The handling of URLs in this script looks suspicious. In particular, the GET and POST that get manufactured use the full "http://host" URL form. This might work for whatever web servers Sprint uses (Netscape Enterprise 3.6 when I looked a moment ago), but other web servers might take umbrage at seeing a request come in that starts with "http://". I see these periodically in my Apache logs, and a bit of experimenting suggests that Apache won't abide by this scheme.