I think of Standard ML as a kind-of "Scheme for production", so there's that. Beyond that, there are languages better for some domains, like APL for math.. And I like D for low-level bit banging, but this kind of work is intrinsically slow so best kept to a minimum.