2 ms·
A long time ago I posed md5's to usenet test groups when I wanted to prove I wrote something. You have to worry about people breaking things. Raw MD5 is consi
by danielweber 14y ago
A long time ago I posed md5's to usenet test groups when I wanted to prove I wrote something.
You have to worry about people breaking things. Raw MD5 is considered useless today, since there are tools that make two documents with the same MD5.
This would not necessarily invalidate all prior MD5s; if you published like that 10 years before people starting breaking MD5, that would be pretty good evidence you really did do it, unless your idea is so amazingly valuable that it was more worthwhile to try to forge documents instead of publicizing your research.
Using a few different HMACs in parallel would probably give you good proof.
- Evbn 14y agoSpoofed MD5s are a vulnerability when you execute software blindly, but the padding is obvious if you examine the source data.
- dllthomas 14y agoThat's no reason not to use a better algorithm. Weaknesses in MD5 are already well enough understood to make it useless for many applications, and weak crypto algorithms only get weaker, as new attacks are discovered and as technology progresses to make a broader range of things feasible.