3 ms·
> We may not know how many bit flips occurred, but we can probably assume it's in the low single digits, right? So, try flipping each bit, and see if any one of
by gspr 6y ago
> We may not know how many bit flips occurred, but we can probably assume it's in the low single digits, right? So, try flipping each bit, and see if any one of them results in a very large improvement. If there are no matches, try every combination of exactly two bits, then every combination of three, and so on.
There are almost 10²⁰ ways to flip 3 bits in 1 MB. This approach seems unfeasible if any more than 1 bit has been flipped.
- faeyanpiraat 6y agoHave we found a new, and actually useful proof of work algorithm?
- Wowfunhappy 6y agoYeah, I guess I didn't do the math on that. Even the ability to correct single bit flips would be useful though. I can also imagine much faster ways to do it... for instance, since the tops of the images are fine, the problem is likely in the 8x8 block where the corruption began (and the user could point to this).