[Standards-JIG] [Fwd: I-D ACTION:draft-saintandre-jabberid-01.txt]

Dave Cridland dave at cridland.net
Thu Jul 20 13:07:41 UTC 2006

On Thu Jul 20 13:42:29 2006, Hal Rottenberg wrote:
> 5.  Security Considerations
>  "A forged Jabber-ID
>   header may break automated processing; therefore the Jabber-ID 
> header
>   SHOULD NOT be depended upon to indicate the authenticity of the
>   message or the identity of the sender."
> Should you mention here that the JID could be validated out-of-band 
> using xmpp?
Probably not - you can validate that the JID's domain exists, but I'm 
not sure you can do much more automatically anyway, can you? Even if 
you could validate that the user exists, I'm not convinced this gains 
you much.

More comments:

The draft has a lot of normative references, many of which are only 
referenced in informative ways. I think only RFC2822, XMPP-URI, and 
RFC4234 are actually needed to implement, RFC2119 is only needed 
because of two uses of MUST and SHOULD NOT, one of which is actually 
in text that's really informative. (Many other specifications are 
needed in practise, but are required only for RFC2822 and XMPP-URI).

For example, I don't like the use of the UNICODE and US-ASCII 
references - they're embedded in informative text and really aren't 
relevant beyond the discussion of why Jabber-Id uses rules from the 
URI spec. To put it another way, the entire second paragraph could be 
stripped from section 2, and it'd still be a valid spec, so it's at 
best an informative reference.

That leaves you with just the SHOULD NOT in the security 
considerations - a slight rewording will leave you without any 
RFC2119 language at all, which is, I think, a sign of a very clear 
specification. (I only know of one other recently, RFC4469).

Dave Cridland - mailto:dave at cridland.net - xmpp:dwd at jabber.org
  - acap://acap.dave.cridland.net/byowner/user/dwd/bookmarks/
  - http://dave.cridland.net/
Infotrope Polymer - ACAP, IMAP, ESMTP, and Lemonade

More information about the Standards mailing list