Hacker News new | past | comments | ask | show | jobs | submit login

It's inventive, but what if instead of doing it at the byte level, we did it at the bit level? Instead of splitting files on every "5" encountered, we split them at every 1 in binary, which gives about 50% of "compression".



That would essentially be a kind of run-length encoding. I doubt if a 50% compression ratio is possible, though.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: