Well, a relative path is relative
to the current directory...
say you're in /shu then 'sha' would be /shu/sha.
An absolute path, which always starts with a / would be
/sha. this is not /shu/sha, but /sha. Somewhere else enterly.
Note: Absolute paths in urls are from the webroot.
Absolute paths in cgi scripts (including perl and php) are from the webservers internal root.
On most shared hosting, your webroot is deep down in something like /usr/customers/joeWebmaster/
Btw, some claim that using absolute paths for urls is slower than relative ones, because the browser supposedly creates a new connection each time he encounters an absolute path.
I'd take this info with a grain of salt, because in http 1.0, each request creates a new connection, anyway, and with http 1.1, I'd believe browsers to be smart enough to handle this correctly (ie. by not opening a new connection to the same server).
