[emmelaich]

This is off topic but why do tools built for large scale analytics or distributed processing have to be so slow at small scale / at the low end.

e.g. R, Erlang, Hadoop.

Feel free to ignore me but please don't downvote without comment :-)

[omni]

Slightly OT, but I wouldn't say that R was "built for large scale analytics." You have to go through some serious hoops to work on anything in R that takes up more than 50% or so of available RAM. Revolution Analytics's R build is designed for it, and some standalone packages like bigmemory can make R work better with large data sets, but vanilla R is awful for the task.