Dear guys, I'm wondering what's diff between scalding, hadoop, spark?
For scalding and hadoop,
scalding is almost pure domain logic with very little boilerplate compared to hadoop, when writing mapreduce job.
it with less infrastructure types, less configuring, more focused on the algorithm, maximize expressiveness and extensibility,
that's its advantage when compared to hadoop,
There comes my question: Q1? Except for wrapping on hadoop, scalding get improved on efficiency compared to hadoop?
Q2? what's diff between scalding and spark? which is more fit for machine learning?
thanks,
stephen