Sample

Dorn/cuda/98

Human rating uses the original dataset scale.

DatasetDorn
Human2.01
Lines10

Scores

MethodScoreScaleDataset rank
(0–100)
Posnett00–1; higher is more readable
25.9Δ 22.0
Scalabrino0.0450–1; higher is more readable
2.3Δ 1.6
Dorn0.0790–1; higher is more readable
6.4Δ 2.5
Mi0.1990–1; higher is more readable
20.6Δ 16.7
LLOC4Logical lines; fewer is more readable
82.7Δ 78.8
DeepSeek V4 Pro2.670–20; mean of three runs
2.1Δ 1.8
GPT-6.1 Sol5.330–20; mean of three runs
0.1Δ 3.8
BTTF0.2580–1; higher is more readable
11.4Δ 7.5
Human2.011–5; original rating
3.9

Views

Download official ZIP ↗

Select the downloaded ZIP or its extracted folder. Files stay in your browser.

Loading feature diagnosis…

Source regions

Feature contributions

Feature contributions

FeatureValueContribution
(log-odds)
Halstead volumeHow much symbolic information the code contains. It combines the total number of operators and values with the number of distinct ones.929.177−30.663
Byte entropyHow varied the bytes in the source text are. Repeated bytes give a lower value; a more even mixture gives a higher value.4.872−7.309
Source linesThe number of non-empty source lines.10+4

Intercept 8.87 + feature contributions = -25.102 (log-odds).
Score = sigmoid(-25.102) = 0.

Score0.045

Feature contributions

FeatureValueContribution
(log-odds)
Operator visual areaThe share of the code’s visual area occupied by operators such as +, =, and *.0.203−1.03
Identifier layout (horizontal)How rapidly the visual pattern of identifier characters changes from left to right. It measures spatial frequency, not the meaning of the names.107.091−0.871
Comment layout (horizontal)How rapidly the visual pattern of comment characters changes from left to right.0−0.804
Comment layout frequencyHow frequently the layout switches between lines containing only comments or whitespace and lines containing code.11+0.647
Indentation frequencyHow rapidly indentation changes from line to line.10−0.391
Identifier layout (vertical)How rapidly the visual pattern of identifier characters changes from top to bottom.9.827−0.272
Number layout frequencyHow rapidly the number of numeric values per line changes through the code.10+0.191

Intercept 0.07 + feature contributions = -2.46 (log-odds).
Score = sigmoid(-2.46) = 0.079.

Score0.199
Score4
Run 1

The code is severely unreadable due to multiple critical issues. It appears to be a broken macro fragment missing its opening, with a dangling line before the first statement. The variable names (b0-b15, p0-p15) are completely non-descriptive, giving zero indication of their purpose. The macro uses token-pasting (##length##) and references undefined variables like 'pass_length', 'sharedBitmap', and 'DEVICE_HashTable' without context. Multiple operations are crammed onto single lines with no separation, and the function call has 23 parameters, making it impossible to track arguments. The closing braces suggest this is a fragment of a larger macro definition, but without context it's incomprehensible. Line continuations with backslashes add visual noise. While the code follows a consistent pattern (transform, check, increment), the lack of meaningful names, extreme parameter counts, and incomplete context make it nearly unreadable.

Run 2

This code fragment is extremely unreadable. It appears to be a macro definition using token pasting (##length##), yet the macro opening is missing from the snippet. The structure is tangled: it mixes variable declarations, function-like macro calls (SHA_TRANSFORM_SMALL, checkHashMultiSHA1, incrementCounters##length##Multi), and counter increments with no separation of concerns. Nesting is not deep, but the dense, unbroken sequence of positional arguments (b0 through b15, p0 through p15) obscures meaning. Clarity of intent is very low—readers cannot discern purpose without surrounding context. Code length is excessive due to repeated variable lists, though this may be macro-generated. Action granularity is poor: many parameters and operations are compressed onto single lines, with little logical grouping. Reading flow is disorganized, jumping from transformation to checksum to counter increment without natural progression. Overall, this code is cryptic, context-dependent, and fails nearly all readability criteria.

Run 3

This code is extremely difficult to read and understand. It appears to be a macro fragment with severe formatting issues, including a dangling closing brace on the first line and excessive line continuation characters. The variable names (b0-b15, p0-p15, a-e) provide no semantic meaning, making intent completely opaque. While the code follows a somewhat linear structure, the massive function calls with 20+ parameters create extreme cognitive load. The macro concatenation (##length##) adds another layer of indirection. The line continuation backslashes make it visually noisy. Although each major operation appears to be on separate lines, the parameters themselves are unreadable. This looks like auto-generated or heavily optimized cryptographic code that prioritizes performance over human readability.

Run 1

Code Structure: The fragment begins mid-statement and ends with unmatched contextual braces, obscuring its organization. Nesting: Little nesting is visible, but missing surrounding code makes scope difficult to determine. Clarity of intent: SHA transformation, hash checking, and counter updates suggest a password-hashing loop, but cryptic variable names and token-pasted macros conceal details. Code length: The operational sequence is short, yet extensive positional argument lists make it visually cumbersome. Action granularity: Most statements perform one operation, although the macros may hide additional actions. Reading flow: The sequence of transformation, checking, and incrementing is logical, but inconsistent indentation, continuation backslashes, and the truncated opening significantly disrupt reading.

Run 2

Code Structure: The fragment appears to follow a hashing, checking, and counter-update sequence, but missing surrounding context and macro continuation syntax obscure its organization. Nesting: Visible nesting is limited, although unmatched closing braces prevent assessment of the full structure. Clarity of intent: SHA_TRANSFORM_SMALL and checkHashMultiSHA1 suggest SHA-1 password checking, but terse variable names and extensive positional arguments make data roles unclear. Code length: Few operations are shown, yet repetitive argument lists create substantial visual clutter. Action granularity: Most statements perform one action each, which helps readability. Reading flow: Inconsistent indentation, long lines, token-pasting syntax, and an opening partial statement interrupt otherwise sequential execution.

Run 3

Code Structure: The fragment shows a sequential hashing workflow, but missing surrounding context, macro continuations, and unmatched visible boundaries obscure its organization. Nesting: Little nesting is visible, although closing braces imply an enclosing structure that cannot be assessed. Clarity of intent: SHA transformation, hash checking, and counter updates suggest the purpose, but terse variable names, token-pasted identifiers, and large positional argument lists make details difficult to understand. Code length: Few statements keep it short, but repetitive argument lists create substantial visual clutter. Action granularity: Each visible statement generally performs one action, though macro expansions may conceal additional operations. Reading flow: Inconsistent indentation, long lines, continuation backslashes, and the incomplete opening statement disrupt otherwise sequential execution.