"Alan`" wrote in message <isgh38$ils$1@newscl01ah.mathworks.com>...
> so i'm making this program to find the basis for the null space of A and B where B=A tranpose
> ...........
> when i go run the program, nullB appears as empty matrix 5 by 0 so i'm trying to figure out what i can do to fix this. In my class they don't really teach us Matlab, it's a linear algebra class so we gotta do all the learning on our own. Thanks for he help in advance.
          
I have just been rereading the documentation for matlab's 'null' function, and I don't wonder that you have been confused about it, Alan. It definitely needs improvement.
For their p by n input argument, A, each of the p rows is considered a vector in R^n or C^n space. For the n by q output argument, Z, each column is considered a vector in R^n or C^n space and these are to constitute an orthonormal basis for the null space of the p rows of A. If the rows of A are linearly independent, then q = np, and otherwise q is greater than that depending on the degree of dependency among the A rows.
This is implicit in their requirements, A*Z = 0 (or, as they delicately phrased it, "negligible elements") and Z'*Z = I, but Mathworks really should have made that much more clear.
I think I have worn out my welcome in Mathworks' documentation department with my numerous complaints about various functions in the past, but if their description of 'null' did confuse you, I would advise your voicing a complaint to them about it.
Roger Stafford
