Design memo: why we simulate apps instead of using real ones

Design memos document real decisions in Wiz Kids and the research behind them.

The most common first question from technically minded adults: why build a pretend browser, inbox, file explorer, spreadsheet, terminal and video editor when real ones are right there, free? It's a fair question — the simulations were most of our engineering. Here's the full reasoning.

1. Assessment: a simulation can verify the doing

Our deepest commitment is that skills are demonstrated by performance, not quizzes — and performance assessment needs the environment to know what happened. Real applications are black boxes: no classroom tool can check that a child's reply-all went to the right people in Gmail, that page 2 (only) reached the printer, or that a quoted search was used. In our simulations, every task is outcome-verified — the file arrived attached, the total uses SUM, only necessary cookies were accepted — with pace and efficiency observable, mastery checkable, and reviews enforceable. The simulation is the assessment instrument.

2. Safety: consequence-free by construction

Children practice exactly the things that are dangerous to practice live: replying to phishing, clicking scam popups, mishandling consent gates, group-chat pile-ons. In simulation, the phishing owl is authored, the "214 partners" popup shares nothing, the chat has no real children in it — so a wrong call costs a retry and a coaching line instead of anything real. This is standard reasoning in every high-stakes training field (nobody learns emergencies in a real aircraft), and the zero-PII stance rides on it too: no real email accounts, no real messaging, no channel to moderate.

3. Pedagogy: instruction lives inside the interface

Cognitive load research is blunt about split attention: instructions separated from the interface they describe tax working memory just by being elsewhere. Real apps can't host our instruction cards, coach at the moment of error, hide chords in recall mode, pause mid-run to ask for input, or slow a drag-payload down so a child sees what's happening. Simulations let the teaching live in the tool — and let us remove the teaching as expertise grows, which the fading literature requires.

4. Determinism: same lesson for every child, forever

Real apps A/B-test, redesign, and inject "new!" popups weekly; a lesson written against today's Gmail is stale by term's end, and two children may literally see different interfaces the same morning. Curriculum requires the environment to hold still. Ours holds still.

The transfer question (the honest cost)

The real objection isn't any of the above — it's transfer: does skill in a simulated inbox move to a real one? The transfer literature's consistent answer: near-transfer tracks structural similarity, so we simulate the genre conventions every real app shares — To/Subject/attach, address bars and tabs, cells and formulas, trim handles — rather than any vendor's pixels. The concepts (reply-all goes to everyone; a formula recalculates; quotes narrow a search) are the durable layer; vendors reskin surfaces, not semantics. Two honest limits remain: children still need a few sessions of your school's actual tools before real-world use (we say so in teacher materials — the last mile is local), and simulation fidelity is permanent engineering upkeep on our side. We also deliberately don't simulate what shouldn't be safe-feeling — reserved OS shortcuts that no web page can intercept are taught as scenarios instead, because a simulation that lies about what keys do would train the wrong reflex.

The line we won't cross

Simulation is for practice, not for walling children off from reality. The curriculum's stated endpoint is confident use of real tools — the simulations are training wheels engineered to come off, and every realm's fiction points outward ("this is how the fastest people you'll ever meet work"). A product that kept children inside its garden forever would have optimized for itself, not for them.

References

• Sweller, J., et al. — split-attention and worked-example effects, Educational Psychology Review (1998; 2019 update).
• Messick, S. (1995). Validity of psychological assessment. American Psychologist, 50(9) — the assessment-instrument argument.
• Fraillon, J., et al. (2019). ICILS 2018 — precedent: the serious international assessment of these skills is itself conducted in purpose-built simulated environments, for the same reasons.

© Glu IO Pty. Ltd. — Wiz Kids (wiz.kids). Link freely; republication requires permission — see terms. Found an error in our reading of the research? We correct fast: tell any teacher piloting Wiz Kids.