The language itself can be criticized, but when academic statisticians publish a new method, they often post an R package, too, so it has recent functionality unavailable elsewhere.