Electromagnetic integral equations requiring small numbers of Krylov-subspace iterations

I was looking at Cat's website, and I found the page proofs for this paper, which I believe is Lemming's first publication. Congrats!

I only have one question: What happens if you consider random cracks?