On Thu, May 20, 2004 at 08:06:25PM -0700, Russ Allbery wrote:
> Isaac To <kkto@csis.hku.hk> writes:
>
> > Unluckily, the thread started by John is not a "thread gone mad", or at
> > least I don't think it is. The underlying views can be summarized as
> > one of
>
> > a. SPF is a bloodily stupid idea that adds to user inconvenience without
> > much gain to the net as whole. So to prevent SPF from gaining grounds,
> > Debian should actively not implement anything to make sure mails from
> > SPF enabled sites to actually get delivered. Then users of SPF will
> > one day understand how stupid SPF actually is.
>
> > b. SPF is a good idea, and Debian should support it even when very few
> > others are using SPF. So it really should implement SRS to make sure
> > sites with SPF enabled, and users with who filter mails based on SPF
> > rules, will be served nicely. Debian should itself publish SPF records
> > as well.
>
> > c. We don't know whether SPF is a good idea, and if enough people do use
> > it, we should not bar them from access to the Debian list. So even if
> > Debian might not implement SPF (i.e., not publish a SPF record in the
> > Debian DNS server), it should implement at least SRS to make sure other
> > users of SPF is served nicely.
>
> I think there's a fourth one:
>
> d. It's not clear whether SPF will actually help anything, and some
> work is required to implement SRS and deal with forwarding back any
> bounces, not to mention possibly breaking user assumptions about
> envelope senders (for example SRS will turn unresolvable envelope
> senders that could be rejected as spam into resolvable envelope
> senders unless you take them apart again). So a wait and see
> approach may be indicated before actually taking time to do the work.
e. SPF is a good idea, but it is not appropriate for Debian to
publish SPF records. We should largely ignore it.
--
.''`. ** Debian GNU/Linux ** | Andrew Suffield
: :' : http://www.debian.org/ |
`. `' |
`- -><- |
Attachment:
signature.asc
Description: Digital signature