Not really broken links, so to speak. More like a file comparison.
A program that starts from the home page and crawls all links, creates a list of all files (htm, gifs, jpgs, pdfs, whatever) that are linked or "touched" by the spider.
Then another process that finds all the files in those folders that aren't touched. Those files might also have links in them, but a person would never hit the page. Its an unneeded file since it can't be reached through any links via the home page or subsequent links from there.
I think I might have found a tool at
http://www.lithopssoft.com/hlv/validator_filelists.html which will do the trick in a round about way.
Thanks.