From: Henry Spencer (email@example.com)
Date: Thu May 09 2002 - 12:53:29 CDT
On 9 May 2002, Erland Sommarskog wrote:
> > >6.16 "An agent MUST use the same serving-name in Xref-headers as the path-
> > > identity it uses in Path-headers."
> Maybe there is some good argument for having the same name in Path and
> in Xref, but it is not obvious to me what it would be.
The issue is whether the Xref header is current -- generated by the server
on which the article resides -- or is a leftover from some other server,
not reflecting reality on the current server. This is important to a
reader agent. Which is why son-of-1036 made this name constraint a MUST,
so that a reader agent can check.
Xref has been so widespread for so long, mind you, that this may be less
of an issue than it once was. An alternative would be to require that a
server always delete an old Xref, whether or not it is inserting a new one.
Mind you, there's a potential problem with that: if a reader agent *is*
checking, it won't believe Xref headers with a different name, no matter
how much care has been taken to ensure that they are valid. I have no
data on how widespread such checking is.