On Mon, 1 Mar 2010, comex wrote: > On Mon, Mar 1, 2010 at 3:03 PM, Kerim Aydin <[email protected]> wrote: >> Not that interesting. If the two messages are exact endpoints of a 30-day >> time period, then neither message is "within" the R869 30-day limit and >> so registration is not prohibited. It might be more interesting if the >> rule read "cannot register until 30 days have passed." "Within" to me >> implies (t1,t2) while "have passed" could be (t1,t2]. -G. > > I'm not sure about that at all. e.g. if I somehow simultaneously > create the obligation to do something, and do it, I don't think it's > clear at all whether I've satisfied the obligation.
In case of multiple events in one message we have ordering precedents for that. If you manage to send to separate messages with the same time stamp then I agree, it might fail due to ambiguity. But this is not the situation where we have a starting time stamp, and an ending time set by the starting time stamp. > However, in the > real world the message cannot be sent *exactly* 30 days later; it must > be either before or after the deadline, but if the difference is > subsecond, we don't know which. Thus a CFJ on whether Wooble is a > player would be easily UNDETERMINED and e might be able to have some > fun. Legally, I think we have to assume that the message was sent *exactly* to the resolution of the timestamp (e.g. subseconds on down .00000...). This isn't the real word time of sending, but it is the legal world time of sending. It seems common-sense-ish to me, but I'll agree that there's no precedent on that. And I think the "within" langauge applies to being within that legal (not actual) window. What would be interesting would be if two different people set up cron jobs to do something contradictory (e.g. where order mattered) with identical times to the resolution of the timestamp. There, neither person would have performed an ambiguous action, but together they would have. -G.
