Sorry for self-reply, but ran out of edit time: a very valid concern about Julia is the lack of packages because it is new - a catch-22 situation that all new languages. This has been recognized by developers and Julia has now got some great inter-language interop packages and functionality, allowing it to harness a world of packages in other languages.

* Very efficient interaction with C and Fortran libraries is a core Julia feature, with the ability to pass arrays and structs directly between C and Julia.

Calling C: http://docs.julialang.org/en/latest/manual/calling-c-and-for...

* Python libraries can be used through PyCall.jl, which is used most famously to call matplotlib (PyPlot.jl) but also other libraries like pandas (Pandas.jl).

PyCall.jl: https://github.com/stevengj/PyCall.jl

PyPlot.jl: https://github.com/stevengj/PyPlot.jl

Pandas.jl: https://github.com/malmaud/Pandas.jl

* Java libraries can be used through JavaCall.jl, which is used in Taro.jl to call Apache Tika and Apache POI

JavaCall.jl: https://github.com/aviks/JavaCall.jl

Taro.jl: https://github.com/aviks/Taro.jl

* R libraries can be called through Rif.jl

Rif.jl: https://github.com/lgautier/Rif.jl