The other thing that Flavia and I discussed, probably as a later step, is that we should be able to cache the scanned BDA for a scanned archive - this makes a lot of sense for things like installed libraries. Though, with the caching we get from MC of the real scanning work, perhaps it won't have any impact...