Most compression algorithms, like ZIP or RLE, fail on very small datasets. For example, a 52-byte dataset often ends up larger with ZIP due to headers and metadata.
FSP (Find Similar Patterns) works differently: it identifies similar blocks, stores a single base block, and records only the differences.
In a test:
Original size: 52 bytes
Compressed size: 29 bytes
Compression ratio: 1.79×
Perfect decompression — no data loss
The GitHub repository contains:
README.txt — explains the algorithm with clear examples
Python test script — shows compression, decompression, and computes compression ratio
GitHub: https://github.com/Ferki-git-creator/fsp
Website(full info): https://ferki-git-creator.github.io/fsp/
FSP works on any data: text, logs, versioned files, images, or video frames. It’s simple, universal, and the first algorithm in the world that can compress datasets smaller than 100 bytes into an even smaller size.
Comments URL: https://news.ycombinator.com/item?id=45157116
Points: 1
# Comments: 0