From: Charles Lindsey (chl@clw.cs.man.ac.uk)
Date: Mon Oct 04 1999 - 06:44:19 CDT
In <ylu2o9nlvt.fsf@windlord.stanford.edu> Russ Allbery <rra@stanford.edu> writes:
>Ralph Babel <rbabel@babylon.pfm-mainz.de> writes:
>> Charles Lindsey wrote:
>>> if "FOO", "Foo" and "foo" are actually different UUCP names in the
>>> world maps, and the sites in question choose to use them in Path
>>> headers, then propagation may suffer.
>> => clearly a "SHOULD NOT" according to RFC 2119.
>I wouldn't object to changing that to a SHOULD NOT for UUCP names and a
>MUST NOT for everything else.
Actually, this is an unenforceable restriction. If some site decides its
Path-entry is "Foo" (that being its UUCP address) and some other site, in
a different continent, decides to use "FOO" for the same reason, then it
is most unlikely that the one knows about the other. So it is unreasonable
to specify that one of them (whichever it is) MUST NOT use that name. So I
would suggest we reduce it to SHOULD NOT in all cases.
The result will be that if that situation does arise, then articles that
have passed through Foo may be rejected at FOO (if Foo's software is case
insensitive). So FOO will miss some articles, though the flooding
algorithm will ensure that the problem does not affect other sites escept
maybe some of FOO's closest neighbours.
The only other thing we could do is to declare that path-entries are
always case-insensitive (but that might break other things). But my
opinion is that the problem just has to be lived with - though I would
recommend demoting to SHOULD NOT for the reasons stated.
-- Charles H. Lindsey ---------At Home, doing my own thing------------------------ Email: chl@clw.cs.man.ac.uk Web: http://www.cs.man.ac.uk/~chl Voice/Fax: +44 161 437 4506 Snail: 5 Clerewood Ave, CHEADLE, SK8 3JU, U.K. PGP: 2C15F1A9 Fingerprint: 73 6D C2 51 93 A0 01 E7 65 E8 64 7E 14 A4 AB A5