wilcox_R: What the package does (short line)

sm2strv7 <-
function(xx,y,iv=c(1,2),nboot=100,SEED=TRUE,xout=FALSE,outfun=outpro,
STAND=TRUE,...){
#
# Compare robust measures of association of two predictors
# based on a smooth
#
# x is a matrix with two columns
# robust explanatory of  x[,1] with y is compared to x[,2] with y.
# xout=T eliminates any leverage points found with outfun, which
# defaults to outpro, a projecion method for detecting outliers.
#
#  iv: indicates the two columns of x that will be used. By default, col 1 and 2 are used.
#
if(!is.matrix(xx))stop("x should be a matrix with 2 or more columns")
if(ncol(xx)<2)stop("x should be a matrix with 2 or more columns")
val1=NA
val2=NA
x=xx[,iv]
xy=elimna(cbind(x,y))
x=xy[,1:2]
y=xy[,3]
if(xout){
x<-as.matrix(x)
flag<-outfun(x,...)$keep
x<-x[flag,]
y<-y[flag]
x<-as.matrix(x)
}
if(SEED)set.seed(2)
data1<-matrix(sample(length(y),size=length(y)*nboot,replace=TRUE),nrow=nboot)
data2<-matrix(sample(length(y),size=length(y)*nboot,replace=TRUE),nrow=nboot)
bvec1=apply(data1,1,sm2str.sub,x[,1],y) #  2 by nboot matrix
bvec2=apply(data2,1,sm2str.sub,x[,2],y) #  2 by nboot matrix
bvecd=bvec1-bvec2
pv=akerdcdf(bvecd,pts=0)
vcor=cor(x,method="kendall")
pv=2*min(c(pv,1-pv))
p.crit=.25*abs(vcor[1,2])+.05+(100-length(y))/10000
p.crit=max(c(.05,p.crit))
list(p.value=pv,p.crit=p.crit)
}