A Repeatable Test for an Emulated Computer

Filed under: Workshop Notes
A familiar startup screen is a useful milestone. For a workshop about emulation, the next milestone is more specific: a known input produces a response the group has described in advance.
The October 26, 2026 Raspberry Pi Apple II tutorial introduces LinApple and includes a short BASIC example that moves a square with keyboard input. Its worked example offers a starting point for examining behaviour. The installation instructions and software choices belong to that dated tutorial; the exercise here concerns how to frame an observation.
Write an expectation you can inspect
Choose one small behaviour from a documented program the group is entitled to use. A movement example is convenient because it exposes position, direction and the difference between a press and a held key.
Record the starting position and the input sequence before running it. For example: press the movement key once, release it, then press the opposite movement key once. Predict the return position relative to the recorded starting position, and note how you will recognise a difference on the displayed image. Also write the assumption behind that prediction: each press produces the same number of movement steps in its respective direction.
That assumption is part of the test. If a key remains held long enough to repeat, the two inputs may no longer represent matching steps. A mismatch is then a reason to inspect input handling, rather than immediate evidence of a fault in the emulated processor.
Trace the boundary before blaming the emulator
The tutorial’s printed BASIC listing gives a concrete paper check. It starts at X = 140 with a movement step of 3. D adds 3 to X; A subtracts 3. The listing then clamps X between 0 and 272.
| Starting X | First key and resulting X | Opposite key and resulting X |
|---|---|---|
| 140 | D → 143 | A → 140 |
| 0, at the left edge | A → 0 after clamping | D → 3 |
The interior pair returns to its start; the edge pair does not. That difference follows from the program’s boundary rule, even when each key produces exactly one step. Check position limits as well as repeated input before attributing a mismatch to emulation. This is a static walkthrough of the published listing; no emulator was run for this article.
Keep a short observation record
- Program: identify the exact example and any changes you made.
- Setup: record the emulator version, selected machine configuration and relevant input settings.
- Action: describe presses, releases and any key that remained held.
- Response: state what changed on screen, including a result that differed from the prediction.
- Interpretation: identify one explanation to investigate next.
Keep the observation and interpretation in separate sentences. “The square moved twice” is a report of a visible response. “The emulator repeated a key” is a proposed account of why that response occurred. The second sentence needs another check before it becomes a conclusion.
A screen recording can help the group revisit the sequence, provided it captures the information needed for the question. A recording of the display alone may leave the actual key actions uncertain. Make that limitation part of the record.
Broaden one question at a time
After a repeatable result, change one relevant condition. You might compare a short press with a held key while preserving the program and configuration. Decide which visible difference would answer the new question before collecting another example.
Several successful checks supply evidence about those named behaviours under the recorded conditions. They still leave many other behaviours unexamined: another program, a different peripheral or a different display path could pose a new question. A workshop does not need to settle all of them to produce a useful result.
Finish with a test description another group could repeat. That gives the familiar screen an evidential companion: an account of what the setup actually did under named conditions.
This article provides a proposed observation method, not a new emulator benchmark or an installation guide.