A couple of little things recently done in CVS that may be
handy for script writers.
1) You can now give an "anonymous" list as an argument to a
user-defined function, as in
retval = foo(logs(X))
for X a named list, or
retval = foo($xlist)
where $xlist gets the list of regressors from the last model.
Up till now only named lists were accepted as function
arguments (and you'd get a misleading error message if you
tried giving an anonymous list).
2) You can now "dot-assign" a scalar value to a pre-existing
matrix, as in
matrix m = ones(3,2)
m .= 0
The second line sets all elements of m to zero, preserving the
original dimensions of m. Up till now there was no direct way
to do this. You could say
m = 0
but that would give you a 1x1 zero matrix.
Show replies by date