There seems to be an error in Veech group computations reported by Andrea Thevis (Saarbrücken). In the example below, there is a weird -1 twist. ``` sage: from surface_dynamics import * sage: o = Origami("(1,6)(2,5,4)","(1,3,2,4,5)") sage: g = o.veech_group() sage: l = g.L() sage: r = g.R() sage: s2 = g.S2() sage: s3 = g.S3() sage: m = s2**2 # minus identity sage: i = m**2 # identity sage: m == s3**3 # OK True sage: (s2*l)**3 == i # WRONG: should be m! False sage: (s2*r)**3 == m # WRONG: should be i! True ```