Programming Pedagogy for Learners without Home Computer Access: A Feasibility Synthesis

A non-trivial fraction of Vietnamese K-12 learners enter informatics instruction without a computer at home. The pedagogical question is operational, not aspirational: which configurations of unplugged activity, paper-coding, smartphonebased programming, and supervised lab access actually let such a learner reach a defensible programming skill-floor, sustain motivation across a school year, and transfer cleanly to full computer access when it becomes available? This paper synthesises the literature on CS Unplugged, paper-tracing and desk-checking pedagogy, mobile-as-IDE research, browserbased low-install platforms, tangible / off-the-page programming, equity-in-CS-education work, and the systematic-review evidence on unplugged-only as standalone pedagogy into four named pedagogical modalities feasible under the no-home-computer constraint: M1 unplugged-only; M2 paper-coding with weekly lab access; M3 mobile-only smartphone IDE; M4 hybrid schoollab plus textbook self-study. Each modality is evaluated against five feasibility criteria — skill-floor achievable, transferable-tocomputer-later guarantee, motivation sustainability, infrastructure dependency, equity — and contrasted on prerequisites, common failure modes, and motivation profile. The synthesis finds that no single modality is sufficient at all skill bands; M1 alone fails for any skill band above Level 1 of the program's SpineCluster ladder; M2 is the most robust baseline for Levels 1–2; M3 supplements rather than replaces because typing-on-glass latency caps practice depth; M4 carries the strongest transfer guarantee when the lab session exceeds two hours per week. A layered approach — M1 as on-ramp, M2 as default baseline, M3 as supplement where a household smartphone is available — is recommended for the FS2P, PP, and CPP rural deployment design. Three open questions are flagged: the exact lab-hours threshold below which transfer fails, whether AI tutors on shared low-end devices change the feasibility envelope, and the upper algorithmic-fluency ceiling reachable without daily silicon access.