Universal prefix code for non-negative integers. Recursively encodes the bit-length of the bit-length (an iterated-logarithm chain) terminated by zero, so arbitrarily large integers can be represented with a self-delimiting code.
| Property | Value |
|---|---|
| Category | Compression Algorithms |
| Sub-category | Universal Codes |
| Security status | Not classified |
| Complexity | Intermediate |
| Inventor | Vladimir I. Levenshtein |
| Year | 1968 |
| Origin | 🇷🇺 Russia |
| Source | algorithms/compression/levenshtein-coding.js |
Status: not classified — treat as unverified.
No vulnerabilities are recorded for this implementation.
4 vectors ship with this algorithm and run in the test suite. Byte values are hexadecimal.
Vector 1 — Empty input
| Field | Value |
|---|---|
input |
(empty) |
expected |
00000000 |
Vector 2 — Single small value - byte 1 (encodes value 2, code ‘1100’)
| Field | Value |
|---|---|
input |
01 |
expected |
01000000c0 |
Vector 3 — Canonical worked examples - byte values 0,1,2,3,7,8
| Field | Value |
|---|---|
input |
000102030708 |
expected |
06000000b378747480 |
Vector 4 — Text sample - ‘AB3’
| Field | Value |
|---|---|
input |
414233 |
expected |
03000000f20bc83f1a00 |