How about advising those who might want to save a link away somewhere
to actually check the link they are saving:
if the BASE differs from the URL used to obtain the resource, then
HEAD the BASE URL to see if it seems to point at the same document.
This would be advice for clients that might want to save away a
'bookmark', and robots that want to index a site and remember
The extra HEAD call wouldn't be necessary for links that either (a)
did not contain a BASE URL (b) where the BASE URL was obviously
equivalent to the retrieval URL (e.g., different host names which
resolve to the same IP address).