fahree
2017-3-4 02:59:49

Will there be a systematic move away from md5 and sha1 in the racket community? Maybe even issuing warnings or errors when the functions are used?


lexi.lambda
2017-3-4 03:03:40

There are valid applications of MD5 and SHA–1 that are not cryptographic and for which collisions are not a problem.


fahree
2017-3-4 03:10:46

won’t a simple CRC do better then?


fahree
2017-3-4 03:11:06

without the false sense of security of broken crypto?


fahree
2017-3-4 03:11:39

at the very least, md5 and sha1 functions should be moved to a package “broken-crypto” or something, so whoever uses it knows what he’s getting into.


lexi.lambda
2017-3-4 03:12:45

I would strongly disagree, though a note in the documentation that they shouldn’t be used for crypto is a good idea. Hash functions have lots of very valid uses beyond crypto or integrity checking, though.