Julia: Achieving C++ Speed in High-Level Code

From my POV, as someone who has been paid to write Julia since v0.6: type stability is essential for performance, not just avoiding unnecessary allocation.

Happily, there are lots of profiling tools available and the people who write libraries actually use them, because performance is generally important to people who use Julia.

I disagree with the idea that Julia has "pythonic" syntax. (Yes, you can do list comprehension, but broadcasting is much more efficient.) To my mind Julia shows more influence from Ruby.

Julia's secret sauce is that it easily extensible, on account of its excellent support for macros.