On Thu, 11 Jan 2007, Dave Crocker wrote: > > As has been clear for some time, the OPES topic is both important and > difficult. That sort of combination always makes want to look for some > history of exerience with ways to solve the current problem. In the case > of OPES, I do not know of a qualifying history, nor do I see the effort > to build something based on such experience. The closest history I can think of is the sendmail milter interface, and the clamav and spamassasin interfaces. All of these are implementation- specific: the Milter interface is exported as an API, not a protocol, and the other two are essentially application-internal interfaces. Milters are closest to what OPES is trying to do, but AFIK not quite as ambitious: geared for content *filtering* not *adaptation*. So I agree that OPES doesn't seem to be based on any well-developed experience. > If the current document is intended as a case analysis for a particular > application -- namely email -- to serve as *input* to the design of the OPES > architecture and protocols, then I do not see how the current document > achieves that. I believe it's intended to fit the existing OPES architecture to SMTP. Tony. -- f.a.n.finch <dot@xxxxxxxx> http://dotat.at/ FAEROES: WESTERLY BECOMING CYCLONIC 7 TO SEVERE GALE 9, PERHAPS STORM 10 LATER. VERY HIGH. RAIN OR SHOWERS. MODERATE OR GOOD. _______________________________________________ Ietf@xxxxxxxx https://www1.ietf.org/mailman/listinfo/ietf