This is a proof of concept of a system that screen scrapes
data from multiple shipping information websites (FedEx,
APL, Maersk) and delivers it to Google Earth.

There are two pieces. Index.cgi is a (bare bones) HTML user interface
that takes a shipper and tracking number as input and creates
a Network Link in your Google Earth browser as output.

The Network Link connects to track.cgi, giving it the shipper and
tracking number. Track.cgi gathers the current status of the shipment
from the shipper's website, then formats it into a set of 
objects that Google Earth renders onto the earth. The Network
Link is configured to reload every 24 hours.

This stuff is proof of concept code; it will probably fail on
any input that has not specifically been tested. In particular,
it will fail:

   * when the tracking number is malformed, or results in
     unexpected data

   * when it encounters a location that has not been hardcoded
     into GeoDB.pm. See comments in GeoDB.pm for how I envision
     it working in the future.

When it fails, it makes no effort to give the user a useful
error message. It's a prototype that shows what's possible,
but it's not yet usable for day to day tracking.

Jeff R. Allen
jra@nella.org