| Discoverability | Authors need a recognizable framework question, but a locator must not become another decision object. |
| Framework scale | A narrow pattern set can be useful, but a broad public framework needs connected problem-family coverage, a first use that does not depend on unpublished authoring context, and a stated reason for later refresh rather than a pattern count. |
| Several structures | A decision needs one coherent practice-architecture answer, but Method, Work, subject, description, capability, provider, and cultural structures need not line up one-for-one. |
| Decision memory | Later work needs rationale and consequences, but the DRR is not the accepted answer, performed authoring, or framework edition. |
| Framework detail | Edition, dependency, pattern placement, relations, omissions, the sources that later authors must be able to revisit, and publication consequences matter, but unrelated quality, naming, and package apparatus must stay conditional. |
| Cheap exit | A suitable non-framework product, small access result, or existing-framework contribution may solve the immediate problem without a framework decision. |
| Relation precision | Initial pattern relations may shape the architecture, but a row or schema does not make those relations obtain. |
| Evolution | The answer needs a reopen condition without turning every refresh concern into a mandatory field. |
| Attainable benefit | A useful method can remain unusable because its audience cannot find, understand, learn or enact it; more explanation can also burden an already prepared reader. |
| Whole cost and distribution | Repeated use can repay learning and authoring, while costs, gains and the ability to pay can fall on different participants. |
| Sharing and independence | Shared methods can reduce duplicate work while accumulating exceptions and propagating changes into otherwise independent uses. |