Without some way to choose which rule to use and when to use it, how can a recommendation that has conditional rule usage be implemented? - Ralph On Jun 3, 2008, at Jun 3, 2008,8:50 AM, Simon Josefsson wrote: > Thomas Narten <narten@xxxxxxxxxx> writes: > >> Longest match in 3484 is a hack, ant it only works some of the >> time. The WG most certinaly knew this when we approved the >> document. But it was felt that longest-match was better than no rule >> at all, as it helped on some situations. >> >> The real discussion that should be held is what could we replace it >> with if we pulled out the rule entirely? >> >> We do need something better. What would that be? > > I suggest to describe the longest match as a rule that can be used > some > of time, and describe a new rule, random selection, to be used some of > the time. Then you recommend the latter approach when the application > don't know that the longest match rule is more appropriate. > > This would seem to meet the objective of causing the least surprise > and > the least damage. > > Thanks, > Simon > _______________________________________________ > IETF mailing list > IETF@xxxxxxxx > https://www.ietf.org/mailman/listinfo/ietf _______________________________________________ IETF mailing list IETF@xxxxxxxx https://www.ietf.org/mailman/listinfo/ietf