On Thu, Aug 28, 2008 at 11:13 AM, Steve Simmons <[EMAIL PROTECTED]> wrote: > > On Aug 27, 2008, at 3:38 PM, Russ Allbery wrote: > >> Simon Wilkinson <[EMAIL PROTECTED]> writes: >> >>> Rerunning the election, in whatever form, is likely to be time >>> consuming, and probably result in a vanishingly small turnout the second >>> time round. Should we just give the outgoing chair a casting vote? >> >> Personally, I'd be comfortable with just flipping a coin. >> >> One way to flip a coin would be to give the position to whoever's MD5 hash >> of their name lexicographically sorts earlier. It has the minor potential >> drawback that the tiebreaker would be known in advance, though. > > I'd prefer that tie means fail. This is supposed to be setting standards, > which ought to be agreed at a near-consensus level. If we're tied, clearly > we've got no consensus. >
That seems reasonable to me for protocol decisions, but the protocol decisions are already based on consensus (cf section 2.3.4). The only elections I'm aware of that will occur are those made for personnel. My suggestion is that the remaining chair breaks the tie. Yes, that leaves the decision in the hands of the remaining chair, but it follows the principle of 'the buck stops here', and it also reflects the responsibility in the hands of the chairs and the voters. The remaining chair can always choose to break the tie in any random fashion. -- Steven Jenkins End Point Corporation http://www.endpoint.com/ _______________________________________________ AFS3-standardization mailing list [email protected] http://michigan-openafs-lists.central.org/mailman/listinfo/afs3-standardization
