Design-space studies
Commission
Opened on 2026-10-28 at the operator's request. Explore whether a consequential Commonplace design decision should pass through an explicit design-space study before a proposal or ADR. Such a study is a revisable map of the frame, the constraints and their provenance, decision axes, compatibility rules, candidate regions, decision-critical unknowns and the posture of each open choice. The goal is to find out whether such a study changes decisions, and if so, what the smallest procedure is that does so.
The prompt was a long ChatGPT exploration supplied by the operator. It is kept as background. It is not on the reading path: treat it as a source of candidate ideas, not as a specification. Its citations were generated with it and have not been checked.
Closes when one of two things happens. Either an instruction for running a
design-space study exists in kb/instructions/, with at least one study whose
map exposed something the existing route would have missed (a demoted
constraint, a hidden alternative, a missing axis, or an unknown that changed
the decision). Or the method is rejected with a recorded reason, such as
pilots where the map changed nothing that inverting the request and the
productive-deferral test did not already catch. Theory gained along the way
goes to kb/theory/. Then delete this directory and its entry in
kb/work/README.md.
Coupling with other workshops
- decision-lifecycle-evidence owns the question of what must survive behind an adopted architecture so later work can reconsider alternatives. A design-space map is one candidate answer to that question. Do not decide durable storage for maps here. Hand findings about retention to that workshop.
- conjecture-store is developing cheap refutable records for theory. A map's axes and compatibility rules are conjectures of that kind. Borrow from it without adopting its record type.
- windows-install, closed on 2026-11-03, holds the evidence for the
candidate retrospective pilot. Read its records from git history
(
git ls-tree 600306e1d kb/work/windows-install/); do not restore them.
Library this builds on
- First-principles analysis maps a design space before selecting within it — rests-on: the map, its falsifiability, rival maps
- Productive deferral requires a preserved option, discriminating evidence, and a convergence rule — rests-on: the posture of an open choice
- A framework rule with a boundary-preserving rival is not inherited — rests-on: demoting a supposed constraint
- Solve low-degree-of-freedom subproblems first — draws-on: commitment ordering
- Invert solution-shaped requests — draws-on: the existing framing procedure a study would extend or replace
- Design rationale management and ADR 070 — rests-on: where beliefs, selections and state live today
Sources ingested for this workshop
- Lane 1990, design spaces and rules: dimensions plus weighted rules, tested against held-out systems
- Ritchey, General Morphological Analysis: pairwise cross-consistency with typed exclusions
- ATAM: scenario evaluation, sensitivity and tradeoff points, assumption-bound judgments
- Already in the KB: set-based rework, real options, Build Systems à la Carte, Parnas 1972
Not yet ingested:
- NASA SE Handbook §6.8 (trade studies, value of information): blocked because Smart App Control refuses the
trafilaturalauncher. - ArchHypo (IEEE TSE 2025, open access): IEEE blocks scripted download, so it needs a manually downloaded PDF.
Files
- operationalizing.md — initial agent thoughts on how to test and operationalize the idea; a proposal awaiting the operator's direction, not an adopted plan
- background/ — the original exploration; background only
Conventions
Commits carry Workshop: kb/work/design-space-studies.
Complete file listing (generated at build time)