Or, you known, docker run -i and use streaming still :) Edit: that said, there's a ton of stuff you could work on there. Not convinced I'd abandon hadoop so qu

Or, you known, docker run -i and use streaming still :) Edit: that said, there's a ton of stuff you could work on there. Not convinced I'd abandon hadoop so quickly just to avoid wrapping my scripts in `docker run --rm -i foo ..`, but that's not the only api that's possible. But even then, not sure I'd want to tie myself so closely to docker still :)