Here is a change log of draft-mengwong-spf-00 vs -01
* IETF IPR boilerplate updated.
* -00 uses "RFC2821" type references, while -01 uses "[RFC2821]"
* minor grammatical changes
* -01 has a note saying entire version is compared, so "v=spf1" != "v=spf10"
* if you need additional records, -00 RECOMMENDS using the _spf
subdomain, while -00 says MAY.
* -01 defines how multiple strings in a single TXT RR should be
handled, along with a few usage notes about this stuff.
* -01 adds the option of allowing HELO checks even if the MAIL FROM is
not <>
* section 2.2.2. "Lookup" changed NXDOMAIN from -00's "fail or none" to
-01's "unknown"
* -01 adds support for zone cuts
* "Error" changed from "MAY reject" to "SHOULD reject" (both using 450
SMTP code)
* The version number in "v=spf" changed from allowing a minor number
to just a number. That is, support for minor version numbers was
dropped.
* Some ABNF cleanup
* '/' is used instead of '|' for alternation
* mechanisms and modifiers names are defined as
alpha/digit/-/_/. instead of just alpha. (Yeah, ip4 was illegal
under the -00 ABNF, even though it was defined in the draft)
* mechanisms and modifiers are defined to use "macro-string" instead
of just *VCHAR
* double quotes are used instead of single quotes for literals.
* part of the definition of <macro-char> is broken out into its own
ABNF token called <transformer>
* Clarification that modifiers can appear after things like -all.
(earlier versions of the SPF spec said they *had* to be last, but
this was dropped.)
* In -00, section 3.2 says the default result is "-", while in section
3.3, it says the default value is "Neutral." -01 just refers to
section 3.3
* -01 adds a note that mechanism and modifier names are case
insensitive. (The ABNF already requires this, but you have to read
RFC2234 to know this.)
* -01 notes that other mechanisms can be defined in the future.
* -01 defines the include mechanism to not match on a softfail or
neutral. (it was undefined in -00)
* -01 notes that the "default" modifier is deprecated and says that
"match_subdomains" is a future modifier
* -01 defines an "accredit" extention modifier
* -01 says the exp= result "SHOULD" be shown instead of "can" be.
* -01 notes that the %{i} macro variable uses "nibble format" for
ipv6.
* -01 defines the new %{c} macro variable for use in the exp text (and
only there)
* -01 changes %{t} to only be defined in exp text
* -01 defines the new %{r} to be usable everywhere.
* -01 corrects that in macro variables, the strip digits transformer
is optional. (The ABNF and SPF implementations said they were
optional.) Also, it says implementations don't have to support an
infinite number of digits, as long as it supports the value of '9'.
* The text explaining how macro variable transformers operate has been
edited to clarify things.
* section 8.3 "Conformance with regard to sending e-mail systems" has
been substaintially edited.
* -01 clarifies in section 8.7 that SPF clients can do whatever they
want with the results of a "fail"
* section 8.9 "Recommendations" says that a compliant SPF sender
SHOULD use FQDNs on their HELO domains. (This was also part of the
editing of section 8.3, mentioned above.)
* clarifications made to 9.4 "Verbatim Forwarding"
* in section 10 "Security Considerations", the include/redirect limit
was changed from a maximum of 20, to a minimum recursion depth of 20.
* The acknowledgement section was edited and updated with new
contributors.
* some light editing was done on the extended exampnles
* -01 added a recommendation to join spf-devel and read
http://spf.pobox.com/developers-guide.html
Well, the list was slightly longer than I remembered. I guess it
might be good to do a change log from, say, draft-mengwong-spf.02.9.4
(the first "frozen" spec, as of Dec 14 2003) and draft-mengwong-spf-00
-wayne