Good suggestions. I do agree with explaining why a subset was tested and how that subset was chosen. In some cases, however, listing all known implementations would be onerous, ambiguous (an implementation that shipped (or went beta) with partial (or untested) possibly unannounced standards support) or even contrary to privacy agreements. Lisa On Tue, May 26, 2009 at 12:38 PM, Stephane Bortzmeyer <bortzmeyer@xxxxxx> wrote: > On Tue, May 26, 2009 at 09:50:52AM -0700, > Lisa Dusseault <lisa.dusseault@xxxxxxxxx> wrote > a message of 43 lines which said: > >> Do you have any suggestions for criteria that could be broadly >> applicable and useful? > > Following Paul Hoffman, you could put in the I-D something like: > > All known (by the author of the report) implementations SHOULD be > listed in the report and, if only a subset is tested, the reasons for > selecting this subset SHOULD be written down. For instance, "We tested > Foobar and Example because they are widely regarded as the two most > common implementations." or "We tested Foobar and Example because they > are the only ones available under a free software licence." > > (Paul's message contains other good examples.) > > _______________________________________________ Ietf@xxxxxxxx https://www.ietf.org/mailman/listinfo/ietf