--On Samstag, Juli 10, 2004 22:00:29 +0300 "Andrew G. Tereschenko"
<spf-discuss(_at_)spam(_dot_)24(_dot_)odessa(_dot_)ua> wrote:
Ralf,
Thank you for your question.
I understand your argumentation against "spf" prefix. But this has nothing
against prefixed values as whole.
ACK on that. But the same argument will be valid every valid hgostname. So
the domainname to be used for SPF information should not be a valid
hostname or at least not ione that could normally be used with a CNAME RR
(which is impossible for 2LDs as they require at least two NS RRs).
[...]
I agree with you that we need to reduce this conflict probability. How
about using "spfv1" prefix ?
As soon as you use the prefix for this kind of information, the same name
lends itself for an informational website. If you call something "FOO",
foo.example.net seems to be a logical place for infirmation about this kind
of data. The only way to avoid this is IIMHO the use of non-hostname
domainnames for this kind of data.
[...]
This is perfect prefix - little, has a meaning SPF v 1, versioned (think
about future versions), while still legal for DNS (no "_").
The underscore ist perfectly valid for RRs in DNS, it is only forbidden for
hostnames. And as we are using just TXT RRs ...
I will defend prefixed SPF TXT records technology until somebody will
clearly show that there is real _technical_ problems with this.
There are no technical reasons. It is just a matter of admnistration and
acceptance. Many hosting services also provide DNS service. A growing
number supports user settable TXT RRs, but still many of them do not
support underscores in domainnames. Also some of them do not support
subdomains at all (just the plain 2LD and www.2LD). Unless these providers
can be convinced to enable their customers to set TXT RRs for arbitrary
domainnames, those customers would no be able to use SPF records for their
domains. And it would be very hard to convice those customers to change
their provider just to be able to publish SPF records when we have to
convince them to publish those records in the first place.
IMHO it is iomportant to keep publishing SPF information simple for the end
users if we want to get widespread acceptance among non-geeks.
Ralf Döblitz