Dear All, I wish to propose for a discussion a small patch to indexed.cpp. I looked at the raised issue with slow performance in contracting Dirac gammas. The present algorithm does this recursively, in Vladyslav's example the deepness may reach 140 nested calls. The proposed patch reduced it around to 100 levels by avoiding to call simplify_indexed() if it will re-direct to simplify_indexed_product() anyway. However, I shall say that the patch only marginally improves timing: by a couple of percents at best. Best wishes, Vladimir -- Vladimir V. Kisil email: kisilv@maths.leeds.ac.uk www: http://www.maths.leeds.ac.uk/~kisilv/ Book: Geometry of Mobius Transformations http://www.worldscientific.com/worldscibooks/10.1142/p835