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
Reduced moment scaled by one hundred · case 01
The reduced moment read off the loading graph is ten times too large.
Reduced moment scaled by one hundred · case 02
The reduced moment read off the loading graph is ten times too large.
Reduced moment scaled by one hundred · case 03
The reduced moment read off the loading graph is ten times too large.
Reduced moment scaled by one hundred · case 04
The reduced moment read off the loading graph is ten times too large.
Reduced moment scaled by one hundred · case 05
The reduced moment read off the loading graph is ten times too large.
CG divided from the reduced moment · case 01
The loaded CG is reported near zero inches.
CG divided from the reduced moment · case 02
The loaded CG is reported near zero inches.
CG divided from the reduced moment · case 03
The loaded CG is reported near zero inches.
CG divided from the reduced moment · case 04
The loaded CG is reported near zero inches.
CG divided from the reduced moment · case 05
The loaded CG is reported near zero inches.
Unknown units silently treated as pounds and inches · case 01
A stone or metre entry is added at face value instead of being rejected.
Unknown units silently treated as pounds and inches · case 02
A stone or metre entry is added at face value instead of being rejected.
Unknown units silently treated as pounds and inches · case 03
A stone or metre entry is added at face value instead of being rejected.
Unknown units silently treated as pounds and inches · case 04
A stone or metre entry is added at face value instead of being rejected.
Unknown units silently treated as pounds and inches · case 05
A stone or metre entry is added at face value instead of being rejected.
Floor pressure computed over the whole compartment floor · case 01
A heavy engine on a small skid passes the floor-loading check.
Floor pressure computed over the whole compartment floor · case 02
A heavy engine on a small skid passes the floor-loading check.
Floor pressure computed over the whole compartment floor · case 03
A heavy engine on a small skid passes the floor-loading check.
Floor pressure computed over the whole compartment floor · case 04
A heavy engine on a small skid passes the floor-loading check.
Floor pressure computed over the whole compartment floor · case 05
A heavy engine on a small skid passes the floor-loading check.
Compartment loaded exactly to its limit flagged over · case 01
Filling the bulk hold to exactly 800 lb raises an OVER finding.
Compartment loaded exactly to its limit flagged over · case 02
Filling the bulk hold to exactly 800 lb raises an OVER finding.
Compartment loaded exactly to its limit flagged over · case 03
Filling the bulk hold to exactly 800 lb raises an OVER finding.
Compartment loaded exactly to its limit flagged over · case 04
Filling the bulk hold to exactly 800 lb raises an OVER finding.
Compartment loaded exactly to its limit flagged over · case 05
Filling the bulk hold to exactly 800 lb raises an OVER finding.
Floor pressure uses the running compartment weight · case 01
Later pieces in a compartment are flagged for floor pressure caused by earlier pieces.
Floor pressure uses the running compartment weight · case 02
Later pieces in a compartment are flagged for floor pressure caused by earlier pieces.
Floor pressure uses the running compartment weight · case 03
Later pieces in a compartment are flagged for floor pressure caused by earlier pieces.
Floor pressure uses the running compartment weight · case 04
Later pieces in a compartment are flagged for floor pressure caused by earlier pieces.
Floor pressure uses the running compartment weight · case 05
Later pieces in a compartment are flagged for floor pressure caused by earlier pieces.
Floor-pressure finding cites a per-compartment index · case 01
Load controllers are pointed at the wrong manifest line for over-pressure pieces.
Floor-pressure finding cites a per-compartment index · case 02
Load controllers are pointed at the wrong manifest line for over-pressure pieces.
Floor-pressure finding cites a per-compartment index · case 03
Load controllers are pointed at the wrong manifest line for over-pressure pieces.
Floor-pressure finding cites a per-compartment index · case 04
Load controllers are pointed at the wrong manifest line for over-pressure pieces.
Floor-pressure finding cites a per-compartment index · case 05
Load controllers are pointed at the wrong manifest line for over-pressure pieces.
Pieces with invalid footprints counted in hold totals · case 01
A rejected zero-footprint piece still pushes its compartment into OVER.
Pieces with invalid footprints counted in hold totals · case 02
A rejected zero-footprint piece still pushes its compartment into OVER.
Pieces with invalid footprints counted in hold totals · case 03
A rejected zero-footprint piece still pushes its compartment into OVER.
Pieces with invalid footprints counted in hold totals · case 04
A rejected zero-footprint piece still pushes its compartment into OVER.
Pieces with invalid footprints counted in hold totals · case 05
A rejected zero-footprint piece still pushes its compartment into OVER.
Trim interpolation measured from the upper node · case 01
Values between nodes are mirrored: a CG just aft of a node gets nearly the next node trim.
Trim interpolation measured from the upper node · case 02
Values between nodes are mirrored: a CG just aft of a node gets nearly the next node trim.
Trim interpolation measured from the upper node · case 03
Values between nodes are mirrored: a CG just aft of a node gets nearly the next node trim.
Trim interpolation measured from the upper node · case 04
Values between nodes are mirrored: a CG just aft of a node gets nearly the next node trim.
Trim interpolation measured from the upper node · case 05
Values between nodes are mirrored: a CG just aft of a node gets nearly the next node trim.
Table end points rejected as out of range · case 01
A CG exactly at the forward or aft table limit returns ERR_RANGE.
Table end points rejected as out of range · case 02
A CG exactly at the forward or aft table limit returns ERR_RANGE.
Table end points rejected as out of range · case 03
A CG exactly at the forward or aft table limit returns ERR_RANGE.
Table end points rejected as out of range · case 04
A CG exactly at the forward or aft table limit returns ERR_RANGE.
Table end points rejected as out of range · case 05
A CG exactly at the forward or aft table limit returns ERR_RANGE.
Out-of-table CG clamped to the nearest trim value · case 01
A CG beyond the certified trim table produces a plausible-looking trim setting.
Out-of-table CG clamped to the nearest trim value · case 02
A CG beyond the certified trim table produces a plausible-looking trim setting.
Out-of-table CG clamped to the nearest trim value · case 03
A CG beyond the certified trim table produces a plausible-looking trim setting.
Out-of-table CG clamped to the nearest trim value · case 04
A CG beyond the certified trim table produces a plausible-looking trim setting.
Out-of-table CG clamped to the nearest trim value · case 05
A CG beyond the certified trim table produces a plausible-looking trim setting.
%MAC rounded to a whole percent before lookup · case 01
Trim jumps in steps and differs by up to half a percent of MAC worth of trim.
%MAC rounded to a whole percent before lookup · case 02
Trim jumps in steps and differs by up to half a percent of MAC worth of trim.
%MAC rounded to a whole percent before lookup · case 03
Trim jumps in steps and differs by up to half a percent of MAC worth of trim.
%MAC rounded to a whole percent before lookup · case 04
Trim jumps in steps and differs by up to half a percent of MAC worth of trim.
%MAC rounded to a whole percent before lookup · case 05
Trim jumps in steps and differs by up to half a percent of MAC worth of trim.
Unsupported flap setting falls back to another table · case 01
Flaps 10 returns a trim from the flaps 15 table instead of an error.
Unsupported flap setting falls back to another table · case 02
Flaps 10 returns a trim from the flaps 15 table instead of an error.
Unsupported flap setting falls back to another table · case 03
Flaps 10 returns a trim from the flaps 15 table instead of an error.
Unsupported flap setting falls back to another table · case 04
Flaps 10 returns a trim from the flaps 15 table instead of an error.
Unsupported flap setting falls back to another table · case 05
Flaps 10 returns a trim from the flaps 15 table instead of an error.
Left and right lateral moments added instead of cancelling · case 01
Symmetric left and right loads produce a large lateral CG offset.
Left and right lateral moments added instead of cancelling · case 02
Symmetric left and right loads produce a large lateral CG offset.
Left and right lateral moments added instead of cancelling · case 03
Symmetric left and right loads produce a large lateral CG offset.
Left and right lateral moments added instead of cancelling · case 04
Symmetric left and right loads produce a large lateral CG offset.
Left and right lateral moments added instead of cancelling · case 05
Symmetric left and right loads produce a large lateral CG offset.
Lateral limit enforced only to the right · case 01
A heavy left-side hoist load is reported within limits.
Lateral limit enforced only to the right · case 02
A heavy left-side hoist load is reported within limits.
Lateral limit enforced only to the right · case 03
A heavy left-side hoist load is reported within limits.
Lateral limit enforced only to the right · case 04
A heavy left-side hoist load is reported within limits.
Lateral limit enforced only to the right · case 05
A heavy left-side hoist load is reported within limits.
Simultaneous longitudinal and lateral exceedance reported as LONG only · case 01
The crew fixes the longitudinal problem and departs with the lateral CG still out of limits.
Simultaneous longitudinal and lateral exceedance reported as LONG only · case 02
The crew fixes the longitudinal problem and departs with the lateral CG still out of limits.
Simultaneous longitudinal and lateral exceedance reported as LONG only · case 03
The crew fixes the longitudinal problem and departs with the lateral CG still out of limits.
Simultaneous longitudinal and lateral exceedance reported as LONG only · case 04
The crew fixes the longitudinal problem and departs with the lateral CG still out of limits.
Simultaneous longitudinal and lateral exceedance reported as LONG only · case 05
The crew fixes the longitudinal problem and departs with the lateral CG still out of limits.
Lateral CG exactly on the limit rejected · case 01
A load that places the lateral CG exactly at the limit is reported LAT.
Lateral CG exactly on the limit rejected · case 02
A load that places the lateral CG exactly at the limit is reported LAT.
Lateral CG exactly on the limit rejected · case 03
A load that places the lateral CG exactly at the limit is reported LAT.
Lateral CG exactly on the limit rejected · case 04
A load that places the lateral CG exactly at the limit is reported LAT.
Lateral CG exactly on the limit rejected · case 05
A load that places the lateral CG exactly at the limit is reported LAT.
Lateral CG divided by off-centerline weight only · case 01
One passenger in a side seat produces a lateral CG equal to the seat arm.
Lateral CG divided by off-centerline weight only · case 02
One passenger in a side seat produces a lateral CG equal to the seat arm.
Lateral CG divided by off-centerline weight only · case 03
One passenger in a side seat produces a lateral CG equal to the seat arm.
Lateral CG divided by off-centerline weight only · case 04
One passenger in a side seat produces a lateral CG equal to the seat arm.
Lateral CG divided by off-centerline weight only · case 05
One passenger in a side seat produces a lateral CG equal to the seat arm.
Offsetting adds and removes net out of the LMC tolerance · case 01
Swapping 300 kg of cargo for 290 kg is annotated as an LMC instead of requiring a new loadsheet.
Offsetting adds and removes net out of the LMC tolerance · case 02
Swapping 300 kg of cargo for 290 kg is annotated as an LMC instead of requiring a new loadsheet.
Offsetting adds and removes net out of the LMC tolerance · case 03
Swapping 300 kg of cargo for 290 kg is annotated as an LMC instead of requiring a new loadsheet.
Offsetting adds and removes net out of the LMC tolerance · case 04
Swapping 300 kg of cargo for 290 kg is annotated as an LMC instead of requiring a new loadsheet.
Offsetting adds and removes net out of the LMC tolerance · case 05
Swapping 300 kg of cargo for 290 kg is annotated as an LMC instead of requiring a new loadsheet.
Removed items shift the index as if added · case 01
Offloading a forward bag moves the index the same way as loading it.
Removed items shift the index as if added · case 02
Offloading a forward bag moves the index the same way as loading it.
Removed items shift the index as if added · case 03
Offloading a forward bag moves the index the same way as loading it.
Removed items shift the index as if added · case 04
Offloading a forward bag moves the index the same way as loading it.
Removed items shift the index as if added · case 05
Offloading a forward bag moves the index the same way as loading it.
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 ↗