I am looking to find a distribution function (both PDF and CDF) for the distance between two points on the unit square.

Each point will be uniformly distributed on the [0,1] interval for both the x and y axes.  The distance between these points (d_ij) will of course be:

d[i, j] := sqrt((x[i] - x[j])^2 + (y[i] - y[j])^2)

I think that using a convolution may be required, but this is over my head.  If anybody can show me (preferably via a Maple worksheet), I would be very appreciative.

Thank you.