wilcox_R: What the package does (short line)

medpb2 <-
function(x,y=NULL,alpha=.05,nboot=2000,SEED=TRUE){
#
#   Compare 2 independent groups using medians.
#
#   A percentile bootstrap method is used, which performs well when
#   there are tied values.
#
#   The data are assumed to be stored in x and y
#
#   Missing values are automatically removed.
#
if(is.null(y)){
if(is.matrix(x) || is.data.frame(x)){
y=x[,2]
x=x[,1]
}
if(is.list(x)){
y=x[[2]]
x=x[[1]]
}
}
x=elimna(x)
y=elimna(y)
xx<-list()
xx[[1]]<-x
xx[[2]]<-y
if(SEED)set.seed(2) # set seed of random number generator so that
#             results can be duplicated.
est1=median(xx[[1]])
est2=median(xx[[2]])
est.dif<-median(xx[[1]])-median(xx[[2]])
crit<-alpha/2
temp<-round(crit*nboot)
icl<-temp+1
icu<-nboot-temp
bvec<-matrix(NA,nrow=2,ncol=nboot)
if(SEED)set.seed(2) # set seed of random number generator so that
#             results can be duplicated.
for(j in 1:2){
data<-matrix(sample(xx[[j]],size=length(xx[[j]])*nboot,replace=TRUE),nrow=nboot)
bvec[j,]<-apply(data,1,median) # Bootstrapped medians for jth group
}
top<-bvec[1,]-bvec[2,]
test<-sum(top<0)/nboot+.5*sum(top==0)/nboot
if(test > .5)test<-1-test
top<-sort(top)
ci<-NA
ci[1]<-top[icl]
ci[2]<-top[icu]
list(n1=length(x),n2=length(y),p.value=2*test,ci=ci,est1=est1,est2=est2,
est.dif=est.dif)
}