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
S-P nomogram local magnitude: distance term weight · case 01
Distant events get far too large a correction.
S-P nomogram local magnitude: distance term weight · case 02
Distant events get far too large a correction.
S-P nomogram local magnitude: distance term weight · case 03
Distant events get far too large a correction.
S-P nomogram local magnitude: distance term weight · case 04
Distant events get far too large a correction.
S-P nomogram local magnitude: distance term weight · case 05
Distant events get far too large a correction.
S-P nomogram local magnitude: nomogram constant · case 01
All nomogram magnitudes are 5.84 units high.
S-P nomogram local magnitude: nomogram constant · case 02
All nomogram magnitudes are 5.84 units high.
S-P nomogram local magnitude: nomogram constant · case 03
All nomogram magnitudes are 5.84 units high.
S-P nomogram local magnitude: nomogram constant · case 04
All nomogram magnitudes are 5.84 units high.
S-P nomogram local magnitude: nomogram constant · case 05
All nomogram magnitudes are 5.84 units high.
Radiated energy magnitude: free-surface correction · case 01
Energy is eight times too large and Me 0.6 high.
Radiated energy magnitude: free-surface correction · case 02
Energy is eight times too large and Me 0.6 high.
Radiated energy magnitude: free-surface correction · case 03
Energy is eight times too large and Me 0.6 high.
Radiated energy magnitude: free-surface correction · case 04
Energy is eight times too large and Me 0.6 high.
Radiated energy magnitude: free-surface correction · case 05
Energy is eight times too large and Me 0.6 high.
Radiated energy magnitude: distance units · case 01
Energy is a million times too small.
Radiated energy magnitude: distance units · case 02
Energy is a million times too small.
Radiated energy magnitude: distance units · case 03
Energy is a million times too small.
Radiated energy magnitude: distance units · case 04
Energy is a million times too small.
Radiated energy magnitude: distance units · case 05
Energy is a million times too small.
Radiated energy magnitude: squared velocity · case 01
Energy grows with the square of the record length.
Radiated energy magnitude: squared velocity · case 02
Energy grows with the square of the record length.
Radiated energy magnitude: squared velocity · case 03
Energy grows with the square of the record length.
Radiated energy magnitude: squared velocity · case 04
Energy grows with the square of the record length.
Radiated energy magnitude: squared velocity · case 05
Energy grows with the square of the record length.
Radiated energy magnitude: energy magnitude slope · case 01
Energy magnitudes exceed 10 for moderate events.
Radiated energy magnitude: energy magnitude slope · case 02
Energy magnitudes exceed 10 for moderate events.
Radiated energy magnitude: energy magnitude slope · case 03
Energy magnitudes exceed 10 for moderate events.
Radiated energy magnitude: energy magnitude slope · case 04
Energy magnitudes exceed 10 for moderate events.
Radiated energy magnitude: energy magnitude slope · case 05
Energy magnitudes exceed 10 for moderate events.
Magnitude from surface rupture length: slip type coefficients · case 01
Rupture-length magnitudes are nonsensical, around 1 to 6 units off.
Magnitude from surface rupture length: slip type coefficients · case 02
Rupture-length magnitudes are nonsensical, around 1 to 6 units off.
Magnitude from surface rupture length: slip type coefficients · case 03
Rupture-length magnitudes are nonsensical, around 1 to 6 units off.
Magnitude from surface rupture length: slip type coefficients · case 04
Rupture-length magnitudes are nonsensical, around 1 to 6 units off.
Magnitude from surface rupture length: slip type coefficients · case 05
Rupture-length magnitudes are nonsensical, around 1 to 6 units off.
Magnitude from surface rupture length: logarithm base · case 01
Long ruptures are assigned magnitudes above 11.
Magnitude from surface rupture length: logarithm base · case 02
Long ruptures are assigned magnitudes above 11.
Magnitude from surface rupture length: logarithm base · case 03
Long ruptures are assigned magnitudes above 11.
Magnitude from surface rupture length: logarithm base · case 04
Long ruptures are assigned magnitudes above 11.
Magnitude from surface rupture length: logarithm base · case 05
Long ruptures are assigned magnitudes above 11.
Hamming correction flips the bit after the syndrome position · case 01
Single-bit errors are moved to the neighbouring bit instead of being repaired.
Hamming correction flips the bit after the syndrome position · case 02
Single-bit errors are moved to the neighbouring bit instead of being repaired.
Hamming correction flips the bit after the syndrome position · case 03
Single-bit errors are moved to the neighbouring bit instead of being repaired.
Hamming correction flips the bit after the syndrome position · case 04
Single-bit errors are moved to the neighbouring bit instead of being repaired.
Hamming correction flips the bit after the syndrome position · case 05
Single-bit errors are moved to the neighbouring bit instead of being repaired.
Hamming syndrome adds positions arithmetically · case 01
Clean codewords report nonzero syndromes and get corrupted.
Hamming syndrome adds positions arithmetically · case 02
Clean codewords report nonzero syndromes and get corrupted.
Hamming syndrome adds positions arithmetically · case 03
Clean codewords report nonzero syndromes and get corrupted.
Hamming syndrome adds positions arithmetically · case 04
Clean codewords report nonzero syndromes and get corrupted.
Hamming syndrome adds positions arithmetically · case 05
Clean codewords report nonzero syndromes and get corrupted.
Hamming data extraction keeps parity position four · case 01
Decoded data contains check bits and is one bit too long.
Hamming data extraction keeps parity position four · case 02
Decoded data contains check bits and is one bit too long.
Hamming data extraction keeps parity position four · case 03
Decoded data contains check bits and is one bit too long.
Hamming data extraction keeps parity position four · case 04
Decoded data contains check bits and is one bit too long.
Hamming data extraction keeps parity position four · case 05
Decoded data contains check bits and is one bit too long.
Shortened Hamming wraps an out-of-range syndrome · case 01
Double errors in a shortened code are silently "corrected" at a wrong bit.
Shortened Hamming wraps an out-of-range syndrome · case 02
Double errors in a shortened code are silently "corrected" at a wrong bit.
Shortened Hamming wraps an out-of-range syndrome · case 03
Double errors in a shortened code are silently "corrected" at a wrong bit.
Shortened Hamming wraps an out-of-range syndrome · case 04
Double errors in a shortened code are silently "corrected" at a wrong bit.
Shortened Hamming wraps an out-of-range syndrome · case 05
Double errors in a shortened code are silently "corrected" at a wrong bit.
Hamming encoder parity covers multiples of the parity position · case 01
Codewords fail to decode as clean.
Hamming encoder parity covers multiples of the parity position · case 02
Codewords fail to decode as clean.
Hamming encoder parity covers multiples of the parity position · case 03
Codewords fail to decode as clean.
Hamming encoder parity covers multiples of the parity position · case 04
Codewords fail to decode as clean.
Hamming encoder parity covers multiples of the parity position · case 05
Codewords fail to decode as clean.
Hamming encoder writes the nibble in reverse order · case 01
Decoders recover the data bits reversed.
Hamming encoder writes the nibble in reverse order · case 02
Decoders recover the data bits reversed.
Hamming encoder writes the nibble in reverse order · case 03
Decoders recover the data bits reversed.
Hamming encoder writes the nibble in reverse order · case 04
Decoders recover the data bits reversed.
Hamming encoder writes the nibble in reverse order · case 05
Decoders recover the data bits reversed.
SECDED corrects on any nonzero syndrome · case 01
Double-bit errors are "corrected" into wrong data and parity-bit errors are reported as double errors.
SECDED corrects on any nonzero syndrome · case 02
Double-bit errors are "corrected" into wrong data and parity-bit errors are reported as double errors.
SECDED corrects on any nonzero syndrome · case 03
Double-bit errors are "corrected" into wrong data and parity-bit errors are reported as double errors.
SECDED corrects on any nonzero syndrome · case 04
Double-bit errors are "corrected" into wrong data and parity-bit errors are reported as double errors.
SECDED corrects on any nonzero syndrome · case 05
Double-bit errors are "corrected" into wrong data and parity-bit errors are reported as double errors.
SECDED overall parity ignores the parity bit itself · case 01
Clean words with an odd number of data/check ones are misclassified.
SECDED overall parity ignores the parity bit itself · case 02
Clean words with an odd number of data/check ones are misclassified.
SECDED overall parity ignores the parity bit itself · case 03
Clean words with an odd number of data/check ones are misclassified.
SECDED overall parity ignores the parity bit itself · case 04
Clean words with an odd number of data/check ones are misclassified.
SECDED overall parity ignores the parity bit itself · case 05
Clean words with an odd number of data/check ones are misclassified.
SECDED flips the bit before the syndrome position · case 01
Single-bit errors are moved rather than corrected.
SECDED flips the bit before the syndrome position · case 02
Single-bit errors are moved rather than corrected.
SECDED flips the bit before the syndrome position · case 03
Single-bit errors are moved rather than corrected.
SECDED flips the bit before the syndrome position · case 04
Single-bit errors are moved rather than corrected.
SECDED flips the bit before the syndrome position · case 05
Single-bit errors are moved rather than corrected.
Repetition decoder counts erasures as zeros · case 01
Blocks with erasures are biased towards 0.
Repetition decoder counts erasures as zeros · case 02
Blocks with erasures are biased towards 0.
Repetition decoder counts erasures as zeros · case 03
Blocks with erasures are biased towards 0.
Repetition decoder counts erasures as zeros · case 04
Blocks with erasures are biased towards 0.
Repetition decoder counts erasures as zeros · case 05
Blocks with erasures are biased towards 0.
Repetition decoder resolves ties to zero · case 01
Evenly split or fully erased blocks decode to 0 instead of unknown.
Repetition decoder resolves ties to zero · case 02
Evenly split or fully erased blocks decode to 0 instead of unknown.
Repetition decoder resolves ties to zero · case 03
Evenly split or fully erased blocks decode to 0 instead of unknown.
Repetition decoder resolves ties to zero · case 04
Evenly split or fully erased blocks decode to 0 instead of unknown.
Repetition decoder resolves ties to zero · case 05
Evenly split or fully erased blocks decode to 0 instead of unknown.
Repetition decoder reads interleaved copies · case 01
Multi-bit messages decode into scrambled bits.
Repetition decoder reads interleaved copies · case 02
Multi-bit messages decode into scrambled bits.
Repetition decoder reads interleaved copies · case 03
Multi-bit messages decode into scrambled bits.
Repetition decoder reads interleaved copies · case 04
Multi-bit messages decode into scrambled bits.
Repetition decoder reads interleaved copies · case 05
Multi-bit messages decode into scrambled bits.
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 ↗