Hello. Thank you to all of you who participated on this thread. Sorry for the low interaction, but as a victim of another time-zone I had to sleep...
Apparently I have to rethink my approach on using high-level functions, which is somewhat as a guarantee in Mathematica to gain speed, which is an immediate result of the specific evaluation scheme inside Mathematica. I find this already fairly impressive what has been done so far and I'm asking myself if it would make sense to preserve this inside a number theoretical package for Julia; if something like this does not already exist. A NT package could help on decent cryptography implementations and such, similar to the Computational Number Theory package for Mathematica by Bressoud/Wagon of Malacaster college. What comes second to my mind is, if there is already a showcase implementation for the 100 digit SIAM challenge in Julia (which is probably another question for the mailing list)?
