The course-wide rules for access, workload, evidence, attribution and submission that every assessment brief assumes but doesn't repeat.

Policies and support#

These rules apply across all three assessments — Jump Autopsy, Take Control, and One Body, Three Personalities — so each brief can link here instead of restating them.

Access#

Every core task in this course has a non-coding route: the Jump Lab’s sliders, presets, trial log, “Copy configuration” link, and A/B comparison are all built to be operated without writing a line of code, and every assessment can be completed entirely through them. Optional, guided starter-code modifications appear later in the course for students who want them, but no assessment criterion requires code, and none rewards code complexity for its own sake. If a control, label, or piece of feedback is not usable with a keyboard, not readable at your device’s default zoom, or depends on sound to convey information no visual channel also carries, that’s a bug in the platform, not a gap in your submission — report it to your tutor and you will not be penalised for a measurement you couldn’t take because of it.

Workload#

Each session is designed to rehearse, at low stakes, exactly the format its nearest assessment is marked on — attending sessions is the most direct preparation available for the assessment attached to that phase of the course. Session outputs are described in each session page’s spec but are not marked; they exist so the first time you produce a piece of required evidence isn’t also the first time it’s due. Lecture tasks are short enough to complete before the following session and are the input that session’s activities build on.

Evidence#

Every number in a submission must trace back to an actual trial run in the Jump Lab — a value you remember, estimate, or reconstruct from a screenshot taken before you started editing is not acceptable evidence, even if it’s correct. Where a brief asks for a reproducible method, that means naming the exact configuration (preset name, or a “Copy configuration” link, or both), the mode, and the number of trials, in enough detail that a marker could run the same trials and land within your reported range. Where a brief asks for a playtest observation, that means a concrete, written account of what a real person did, not a general impression of how the session went.

Attribution#

Work submitted for this course must be your own measurements, your own written explanations, and your own observations of anyone you playtested with — but the playtesting itself is expected to involve another person, and you should name who you tested with (a first name or initials is enough) in the relevant section of your submission. If you discussed an approach with a classmate before running your own trials, say so briefly; it does not cost you marks and keeps the record honest. Do not present a classmate’s trial data, configuration link, or written observation as your own.

Submission#

Submit through the same channel for all three assessments, as a single document containing your tables, links and written responses in the order each brief lists its criteria — this makes marking against those criteria straightforward and is the fastest way to make sure nothing you produced gets missed. Configuration links and any other URLs should be pasted in full, not shortened, so they can be checked directly. Late submissions and extension requests are handled individually by your tutor; raise them as early as you can, ideally before the deadline rather than after it.