1.  R uses column-major storage, while C/C++ uses row-major.  So one
version of a matrix will be the transpose of the other.

2.

nextIter <- function(itrObj) {
   self <<- itrObj
   itrObj$nxt()
}

resetIter <- function(itrObj) {
   self <<- itrObj
   itrObj$rst()
}

# test example

# create the iterator object
q <- new.env()
q$a <- c(5,12,13)
q$i <- 0
q$nxt <- function() {
   self$i <- self$i + 1
   self$a[self$i]
}
q$rst <- function() {
   self$i <- 0
}

# try it
print(nextIter(q))  # 5
print(nextIter(q))  # 12
print(nextIter(q))  # 13
resetIter(q)
print(nextIter(q))  # 5

3.  

# run this ONCE
setup <- function() {
   source('http://heather.cs.ucdavis.edu/~matloff/145/SuppMaterials/Walk.R')
   download.file('http://heather.cs.ucdavis.edu/testdir.zip','timp')
   unzip('timp')
}

findLtrFile <- function(startDir,startLtr) {
   walk(startDir,lookForLtr,
      list(startLtr=startLtr,found=FALSE,foundDir=NULL,foundFile=NULL))
}

lookForLtr <- function(drname,filelist,arg) {
   if (arg$found) return(arg)
   for (fname in filelist) {
      if (substr(fname,1,1) == arg$startLtr) {
         arg$found <- TRUE
         arg$foundDir <- getwd()
         arg$foundFile <- fname
         return(arg)
      }
   }
   arg
}

setup()
startDir <- getwd()
findLtrFile(startDir, 'y')