Intent drives the assurance problem.
The first question is not which test should exist, but what must be true for the intended product outcome to be supportable.
IDQA is a methodological research proposal for making the path from software intent to quality assurance explicit: derive what must be guaranteed, determine what evidence is appropriate, observe the system, and bound the conclusions to what the evidence can actually support.
Software can be locally correct and globally wrong. A requirement can have an assertion, a test can pass, and an implementation can behave as coded while the available evidence still fails to establish that the higher-level product intent was realized.
The first question is not which test should exist, but what must be true for the intended product outcome to be supportable.
A deterministic assertion, temporal trace, visual observation, source analysis or human judgment may each be appropriate for different claims.
Evidence supporting child/component properties does not automatically establish a parent journey or composed product guarantee.
IDQA investigates whether intent, guarantees and evidence requirements can remain durable while scenarios, automation and executions are generated, reused or discarded according to need.
IDQA is being evaluated through literature mapping, explicit research questions, competing explanations, controlled benchmark cases and a reference implementation. Claims of generality, effectiveness and artifact durability remain hypotheses until the experiments support them.
Costa, Rodrigo Candido. “Intent-Driven Quality Assurance (IDQA): Working Specification and Research Agenda.” Version 0.1. IsenForge Research, 28 September 2026. DOI: 10.5281/zenodo.23032992. https://isenforge.com/idqa/specification/v0.1/