Knuth Fisher-Yates shuffle for Go (golang)

knuth.go

// implements Knuth or Fisher-Yates shuffle
package knuth

import (
    "math/rand"
    "time"
)

func init() {
    rand.Seed(time.Now().UTC().UnixNano())
}

func Shuffle(slc []interface{}) {
    N := len(slc)
    for i := 0; i < N; i++ {
        // choose index uniformly in [i, N-1]
        r := i + rand.Intn(N-i)
        slc[r], slc[i] = slc[i], slc[r]
    }
}

func ShuffleInts(slc []int) {
    N := len(slc)
    for i := 0; i < N; i++ {
        // choose index uniformly in [i, N-1]
        r := i + rand.Intn(N-i)
        slc[r], slc[i] = slc[i], slc[r]
    }
}

func ShuffleStrings(slc []string) {
    N := len(slc)
    for i := 0; i < N; i++ {
        // choose index uniformly in [i, N-1]
        r := i + rand.Intn(N-i)
        slc[r], slc[i] = slc[i], slc[r]
    }
}

I think this has a flaw as it doesn't match the pseudocode:

To shuffle an array a of n elements (indices 0..n-1):
  for i from n − 1 downto 1 do
       j ← random integer with 0 ≤ j ≤ i
       exchange a[j] and a[i]

Try this:

rand.Seed(time.Now().UnixNano())
n := len(slice)
for i := n - 1; i > 0; i-- {
    j := rand.Intn(i + 1)
    slice[i], slice[j] = slice[j], slice[i]
}

You were right - there was a flaw in the original implementation. Your version looks correct. There is also a way to do it by counting up instead of down. I based my revised version on Bob Sedgewick's version: http://algs4.cs.princeton.edu/11model/Knuth.java.html

rand.Seed should not be called in the package.
If it has to be done, it must be done just once per program in the main package. Because if every package that uses rand calls Seed, it may be called many times, while once is enough.