
1) we have had some problems, but they are solved now
     ---> tests/C-BUG-whimed.R

2) I think its wrong in principle to only work with *integer* weights
      for 'whimed'.
      Further note, that *because* of the integer weights, we need type
      'long long' {because otherwise integer overflow kills everything}.

   --> Change this `everywhere' and make  wgt.himedian() an R function

   NOTE: There is  'weighted.median()' in limma
   	 and I have generalized it ---> inst/ex-funs.R

   But see more general code (different 'tie' strategies;
			      weighted *quantile*s)
   in /u/maechler/R/MM/STATISTICS/robust/weighted-median.R


3) Data Sets --- Valentin Todorov has several of Rousseeuw's
	     in his "rrov" package
	     (and promised me "the rest" when needed)

   I don't like the *.x, *.y  sub datasets:
   They shouldn't be needed when use a *formula*

   In his lts tests, he uses these "data sets from the literature":
   (Note that 'stackloss' is already in "datasets") :
    heart.x,heart.y,	   data(heart)
    stars.x,stars.y,	   data(stars)
    phosphor.x,phosphor.y, data(phosphor)
    stack.x,stack.loss,	   data(stackloss)
    coleman.x,coleman.y,   data(coleman)
    salinity.x,salinity.y, data(salinity)
    aircraft.x,aircraft.y, data(aircraft)
    delivery.x,delivery.y, data(delivery)
    wood.x,wood.y,	   data(wood)
    hbk.x,hbk.y,	   data(hbk)
