This one still needs some work. There should be options for filename/checksum/byte-for-byte comparison as well as a way to limit the search by directories and/or file types. Minimum and maximum file size limits would be nice too (0-byte files should be skipped by default).
This is kinda advanced but in Vista most of it’s files are hard links, i.e. what looks like duplicates is really just two differently named links to the same area on the hard drive. See Hard link - Wikipedia. I’m not sure how you should deal with this besides filtering out the whole winsxs folder… Anyway, if true duplicates are found, there should be an option to hardlink them instead of deleting one or the other. This way you eliminate the duplicate while preserving the directory reference/pointer. This is much better than simply deleting one of the duplicates, especially in the case of system files that need to be in a certain place but I only know of one program that does this - FINDDUPE: Duplicate file detector and eliminator.