How can different distributed processing frameworks be used to process large amounts of data without having to learn the programmatic intricacies of each framework?