Doug, That was discussed at length in the WG and we ended up where we ended up.
I'm sure you'll make the point again in IETF LC, (which is entirely fine), but we don't need to revisit the topic on this list, so please stop raising it here. If you think you have some compelling reason that we should discuss it on-list again, feel free to mail Barry and I off-list. However, I'm pretty sure I'll not be convinced from what I've seen to date. Thanks, Stephen. Douglas Otis wrote: > On Oct 30, 2008, at 9:26 AM, John R Levine wrote: > >>> I think most of the likely candidate domains for using >>> "discardable" would disagree with your assertion John. >> Then I have to say that they don't understand what discardable >> means. Really, it says feel free to throw our mail away if you have >> the least doubt about it. This chronic misconception is the main >> reason that I doubt that discardable will be useful in practice, >> since only a small fraction of people who assert it will truly >> understand what they've said. > > > The term "discard" in RFC 5321 is used to mean the silent dropping of > information. > > In the prior note you said "It's really only useful for banks and > places like Paypal sending out notices about accounts, not for any > domain with individual users." > > As Michael suggested, these domains will not want to have their > messages silently discarded. > > The term "discardable" suggests permission to violate RFC 5321 Section > 2.1 MUST accept responsibility for either delivering the message or > properly reporting the failure. This same MUST is echoed in Section > 3.6.3 , 4.4 and 6.1. Section 6.2 provides advice on making an > exception. > > ADSP will affect _all_ messages from the domain. Few customers will > welcome messages from similar domains by the same institution. > Customers that have opted to go paperless will be negatively affected > by "discardable" since this affects all communications, even those > considered extremely important. Unless DKIM signatures are so robust > that _any_ type of failure provides very high confidence that the > messages are seriously fraudulent, then the silent discard is not the > best choice. If some institution ever had a problem with their DKIM > signatures, how will they become aware of the problem when _all_ their > messages are being silently discarded? > > At least two people on the DKIM list suggested that this term would be > problematic. Instead of "Discardable", "Dismissible" would not imply > that not reporting a failure is now allowed. > > -Doug > _______________________________________________ > NOTE WELL: This list operates according to > http://mipassoc.org/dkim/ietf-list-rules.html > _______________________________________________ NOTE WELL: This list operates according to http://mipassoc.org/dkim/ietf-list-rules.html
