For any who wish to talk about data compression theories, ideas, general purpose or specific.

Data compression and encryption are two topics i love, although i may not fully grok on every level.

Basic ideas for early types of compression is RLE (run length encoding). When many images early on like with Bitmap with 4-16 colors, you had a lot of repeating colors. So you might have for example...


000000
001100
010010
010010
001100
000000

Which is a small circle in a 6x6 image (36). Note there's a LOT of 0's. When you start having say 3 or more of the same color, you could then replace the next character with the length instead. So assuming they are nibbles (4 bits, or hexes) it could transform to something more compact . I'll use {} to denote the length.

000{5}11000{0}1001001001000{0}11000{5}

Reducing from 36 to 30 hexes. (reduction of 16%).

A sliding window would look back and see N spaces back and say if it repeats, course the length as to be 2 to break even, and 3+ to be worth it. Assuming no matches, going back 0 you'd have a length instead of how many are raw. The length could exceed the backtrack of the window too.

So using the same 6x6 image we'd have (assuming i did that right)
{0, raw length}
{spots back, length to copy)

{0,1}0{1,7}{0,2}11{5,3}{4,3}{3,9}{0,1}1{5,3}{1,6}

which would be 22 hexes vs the 36. (Reduction of 39%)

These are only a couple of the simpler concepts of how data could be compressed.