On Wed, Mar 21, 2007 at 08:43:37PM -0700, Ben Pfaff wrote: ... > And then the main loop becomes: > > moments1_create() > correlation_create() > for each case C do: > get x and y from case > moments1_add(&m1_x, x) > moments1_add(&m1_y, y) > correlations_add(x, &m1_x, y, &m1_y) > moments1_calculate() > correlation_calculate() > > This works, no? And then correlations doesn't have a dependency > on moments1 except as a user. We could even get rid of that > minor dependency by changing the interface from taking the > moments1 objects to taking the value of the mean. > > Let me know what you think.
Looks good to me. -Jason _______________________________________________ pspp-dev mailing list [email protected] http://lists.gnu.org/mailman/listinfo/pspp-dev
