On Thu, 17 Nov 2005, Alan Bromborsky wrote:
You probably handle this correctly internally, but the documentation example is not consistent. In section 4.15.1.2, "A Generic Clifford Algebra", you state that e[i]*e[j]+e[j]*e[i] = M(i,j) and then give an example where ex M = diag_matrix(lst(1,-1,0,s)); ex e = clifford_unit(nu,M); and say M(0,0) = 1 -> e[0]^2 = 1. I get 2*e[0]^2 = 1 so that e[0]^2 = 1/2. Is M internally multiplied by 2?
Yes, this is a misprint in the earlier version of the documentation: it should be read as e[i]*e[j]+e[j]*e[i] = 2 * M(i,j). The present version of CVS of GiNaC can work with non-symmetric algebras and the defined equation is given in the ginac.info by e~i e~j + e~j e~i = M(i, j) + M(j, i), which is equivalent to the above for a symmetric M. Best wishes, Vladimir -- Vladimir Kisil email: kisilv@amsta.leeds.ac.uk www: http://amsta.leeds.ac.uk/~kisilv