Hey all, The way I was taught GR, the summation convention applies on terms where an index is repeated strictly with one covariant, one contravariant. But reading through a translation of Einstein's GR foundations paper just now it looks like the index placement doesn't matter (I've seen it this way on Wikipedia too! :P). I've never actually seen a term like, say, a_\mu b_\mu where you have repeated upper indices or repeated lower indices, so as yet this hasn't been an issue, but I'm curious what the consensus on the convention is, and whether it actually matters (are there terms/can there be terms in GR with repeated upper/lower indices?). Thanks!