Compute K-fn of a Point Pattern
Actually computes L = sqrt(K/pi).

Syntax
	Kfn(pp, fs, k=100)
Input
pp  a list such as a pp object, including components `x' and `y'
fs  full scale of the plot
k   number of regularly spaced distances in (0, `fs')

Output
A      list with components
x      vector of distances
y      vector of L-fn values
k      number of distances returned  (may be less than `k' if `fs' is too large)
dmin   minimum distance between pair of points
lm     maximum deviation from L(t) = t

Details
relies on the domain D having been set by `ppinit' or `ppregion'.

See also
`ppinit', `ppregion', `Kaver', `Kenvl'

Example
towns <- ppinit("towns.dat")
par(pty="s")
plot(Kfn(towns, 10), type="s", xlab="distance", ylab="L(t)")