[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Allowing IP addresses as path-identiites (Re: #1047 Path field delimiters and syntax - status)
Harald Tveit Alvestrand wrote:
--On 6. september 2005 13:46 -0400 "Forrest J. Cavalier III"
<forrest@xxxxxxxxxxxxxxx> wrote:
YES - IP addresses should be allowed as path-identity.
NO - IP addresses should not be allowed as path-identity.
I think these poll choices are poorly specified.
I say MUST ACCEPT, MUST NOT generate. Is that YES or NO?
Shucks, I forgot "none of the above" as a choice.....
Are you talking about components of the path, or about identities that can
appear in Xref: and Injection-info:?
Since you replied by asking a question....does that mean you agree your poll
choices are ambigous? This has nothing to do with understanding English.
I observe that your understanding and use of English is excellent, but what's
the reason to avoid RFC2119 terms? If we can't even ask simple questions
precisely, then let's just give up on writing a useful document.
"not allowed" can mean MUST NOT generate, MUST NOT relay, MUST NOT
use to determine propagation. Or does it mean that the USEFOR grammar ends up
having no way to specify IP addresses in path-identity (which means there
may not be any RFC2119 MUST terminology.)
"path-identity" is used in the grammar defining the syntax of three
different header fields. (-- A mistake if you ask me....Quick tell me
what will every clients do with an IPv6 address with colons in the
path-identity Xref header field?)
When someone votes "NO - IP Addresses should not be allowed as path-identity"
are they voting that Relaying servers are supposed to reject articles which
have an IP in the Path: header field? Are they voting that servers are permitted
to reject articles which have an IP in the Path: header field? Or that
the Injection-info-experiment-trying-to-be-a-standard-field cannot use an IP
address for the first terminal after the first CFWS.
I hope people know what they are voting for. I don't.