Re: [Gretl-users] idempotence and symmetry

> Thank you very much Mr Lucchetti. Mr Cottrell also replied and suggested > another way, using > > scalar meq = (m1=m2) with values 1 if right, NA if wrong. 1 if right, 0 if wrong. But that's a test for strict equality. Note that on a digital computer matrices that "should be" equal (as in my example of a given matrix and the inverse of its inverse) will generally _not_ be exactly equal, if any non-trivial calculation is involved.