If the main problem with leap seconds is their future unpredictability, isn't there a compromise option between the status quo and no more leap seconds? Couldn't they come up with a fixed schedule for leap seconds for many centuries at a time, based on current predictions of approximately how many will be needed each century? That should be good enough to prevent real human-noticeable drift between UTC and perceived day/night time. If a correction is needed because current predictions turn out to be wrong, it seems that could be done very rarely, with centuries of lead time in changes to the schedule. Was anything like this considered at the international level? [ I know this is not something the IETF can decide, of course ] -- Cos _______________________________________________ Ietf mailing list Ietf@xxxxxxxx https://www.ietf.org/mailman/listinfo/ietf