[danmux]

I would tend to agree, based on instinct, that Go is poor for data science work (as it happens I am about to find out for real in the coming weeks :| ). To build systems to feed data into some other analysis platform: definitely, but dynamic number crunching, not so much. Rob Pike probably knows this better than most; Sawzall would probably not exist otherwise.

[codygman]

I was just reading an article describing Sawzall needing to sandbox code really makes me think of Haskell, or some other language with an effect system.

[danmux]

And this : http://www.unofficialgoogledatascience.com/2015/12/replacing... replacing Sawzall with Golang :|