[ietf-dkim] Why bother removing features?
sm at resistor.net
Fri Jun 12 21:24:01 PDT 2009
At 17:06 12-06-2009, Steve Atkins wrote:
>I've seen interop problems caused very recently by a deep
>misunderstanding of what g= is for and how it interacts with i=. "We
>tried DKIM and gmail / yahoo said our signatures were invalid" is
>hurting adoption by senders today, and I'm sure some of that is
>because of misunderstanding of what some of the more obscure flags are
>supposed to be used for.
I've seen interoperability problems with most of the features before
RFC 4871 was published. Even if those features are removed, there
will still be people reporting invalid signatures with gmail and
Yahoo. That happened with DomainKeys and it happens to a lesser
extent with DKIM.
There is some operational experience which allows a better
understanding of what could go wrong. We have seen slow adoption of
DKIM for reasons which have not been discussed in this WG. Even if
these features are removed, it won't speed up adoption of DKIM as
there is a missing link which is not part of the work of this WG.
Part of the problem is that most people do not understand how
Internet Mail works in their environment, or to put it another way,
DKIM imposes restrictions on how it should work.
More information about the ietf-dkim