[Cryptography] At what point should people not use TLS?

[Ryan Carboni]

>
> That may not indicate more than these protocols have not received enough attention.  It doesn't prove they are correct, nor does it prove that they are weak.  There is just no evidence.





> Tiger: One of the few unbroken but time-tested hash functions, designed by Anderson and Biham [5] in 1996, Tiger is sometimes recommended as an alternative to MD4-like designs like SHA-1, especially because it is faster than SHA-1 on common platforms. Tiger is in practical use e.g., in decentralized file systems, or in many file sharing protocols and applications, often in a Merkle-tree construction (also known as TigerTree [3]). The best collision attack on Tiger is on 19 rounds [31].




The question ultimately becomes, Rich, that what is easier to
cryptanalyze, a hash function, or a protocol? Furthermore, attacks on
one protocol shouldn't be difficult to apply to other protocols with
similar problems, while applying cryptanalysis is more difficult.