Agreed. I usually end up doing the opposite in node, whenever there is an existing *nix tool for a job, I'll spawn that from my node process rather than reimple

Agreed. I usually end up doing the opposite in node, whenever there is an existing *nix tool for a job, I'll spawn that from my node process rather than reimplementing it. Node's child process machinery is the best I've seen so far. (Disclaimer: I'm a node contributor)