On Mon 25/Nov/2024 17:16:32 +0100 Daniel K. wrote:
The top level "version" element is required by the XSD in RFC 7489, but
not everyone got the memo and only about 40% of the reports I've
received has it in the XML data.
In an earlier release of the aggregate reporting draft, the XSD was
changed to make "version" optional.
I searched the list archives around the time of the change, but did not
find anything about it, and the commit message was also a bit terse.
See https://mailarchive.ietf.org/arch/msg/dmarc/JdRxmT9Aw3HkWM7rr3Av9B3EwRc
Also, "version" is not described in the draft text at all, and I think
some changes should be made to address it. Please help me with some
background to find out what should actually be done or changed.
Matt's argument in the message linked above was:
I wouldn't recommend the removal. RFC 7489 has introduced tag
<version>1.0</version> and existing implementations will rely on its
existence. Without <version>, chances are the report consumer interprets
the report as pre-IETF draft version.
1&1, the only ones I know about to implement this draft adds the
"version" element with value "1.0"
According to RFC 7489.
The draft now specifies an xml namespace to be used for the feedback XML
reports; urn:ietf:params:xml:ns:dmarc-2.0
Here we allude to dmarc-2, which did not come to pass, as the version
number did not need to be increased.
The schema is to be referred per IANA Considerations.
https://www.iana.org/assignments/xml-registry/xml-registry.xhtml
The "2.0" part is to differentiate it from other schemata numbered 1.0, if they
exist. We can still change it...
How is versioning supposed to work going forward?
Will the XML namespace need to be changed on update?
In that case, will the change of namespace identifier be enough to
distinguish between the versions?
If a IETF document updates the schema, increasing that 2.0 will suffice.
The schema is extensible, so it is possible to define child schemata without
changing
Should we change "version" back to be a mandatory element and add text
similar to the text in RFC 7489:
The "version" for reports generated per this
specification MUST be the value 1.0.
That's similar to MIME, isn't it?
or, if namespace must be updated, abolish the "version" element altogether?
See above.
Best
Ale
--
_______________________________________________
dmarc mailing list -- dmarc@ietf.org
To unsubscribe send an email to dmarc-le...@ietf.org
