On Tue 30/Apr/2013 20:02:11 +0200 Edward Lewis wrote:
On Apr 30, 2013, at 12:28, Alessandro Vesely wrote:
...The basic fact that killed the SPF type is the ability to use
TXT as a replacement. There must be an analogous of Gresham's
law: "Bad types drive out good ones."
I disagree with the assertion that what killed SPF is "the ability
to use TXT as a replacement." It has nothing to do with one
option being superior to the other, it was the lack of technical
incentive to switch from one to the other.
I'm not following: Either one option is superior to the other, or
they are more or less equally fit. In the latter case, it becomes a
question of taste and there is no technical incentive.
I post this in the sense that if the root cause is not understood,
no solution will stick. Here is a message with my recounting of
what led us to this point:
http://www.ietf.org/mail-archive/web/dnsext/current/msg12681.html
I agree that security shackles are different from provisioning. Yet,
their effects on deployment are similar: They are part of those minor
problems that always have to be faced when introducing something new.
Such problems are only faced and overcome if the new type provides
some added value, which SPF over TXT does not.
I don't see the death of SPF has a harbinger of things to come.
There's a strong case to be made the failure happened before 2004
and the root cause has since been corrected by changing the RRTYPE
type allocation policies. Still that fix was too late to ever let
SPF sprout wings.
Type allocation policy isn't but one requirement for new types. To
check the syntax and the semantics of data records, and to
conveniently format them on the wire would make for real added value.