At 2:38 PM -0800 3/8/07, Hallam-Baker, Phillip wrote:
I do not see where making the change in the spec is going to reduce
compatibility.
On the contrary I think that if we don't change the spec to say what
we meant we will see more incompatibility.
Or we can change the spec to make it clearer what the spec says.
The whole point of Proposed vs Draft is to catch this type of thing.
Bosh: that isn't even part of the point. Changing the
bits-on-the-wire interpretation is only done when advancing a
standard except when some huge mistake has been made.
We don't have a standard yet.
Bosh again. See
<http://mipassoc.org/pipermail/ietf-dkim/2007q1/007026.html>. It is a
standard *today*.
Its an errata on a proposed standard. It happens.
It happens sometimes, but only in extreme cases; this isn't one of those.
If someone has an implementation that has a problem with this that
would be another issue entirely.
The first message on this thread was an example of that.
At 3:16 PM -0800 3/8/07, Hallam-Baker, Phillip wrote:
How about
"Implementers should note that the ABNF speficied above is in error
and may be changed in future protocol versions. Implementations
should accept key records compliant with the following syntax:"
Absolutely not. What a great way to make people not want to implement the spec.
--Paul Hoffman, Director
--Domain Assurance Council
_______________________________________________
NOTE WELL: This list operates according to
http://mipassoc.org/dkim/ietf-list-rules.html