Understand the failure.
Verify the repair.
Small, reproducible software failures. The broken implementation, the fix that didn’t work, and the one that passed—preserved together.
Explore the cases ↓How results are verified ↗WHAT THE ARCHIVE CONTAINS
100840 executable cases. 20168 are open.
Every case records the implementation that fails, the fix that did not work, and the repair that passed its checks—with recorded outputs and source hashes. This release adds 100840 cases across 20168 failure mechanisms and 254 domains.
The open tier gives you the failure and the unsuccessful fix for one case in every mechanism. The remaining 80672 cases, 5 variants per mechanism, are member-only: the verified repair, its recorded checks, and the full fixture suite are held in the member archive. Read the methodology ↗
A RECORD OF WHAT WENT WRONG
Browse the archive / 100840
Teaching commitment blocks the whole day · case 01
Invigilators teaching one period cannot be assigned any duty that day.
Teaching commitment blocks the whole day · case 02
Invigilators teaching one period cannot be assigned any duty that day.
Teaching commitment blocks the whole day · case 03
Invigilators teaching one period cannot be assigned any duty that day.
Teaching commitment blocks the whole day · case 04
Invigilators teaching one period cannot be assigned any duty that day.
Teaching commitment blocks the whole day · case 05
Invigilators teaching one period cannot be assigned any duty that day.
Travel time found only in the stored direction · case 01
Travelling south-to-north needs no time because only north|south is stored.
Travel time found only in the stored direction · case 02
Travelling south-to-north needs no time because only north|south is stored.
Travel time found only in the stored direction · case 03
Travelling south-to-north needs no time because only north|south is stored.
Travel time found only in the stored direction · case 04
Travelling south-to-north needs no time because only north|south is stored.
Travel time found only in the stored direction · case 05
Travelling south-to-north needs no time because only north|south is stored.
Travel gap measured between exam starts · case 01
Back-to-back sittings on different campuses pass because starts are far apart.
Travel gap measured between exam starts · case 02
Back-to-back sittings on different campuses pass because starts are far apart.
Travel gap measured between exam starts · case 03
Back-to-back sittings on different campuses pass because starts are far apart.
Travel gap measured between exam starts · case 04
Back-to-back sittings on different campuses pass because starts are far apart.
Travel gap measured between exam starts · case 05
Back-to-back sittings on different campuses pass because starts are far apart.
Overlapping sittings on one campus not reported · case 01
A student booked into two overlapping exams in the same building passes.
Overlapping sittings on one campus not reported · case 02
A student booked into two overlapping exams in the same building passes.
Overlapping sittings on one campus not reported · case 03
A student booked into two overlapping exams in the same building passes.
Overlapping sittings on one campus not reported · case 04
A student booked into two overlapping exams in the same building passes.
Overlapping sittings on one campus not reported · case 05
A student booked into two overlapping exams in the same building passes.
Consecutive sittings taken in registration order · case 01
Unordered registrations compare non-consecutive exams.
Consecutive sittings taken in registration order · case 02
Unordered registrations compare non-consecutive exams.
Consecutive sittings taken in registration order · case 03
Unordered registrations compare non-consecutive exams.
Consecutive sittings taken in registration order · case 04
Unordered registrations compare non-consecutive exams.
Consecutive sittings taken in registration order · case 05
Unordered registrations compare non-consecutive exams.
Travel shortfall reported as the full travel time · case 01
Rescheduling suggestions move exams further than necessary.
Travel shortfall reported as the full travel time · case 02
Rescheduling suggestions move exams further than necessary.
Travel shortfall reported as the full travel time · case 03
Rescheduling suggestions move exams further than necessary.
Travel shortfall reported as the full travel time · case 04
Rescheduling suggestions move exams further than necessary.
Travel shortfall reported as the full travel time · case 05
Rescheduling suggestions move exams further than necessary.
Setup buffer placed after the exam · case 01
A booking immediately after another exam passes although the room needs setup first.
Setup buffer placed after the exam · case 02
A booking immediately after another exam passes although the room needs setup first.
Setup buffer placed after the exam · case 03
A booking immediately after another exam passes although the room needs setup first.
Setup buffer placed after the exam · case 04
A booking immediately after another exam passes although the room needs setup first.
Setup buffer placed after the exam · case 05
A booking immediately after another exam passes although the room needs setup first.
Touching room bookings reported as conflicts · case 01
Bookings that exactly meet after buffers are rejected.
Touching room bookings reported as conflicts · case 02
Bookings that exactly meet after buffers are rejected.
Touching room bookings reported as conflicts · case 03
Bookings that exactly meet after buffers are rejected.
Touching room bookings reported as conflicts · case 04
Bookings that exactly meet after buffers are rejected.
Touching room bookings reported as conflicts · case 05
Bookings that exactly meet after buffers are rejected.
Overlap detected only when the second booking starts inside the first · case 01
A later-listed booking that starts earlier and overlaps is missed.
Overlap detected only when the second booking starts inside the first · case 02
A later-listed booking that starts earlier and overlaps is missed.
Overlap detected only when the second booking starts inside the first · case 03
A later-listed booking that starts earlier and overlaps is missed.
Overlap detected only when the second booking starts inside the first · case 04
A later-listed booking that starts earlier and overlaps is missed.
Overlap detected only when the second booking starts inside the first · case 05
A later-listed booking that starts earlier and overlaps is missed.
Conflict pair reported in booking order · case 01
The same conflict appears under different pair orders.
Conflict pair reported in booking order · case 02
The same conflict appears under different pair orders.
Conflict pair reported in booking order · case 03
The same conflict appears under different pair orders.
Conflict pair reported in booking order · case 04
The same conflict appears under different pair orders.
Conflict pair reported in booking order · case 05
The same conflict appears under different pair orders.
Period penalty charged once per room of a split exam · case 01
Splitting an exam across rooms multiplies its period penalty.
Period penalty charged once per room of a split exam · case 02
Splitting an exam across rooms multiplies its period penalty.
Period penalty charged once per room of a split exam · case 03
Splitting an exam across rooms multiplies its period penalty.
Period penalty charged once per room of a split exam · case 04
Splitting an exam across rooms multiplies its period penalty.
Period penalty charged once per room of a split exam · case 05
Splitting an exam across rooms multiplies its period penalty.
Room penalty charged once per room for all exams · case 01
Only the first exam using an unpopular room pays its penalty.
Room penalty charged once per room for all exams · case 02
Only the first exam using an unpopular room pays its penalty.
Room penalty charged once per room for all exams · case 03
Only the first exam using an unpopular room pays its penalty.
Room penalty charged once per room for all exams · case 04
Only the first exam using an unpopular room pays its penalty.
Room penalty charged once per room for all exams · case 05
Only the first exam using an unpopular room pays its penalty.
Repeated assignment row charges the room penalty again · case 01
Duplicate rows from a timetable merge inflate the room cost.
Repeated assignment row charges the room penalty again · case 02
Duplicate rows from a timetable merge inflate the room cost.
Repeated assignment row charges the room penalty again · case 03
Duplicate rows from a timetable merge inflate the room cost.
Repeated assignment row charges the room penalty again · case 04
Duplicate rows from a timetable merge inflate the room cost.
Repeated assignment row charges the room penalty again · case 05
Duplicate rows from a timetable merge inflate the room cost.
Shorter clashing exam sat first · case 01
The long exam starts late in the day and overruns.
Shorter clashing exam sat first · case 02
The long exam starts late in the day and overruns.
Shorter clashing exam sat first · case 03
The long exam starts late in the day and overruns.
Shorter clashing exam sat first · case 04
The long exam starts late in the day and overruns.
Shorter clashing exam sat first · case 05
The long exam starts late in the day and overruns.
Second sitting starts without the supervised break · case 01
Candidates go straight from one paper to the next.
Second sitting starts without the supervised break · case 02
Candidates go straight from one paper to the next.
Second sitting starts without the supervised break · case 03
Candidates go straight from one paper to the next.
Second sitting starts without the supervised break · case 04
Candidates go straight from one paper to the next.
Second sitting starts without the supervised break · case 05
Candidates go straight from one paper to the next.
Second sitting ending exactly at close flagged infeasible · case 01
Sittings that finish at the building closing time are rejected.
Second sitting ending exactly at close flagged infeasible · case 02
Sittings that finish at the building closing time are rejected.
Second sitting ending exactly at close flagged infeasible · case 03
Sittings that finish at the building closing time are rejected.
Second sitting ending exactly at close flagged infeasible · case 04
Sittings that finish at the building closing time are rejected.
Second sitting ending exactly at close flagged infeasible · case 05
Sittings that finish at the building closing time are rejected.
Duplicate enrolment rows inflate conflict counts · case 01
A repeated row creates a self-conflict and double counts the student's pairs.
Duplicate enrolment rows inflate conflict counts · case 02
A repeated row creates a self-conflict and double counts the student's pairs.
Duplicate enrolment rows inflate conflict counts · case 03
A repeated row creates a self-conflict and double counts the student's pairs.
Duplicate enrolment rows inflate conflict counts · case 04
A repeated row creates a self-conflict and double counts the student's pairs.
Duplicate enrolment rows inflate conflict counts · case 05
A repeated row creates a self-conflict and double counts the student's pairs.
Conflict pairs counted in both orders and with themselves · case 01
The matrix contains reversed and self pairs.
Conflict pairs counted in both orders and with themselves · case 02
The matrix contains reversed and self pairs.
Conflict pairs counted in both orders and with themselves · case 03
The matrix contains reversed and self pairs.
Conflict pairs counted in both orders and with themselves · case 04
The matrix contains reversed and self pairs.
Conflict pairs counted in both orders and with themselves · case 05
The matrix contains reversed and self pairs.
Conflict matrix stores a flag instead of a student count · case 01
Heavily shared exam pairs weigh the same as pairs with one shared student.
Conflict matrix stores a flag instead of a student count · case 02
Heavily shared exam pairs weigh the same as pairs with one shared student.
Conflict matrix stores a flag instead of a student count · case 03
Heavily shared exam pairs weigh the same as pairs with one shared student.
Conflict matrix stores a flag instead of a student count · case 04
Heavily shared exam pairs weigh the same as pairs with one shared student.
Conflict matrix stores a flag instead of a student count · case 05
Heavily shared exam pairs weigh the same as pairs with one shared student.
Online exam candidates consume hall seats · case 01
Periods with online exams appear over capacity.
Online exam candidates consume hall seats · case 02
Periods with online exams appear over capacity.
Online exam candidates consume hall seats · case 03
Periods with online exams appear over capacity.
Online exam candidates consume hall seats · case 04
Periods with online exams appear over capacity.
Online exam candidates consume hall seats · case 05
Periods with online exams appear over capacity.
INSPECTABLE BY DESIGN
Every result has a runnable source.
Runnable implementations with recorded outputs, source hashes, and explicit contracts. Related variants share a failure mechanism and belong together in evaluation splits.
Read the methodology ↗