I've noticed recently that some search engines (seem to be mainly meta-search engines - the ones which fetch results from other real search engines and feed them to you) mangle some URLs.
My music database is done in CGI. To view the details for an album, you'd look at a URL like the following:
The problem is that the search engines in question mangle this, changing the ? into a %3F. When the user clicks on it, instead of running the cd.cgi script, the webserver tries to locate a file called "cd.cgi%3F1170", which doesn't exist, and the user gets a nice error page.
Course, it could be a misconfiguration on my server's end... I don't know enough about apache to be able to tell. All I know is that this result is entirely un-useful for both me and the person looking for the info.