Eugene, I got your point.
My intention was not to have a fast password hashing.
Skein is based on Threefish-512 with a safty factor of 2.9 vs. AES with 1.7.
So it would be intersting to have Threefish with Skein. And "fast" could be interesting for the native
Why a additional block cipher? Its for me a question of trust.
The weakness is not only depending on the speed of the hash algorithm, or the brute force resistance.
Its depending on the design of the mathematical algorithms and the potential backdoor by design.
Remember GOST and KGB.
MARS was from IBM's Cryptography Research Group, who is developing codebreaker systems like the IBM 7950 Harvest for the NSA since 1962. (check Bruce Schneier mars attacks)
Serpent was a teamwork from Cambridge University, Haifa University and Bergen. GCHQ is recruiting new employees in Cambridge. The former head of MI6 Sir Richard Dearlove is now at the University. And Haifa University is Mossad playground.
The deciscion for Rijndael was not fallen for the best security, it was fallen for the performance:
"We believe that the effort spent on evaluating the security of the five AES finalists has been
very limited, certainly compared to the 17 man-years spent by IBM on DES in the 1970s.
It is noteworthy that the majority of papers at AES3 dealt with performance evaluation
rather than with security evaluation. ..."
So, why not think about an alternative highly secure NIST free algorithm and password hashing?
It is only a proposal, or even a wish.