[Top] [All Lists]

[smime] draft-dolmatov-dnsext-dnssec-gost-01 technical & editorial issues (fwd)

2009-09-17 07:50:08
I apologize for cross-posting, but the following excerpt of my
post to namedroppers might be of interest for PKIX and SMIME,
the originators of RFCs 4490 and 4491, for which I have filed
Errata -- in part to solicit discussion.

The full message is archived at:

----- Forwarded message -----

From: Alfred Hönes <ah(_at_)TR-Sys(_dot_)de>
To: dol(_at_)cryptocom(_dot_)ru, ran(_at_)cryptocom(_dot_)ru, 
Message-Id: <200909171025(_dot_)MAA15356(_at_)TR-Sys(_dot_)de>
Subject: draft-dolmatov-dnsext-dnssec-gost-01 technical & editorial issues
Date: Thu, 17 Sep 2009 12:25:12 +0200 (MESZ)


I have reviewed draft-dolmatov-dnsext-dnssec-gost-01 and found
a number of issues of varying degree of severeness.
Items (1)..(3) below are the hard stuff, the nits are in (4).


First of all, I'm confused by, and strongly oppose to, the
introduction of little-endian encoding for the on-the-wire
encoding of specific elements of this proposal, as requested
by the last paragraph of Section 2 of this draft.
This is incompatible with the Internet architecture and also
with the ASN.1 framework quoted by the draft, which makes
consistent use of "big-endian" (network) byte order.

Unfortunately, the quoted related RFCs 4490 and 4491
(cf. Section 4 of the draft) already have created bad precedents.
I have filed Errata Notes for these RFCs (EIDs: 1884 and 1885).
As noted there, mixed endianness within one protocol, or even
within a single specification for an extension to a protocol
is a very likely source of confusion, implementation errors,
and, consequentially, interoperability problems.
This practice should be avoided consistently.


----- End of forwarded message -----

Kind regards,
  Alfred Hönes.


| TR-Sys Alfred Hoenes   |  Alfred Hoenes   Dipl.-Math., Dipl.-Phys.  |
| Gerlinger Strasse 12   |  Phone: (+49)7156/9635-0, Fax: -18         |
| D-71254  Ditzingen     |  E-Mail:  ah(_at_)TR-Sys(_dot_)de                    

smime mailing list
<Prev in Thread] Current Thread [Next in Thread>
  • [smime] draft-dolmatov-dnsext-dnssec-gost-01 technical & editorial issues (fwd), Alfred Hönes <=