| CC-E15-1 (Exact change basis). | A conforming use MUST identify the exact predecessor and candidate editions, the change question, and the receiving use. | Makes continuity replayable without a universal context record. |
| CC-E15-2 (Actual delta first). | Delta-Class MUST follow comparison of actual effects on recognition, inputs, actions, conditions, results, norms, ontology, consumers, and assurance; a label, version number, edit operator, or line count MUST NOT decide it. | Prevents small semantic changes and large harmless rewrites from being misclassified. |
| CC-E15-3 (Proportional route). | A direct understood repair MUST NOT be forced through multiple-candidate generation, aggregate scoring, a SoTA harvest, or a new trace record. Alternative comparison MUST be used only when materially plausible alternatives or a consequential current-source choice remain. | Keeps ordinary changes cheap without deleting the stronger branch. |
| CC-E15-4 (Actual affected reach). | Every in-scope locus whose action, interpretation, condition, result, check, or public reference depends on the changed premise MUST be repaired or explicitly unresolved. Mentions and declared links MUST NOT substitute for the actual dependency test. | Makes the repair dependency-closed. |
| CC-E15-5 (Independent predecessor preservation). | A material rewrite MUST be checked against an independently derived inventory of predecessor functions and uses. Every intentional change, move, or retirement MUST be named by its successor decision; an author-authored preservation map alone MUST NOT close completeness. | Prevents compression or restructuring from silently deleting ideas. |
| CC-E15-6 (Valid selection). | Hard constraints MUST remain separate from multi-scale qualities. When alternatives are compared, the selected candidate MUST be non-dominated for the stated intended use and protected trade-offs; an undeclared aggregate score MUST NOT choose it. | Makes “better” a replayable use-relative claim. |
| CC-E15-7 (Record economy and object separation). | The change account MUST reuse the receiving workflow’s governed records by reference. A new episteme or publication MUST NOT be inferred solely from a form, file, self-log, or the fact that authoring occurred. | Avoids duplicate records and false work evidence. |
| CC-E15-8 (Proportionate verification). | Checks MUST be selected from the actual change and claimed conclusion. After material formal or ontological repair, the whole changed pattern MUST receive a precise-plain-language and practitioner-use read in addition to focused semantic checks. | Prevents locally exact repairs from making the pattern unusable. |
| CC-E15-9 (Edition continuity). | Historical editions MUST remain recoverable, and only affected premises or consumers MUST reopen. Publication, acceptance, registration, and landing MUST NOT be inferred from E.15 completion. | Separates semantic continuity from later lifecycle decisions. |
| CC-E15-10 (Problem-specific SoTA). | When external or internal SoTA changes the repair, the account MUST compare serious current alternatives at comparable application effort and state the adopted, adapted, or rejected contribution at the receiving locus. Source popularity, age, official status, or tool availability MUST NOT substitute for that decision. | Keeps research load-bearing and bounded. |