On 4/1/11 9:18 PM, John R. Levine wrote:
I think it can be immensely useful if the list plainly says /why/ the
WG closes. As Rolf noted, DKIM is not (yet) a well refined protocol
that any of us would recommend his grandma to make use of.
If that's the requirement, I think that pretty much every IETF standard
since the dawn of the Internet is a failure. DKIM's main audience is the
people who run mail systems, for MTA-MTA security, not individual users.
it seems to me you don't take Alesandro's statement very serious, by
responding only to this part of his message. Let's face the situation.
Some 90% of all mail sent across the Internet is spam. The vast majority
of this spam is caught by DNSBL based filtering. As we all know, the
upcoming use of IPv6 poses some interesting challenges to the way
current DNSBLs operate/are used.
Over the years, 'the people who run mail systems' have started to use
non-(IETF-)standard anti-spam tools, not bothering too much about the
collateral damage (false positives and delivery delays) they cause. We
all know the examples of call back verification, greylisting, DNS back
and forward checking of IP/names etc. etc. Because these techniques are
not standardized, they cause problems with the delivery of legitimate
mail. Although DKIM is not an anti-spam technique in and by itself, it
is the only spam-related standards track technology around.
By closing down the WG the momentum will be lost; in my view it's
essential to keep momentum and a WG that is actively investigating the
impact of DKIM and further developing the standard based on real-world
usage, can be a way to keep the industry and government interested. Note
that the investigation of the real-world usage of DKIM has led to the
removal of g= and to the proposal to remove i= in 4871bis.
However, if there is consensus to close down the WG, I would like to
suggest to followup this WG by chartering a reputation WG, which will
pick up the work done so far on the domainrep mailing list, to make a
start with 'cashing' on the results achieved with DKIM so far.
/rolf
_______________________________________________
NOTE WELL: This list operates according to
http://mipassoc.org/dkim/ietf-list-rules.html