Examinations & online exams.

The exam hub at /dashboard/grades organises the whole engine into four movements: set up the structure, create the assessments and papers, run the sittings and the marking, and review the results. A vital-signs strip counts the exam types configured, live online sessions, report cards published, and open resits, so the season's state is visible the moment you arrive.

A note on scope. The daily side of assessment — entering continuous scores, building the gradebook, running the continuous-assessment components — has its own handbook (the gradebook). This one is about the examination machinery: how a sitting is scheduled, invigilated, and marked; how a question bank and a paper are built; how online exams stay honest; and how every grade reaches the report card with an approval and an audit behind it.

Paper or online, one engine — and not a single grade without a record of who set it.

Before a single mark is entered, the school declares how it assesses. At /dashboard/grades/setup you define exam types — the named, weighted buckets a term's marks fall into (continuous assessment, sequences, the term exam) — create the named exam sessions (a "Midterm" or a "Final"), set the per-session policy (papers per day, same-subject spacing, cooldowns), and map the right assessment model onto each programme.

When the season opens, declare the window at /dashboard/grades/exam-periods. An exam period carries start and end dates and an optional results embargo — turn it on and report-card publishing is held until the period closes, so no results leak before the school is ready to release them all at once. The grade boundaries themselves — what counts as an A or a pass — come from the school's grading systems, which can be resolved per programme.

A clashing exam timetable is a logistical disaster, so YESS generates one. At /dashboard/grades/schedule the Auto-Generate action takes a session and applies its policy rules to produce a draft schedule, which you review for conflicts and mark complete. Each scheduled paper then needs a place to sit: at /dashboard/grades/exam-rooms you allocate rooms filtered to those with enough capacity free in the slot, set the seat count, and — if your halls mix subjects — turn on a mix policy so two subjects can share one room under controlled rules.

The result is a schedule where no student is double-booked, no room is over capacity, and every paper has a hall and a seat count before the first candidate walks in.

Good papers come from a good question bank. At /dashboard/grades/question-bank the school curates a reusable library, each question tagged with its Bloom's level (remember, understand, apply, analyse, evaluate, create), a difficulty (easy, medium, hard, expert), a type — multiple-choice, true/false, short-answer, essay, matching, fill-in-the-blank or image-based — its mark value, curriculum tags and topic, and, for an essay, a marking rubric. Everything is searchable, so a paper can be assembled from exactly the right spread of recall and reasoning.

Papers are built — and tracked — separately. At /dashboard/grades/papers/oversight a head sees every paper a session needs, which are missing, drafting, or locked, and can nudge a contributor who is late. Authoring a paper at /dashboard/grades/papers/[paperId] runs in one of three modes: a sole author writes the whole paper; split-by-topic assigns each topic to a different teacher; joint contribution has each section authored by a different hand. Questions pull from the bank by difficulty and Bloom's level, and when the paper is ready a head locks it.

The same engine runs exams on screen. At /dashboard/grades/online-exams you create an online exam, link it to an assessment component so its score routes back into the gradebook, set its window, and configure its anti-cheat posture: tab-switch detection (with a tolerance before an attempt is flagged), copy/paste detection, a required camera, a browser lock-down, IP restriction to a test centre, and randomised questions and options so no two students see the same order — alongside paced controls like one-question-per-page and a per-question timer.

During the sitting an invigilator watches a live monitor of every attempt, with suspicious activity flagged in real time and the power to disable an attempt where cheating is clear. The instant a student submits, the objective questions auto-mark and the score streams to the grade pipeline; if the connection drops mid-exam, answers are held and sent when it returns.

In the hall, the invigilator marks the register at /dashboard/grades/exam-attendance — present, arrived late, absent excused, absent unexcused, sick, a family emergency, expelled from the hall, or waived — with a bulk "mark remaining present" to finish fast. Anyone marked absent flips a makeup state to pending and flows automatically into the makeup queue at /dashboard/grades/makeups, where an admin schedules a makeup against a future sitting of the same subject, waives it, or marks it complete once sat.

Resits — repeat assessments for those who didn't pass — have their own machinery at /dashboard/grades/resits. You open a resit window with dates and a registration deadline; candidates move through eligible, registered, scheduled, sat, graded and final, with branches for no-shows, withdrawals and medical reschedules. Crucially, each programme sets how a resit score counts — capped at the pass mark, taken as-is, the best or the average of the two, a full replacement, or the original kept with an asterisk — and the rule is applied automatically when the resit is graded. A programme can even restrict eligibility to borderline scores, the way a French or Cameroonian rattrapage does.

No grade reaches a report card on a single person's say-so. When a teacher submits a component's scores, they queue for review at /dashboard/grades/submissions, where an HOD sees the grading progress and either approves — locking the scores in — or returns them for revision with a reason. Once everything is approved, results are compiled at /dashboard/grades/results: the engine weighs each component, applies the exam-type weights and the programme's structure, resolves the grading system, and writes a compiled result per student — total, weighted percentage, grade letter, rank, and promotion status — ready to publish to report cards. Year-end aggregates live at /dashboard/grades/annual-results.

After approval, scores are protected. A component is sealed at /dashboard/grades/lockout, and any later change must go through /dashboard/grades/edit-requests — a teacher requests the amendment with a justification, an HOD approves or rejects, and the original stands unless it is approved. Every one of these moves — the entry, the submission, the approval, the lock, the edit — is written to the immutable trail at /dashboard/grades/audit.