HCD.2:11 - SoTA-Echoing

The working question is: How should a practitioner compose and compare a development programme for one person’s consequential later Work without reducing it to topics or promising an unobserved effect?

The selected best-known line combines sufficiently whole complex-task practice with expertise-sensitive support, explicit correction and retesting where those mechanisms fit, and configuration-relative finite comparison. The maintained 4C/ID model and its fourth-edition presentation supply the serious complex-task design candidate: whole learning tasks, support for non-routine reasoning, timely procedural information, and additional practice for recurrent parts. Paas and van Merriënboer’s 2020 review (COG-05) supports changing guidance with prior knowledge and task complexity. Winget and Persky’s 2022 review (COG-06) supplies the bounded formative-assessment, corrective-activity, retest, and threshold arrangement. HCD.2 adapts these contributions into §§4.2–4.3; none establishes later professional performance or the effect of a local programme.

The serious ordinary alternative is a syllabus-first or branded-programme comparison: list content and hours, match them to a role profile, then prefer the most complete or familiar package. It is cheaper to describe, but it hides complete-task closure, same-holder starting evidence, contribution allocations, provider dependence, displaced work, and unequal detail. HCD.2 rejects topic count and programme labels as the comparison basis, while retaining a syllabus as one carrier of a candidate once its contribution is made explicit.

C.11.CRC supplies the internal finite-comparison grammar used in §4.4. The HCD evidence chain keeps intended instruction, supported attempts, independent performance, transfer, and retention in their direct pattern owners; HCD.2 carries only the expectations and evidence needs used by the comparison. Bastani and colleagues’ 2025 school-mathematics trial (AI-04) is counterexample evidence that improved AI-assisted practice can fail to improve later unassisted performance; its population and task do not supply an adult workplace effect size or a universal AI prohibition. These limits produce the separate evidence states in §§4.3–4.6 rather than a promised causal programme effect.

This comparison adds design and evidence work, but it is non-dominated for the stated question because it preserves complete-task usefulness, personal fit, support dependence, finite burden, and uncertainty that the cheaper syllabus-first alternative loses. Reopen the selected line if a stronger source changes the complex-task, support, correction, AI-dependence, or transfer boundary, or if representative use shows that the same decision values can be preserved with materially less effort.