A brand-new compression algorithm based on using various irrational numbers.
Imagine an irrational number, Pi, for instance. Let's take its first digits in hexadecimal notation (for convinience). Also, imagine that we have an abstract file. Here they are:
π (in hex) | File contents (in hex) |
---|---|
3.243f6a8885a308d313198a2e03707344a40… |
3f195a3103702e30 |
Then, let's separate this file in segments (called clusters) that can be found in the Pi and encode them in such a form:
[serial number after the point] - [length of the cluster] (all numbers in decimal notation)
Separated | 3f | 19 | 5a | 31 | 0370 | 2e | 30 |
---|---|---|---|---|---|---|---|
Encoded | 3-2 | 19-2 | 10-2 | 16-2 | 25-4 | 23-2 | 12-2 |
This explaination is rough and doesn't disclose some details (particularly, how the file is separated in clusters and how the output file is generated).