My recent post had a misprint in a url, as pointed out by Mike Barr. Here is the correction www.groupoids.org.uk/nonab-a-t.html The book has a section 10.4 on acyclic models for crossed complexes, rather than chain complexes. This means that the theorem readily gives weak equivalences rather than just homology equivalences. Ronnie [For admin and other information see: http://www.mta.ca/~cat-dist/ ]