Why doesn't this program work?

% Weird1:
% Demonstrates some bizaar behavior of MatLAB
% Apparently scalar multiplication flips a vector from row to column.
% Or it is against the rules to put a variable on both sides of the 
% equal sign; for example X(i) = 2*X(i)
% John Kirk email [e-mail]
clear all; clc; close all;

rpm = zeros(1000,1); % Now create all the data      
for ii=1:1000;
    if mod(ii,2) == 0; % is it even?
        rpm(ii)=ii;
    else
        rpm(ii)=0;
    end 
end  %rpm should now be vector with every other element zero

%Remove zero points for ii=2:(n-1);
count(1)=1;
count(1000)=1000;

for ii=2:(1000-1);
    count(ii) = count(ii)+1;

     if rpm(ii)==0;% if rpm(ii) is zero then
         rpm(ii)=(rpm(ii-1)+rpm(ii+1))/2 % use average of non-zero points
     else
      % if it is not zero then do nothing to rpm(ii)   
     end
     % shouldn't this statement be executed for every loop?
     rpm(ii)= rpm(ii)*2;% multiply every element in rpm by 2

     end% of for loop

   plot(count,rpm)

%  END of program *******************************************
%   We have several theories for why this program does not work.
%   One is that referring to rpm(ii) on both sides of the '='
%   sign confuses MatLAB.  The other, and almost certainly true,
%  idea is we do not understand how MatLAB handles vectors and
%  matrix's.   We imagine that we are simply multiplying a
% scalar time each element in a vector when in fact we are
% doing matrix math and flipping a row vector into a column
% vector.

[Jan]

I cannot see any bizarre behavior. X(i)=X(i)*2 does not change the shape of X.

Answer 1

[Matt Fig]

I don't see the problem?? The code appears to behave exactly how it is programmed. The one thing I think stands out is the comment on this line:

rpm(ii)= rpm(ii)*2% multiply every element in rpm by 2

This line does not do what the comment says it does. If you want to multiply every element of rpm by 2, you do:

rpm = rpm*2;  % It's called 'scalar expansion' - a good thing!

What the first bit of code does is multiply only the element at ii by 2. Thus it the comment should read:

% multiply element ii by 2

Is that your concern, or is there something else? When you say the program "doesn't work" I think you mean it "doesn't do what I expect it to do." But there is a big difference between those statements! So what do you expect it to do? (BTW, I think nothing is substantially changed by replacing each instance of 1000 by 10 so that we can keep better track of things....)

Answer 2

[owr]

The only row vector vs. column vector issues I see here (I ran the code) is between "rpm" and "count".

rpm is defined to be a column vector with this:

rpm = zeros(1000,1);

and remains a column vector throughout all your computations.

count in indirectly defined to be a row vector - indirectly because you never initialize it with anything but this:

count(1)=1;
count(1000)=1000;

When you introduce a variable like this MATLAB automatically creates it as a row vector with exactly as many elements as it needs. The first line creates a 1 x 1 row vector, the second stretches it out to a 1 x 1000 element row vector with zeros in positions 2,...,999

[Matt Fig]

Yet this is not a problem or a conflict in any way. In the code, count and rpm do not interact... If count were defined as

count = zeros(1000,1);

before the assigning of the values at 1 and 1000, nothing else about the code would be different whatsoever.