Depending on the number of bars in a histogram.  The bars vary their width. 

Why is it not consistent?

I created a small procedure to select a random number from a list. 

pickrand:=proc(num,list)
local a,b,i:
randomize():
a:=nops(list):
for i from 1 to num do
  b||i:=rand(1..a)():
end do:
seq(list[b||i],i=1..num);
end proc:

a:=[3,6,7,8,3,6,5,7,8,9,1,2,3,4]:

pickrand(3,a);

          [8,8,5]

Now it's possible that the two 8's are in two different positions of...

I can't seem to find anything on cluster anlaysis (kmean etc...) in Maple's help.  I thought graph theory might contain something close but not really.  I haven't tried but I suppose it shouldn't be too hard to create a few procedures for such things?  Can someone provide a few examples?

Of the 3 M's Maple seems to have left this one out?

I think I know why they are the same but they shouldn't be. 

If I pointplot3d with symbol=sphere I get what I expect.  When I use symbol=circle I would expect a flat circle to be drawn.  Is this an option that was missed?  or is it that Maple just didn't carry the command over to 3d land?

If I have a sequence of points occurring at specific times.  How can I plot them on a graph so that they appear at a relative time scale? 

Let's take 4 points as an example. 

a:=[ [3,2,11:15] , [4,5,11:17] , [6,3,11:25] , [7,9,11:49] ]

How would I get these points displayed insequence at their relative times?