

Form Classification¶

Form family: Experiment, Test & Rehearsal

Rationale: Adds explicit anchors at edges and transition points where nearby cases may behave differently, making its operative form a deliberate probe, variation, simulation, or practiced execution used to generate evidence or readiness.

Independent corroboration: The frozen evidence defines Boundary-Value Test Suite as 'Adds explicit anchors at edges and transition points where nearby cases may behave differently', so its operative form is Experiment, Test & Rehearsal.

Review outcome: Independent reviewer agreement; high confidence.

Origin Attribution¶

Primary origin: Computer Science & Software Engineering

Origin pattern: Single lineage

Present-day reach: Multi-domain

Rationale: Software boundary-value analysis concentrates cases immediately inside and outside every cutoff, cap, floor, transition, and interacting corner where behavior can flip.

Related originating lineages:

Statistics & Experimental Design — Statistics contributes sampling, uncertainty, blocking, blinding, controlled comparison, or inferential discipline used here.