On Feb 16, 2009, at 10:55 AM, Eliot Lear wrote: > On 2/16/09 6:28 PM, Jim Fenton wrote: >> (d). My suggested changes to Eliot's proposal are at >> http://mipassoc.org/pipermail/ietf-dkim/2009q1/011153.html >> > > +1.
+1. For what its worth, a reputation system can utilize an opaque i= value to isolate messages identified as abusive, provided the number of such exceptions is kept within a practical range. The alternative might require each receiver to deal with abuse that might have been consolidated by the i= value. Dave's draft goes a bit too far when describing even the domain as being opaque. Is this attempting to suggest the d= value is not that of a domain? While the i= value can be opaque, any language that appears to lessen its relationship to that of on whose behalf the signature had been added should also warn against overlapping real and fictitious namespace if only to better support the uses described by ADSP. -Doug _______________________________________________ NOTE WELL: This list operates according to http://mipassoc.org/dkim/ietf-list-rules.html
