[rfc-i] URL checking

Martin Duerst duerst at it.aoyama.ac.jp
Fri Oct 21 01:04:36 PDT 2005

At 09:19 05/10/20, Joe Touch wrote:

 >However, URLs are more like "book 5, row 6, on Joe's office shelf" -
 >they don't identify the document, they identify WHERE to get the

Well, with things like round-robin A records and Akamai stuff and so on,
the user may have absolutely no clue whether that shelf is in
the US or in Japan or wherever :-). That's a pretty bad analogy
then. Another thing is that it's very easy in HTTP to set up
a redirect from an old to a new location, which is used by
the browser without the user really noticing. Again, the
shelf analogy fails badly.

 >IMO, if you identify the URL sufficiently (IETF BTNS WG at
 >IETF 63), you have more than enough information to find it, and a URL is
 >just a location identifier that WILL change.

Everything will change, except the fact that everything will change.

One advantage of URIs not mentioned in detail is that they can
simply be clicked in many contexts, whereas search requires
time to select the right search terms, time to put these terms
into the right form field, and time to select the best search
result from a long list. Even if it's the top one you're looking
for, some time is needed for deciding that the top one is the
best one.

Regards,    Martin. 

More information about the rfc-interest mailing list