← Back to BLACKWIRE PRISM BUREAU TESTING TAXONOMY Diagram illustrating the dual‑axis test scope model with function, module, and system levels.

The dual‑axis model separates testing scope by programming construct and statefulness, a core proposal from the essay.

TESTING VOCABULARY REDEFINED: NEW CLASSIFICATION SUGGESTED

*The software testing lexicon remains vague despite years of practice. A recent essay maps the gaps and proposes a five‑category framework anchored in scope, state, and purpose. The shift could make test planning actionable rather than vibe‑driven.*

By PRISM Bureau - BLACKWIRE  |  November 22, 2026, 17:00 CET  |  testing, software testing, testing taxonomy, design by contract, test classification

A recent essay posted on August 16, 2025, flags the chronic imprecision of software testing terminology. The author, frustrated by undefined units and vague "vibes" surrounding integration and end‑to‑end tests, maps the current landscape and sketches a new classification. The piece, tagged under testing and clocking a 19‑minute read, argues that without a shared vocabulary, teams cannot reliably scale or prioritize tests. By dissecting scope, state, purpose, and execution method, the author aims to turn testing from an art into a disciplined, resource‑aware practice.

Why existing test labels fall short

The author notes that terms like unit, integration, and end‑to‑end test are loosely defined. Unit tests lack a clear definition of what a "unit" is. Integration tests are described as arbitrarily scoped modules, and E2E tests underspecify both the behavior under test and the methods—whether automated agents, human supervision, predefined actions, or random exploration. Stress, load, and penetration tests exist without a coherent classification, leaving practitioners to rely on "vibes". This lack of precision hampers resource allocation and test upgrade paths.

What are the two axes of test scope?

The essay introduces a dual‑axis model for scope. The primary axis follows programming constructs: function, class/struct/type, module, or system. The secondary axis distinguishes stateful from stateless entities. Most functions are stateless; most systems are stateful. This distinction matters because inputs for functions are arbitrary language structures, while system inputs are serialized directives that naturally narrow the input space. Transforming stateful tests into stateless equivalents is possible by embedding state into inputs, but complex realistic states may be infeasible, prompting a simulation‑style approach.

The concept of a unit in unit test is not really defined.

How does Design by Contract tighten input spaces?

Design by Contract (DbC) appears in C++26, D, Kotlin, Racket, and via libraries in Python and Rust. DbC lets developers annotate assumptions about inputs, turning unchecked pointer dereferences or undocumented promises into runtime or compile‑time checks. By constraining the input domain, DbC reduces the mismatch between a language's permissive type system and the real, narrower set of valid inputs. The author cites a hashmap example where every lookup is technically fallible, yet the program knows all keys exist. DbC, expressive type systems, and better domain modeling resolve this tension.

Can a five‑category testing taxonomy work in practice?

The author proposes roughly five orthogonal categories that can branch. The first category covers scope (function, module, system) and statefulness. The second addresses purpose: functional correctness, performance timing, security resilience, etc. The third concerns hypothesis testing—each test aims to falsify a hypothesis. The fourth looks at execution method: automated agents, human supervision, predefined actions, random exploration. The final category ties resources to test depth, allowing upgrades or downgrades based on availability. While biased toward random testing, the framework promises actionable recipes for any situation.

Frequently asked questions

What are the two axes used to define test scope?

The essay defines scope along a primary axis of programming constructs—function, class/struct/type, module, or system—and a secondary axis of statefulness, distinguishing stateful from stateless entities.

How does Design by Contract help limit the input space of functions?

Design by Contract lets developers annotate input assumptions, turning unchecked cases into runtime or compile‑time checks, thereby narrowing the language's broad input type to the realistic set of valid inputs.

What classification does the author propose for software testing?

The author suggests roughly five orthogonal categories covering scope and state, purpose (functional correctness, performance, security), hypothesis testing, execution method, and resource‑driven test depth.

If the testing community adopts the proposed five‑category framework, the chaotic "vibe" taxonomy could give way to a disciplined, resource‑aware language. Clear scope definitions, state considerations, and purpose‑driven hypotheses would let teams allocate effort where it counts. Until then, developers will continue to wrestle with ambiguous labels, risking misaligned testing strategies and wasted resources.

Sources: Hacker News – A Better Vocabulary for Testing (https://alperenkeles.com/posts/vocab-for-testing/)