I want to rsync on the basis of content and content only; I don't want to copy if the file is no different and I don't want not to copy if the file is at all different (of course). I think the right incantation is with the -c (always checksum) option: rsync -rlvc here/ there But it isn't absolutely clear whether under some circumstances rsync might decide to copy files even if they haven't changed; e.g. perhaps it thinks it would be more efficient to copy small files rather than test them. Is the -I (--ignore-times) option purely a heuristic /efficiency thing that will not affect what is updated? Will the -n (--dry-run) option always provide a completely accurate prediction of what is going to be updated if given with the -c option? Even if going over a network? (ssh). It is possible for -v (verbose) to tell me why an update is being made? (in case I'm not using the -c for instance) Obviously here I'm using rsync for it's 'revision control' features more than for it's efficiency features. Regards, Matt