On Jan 5, 2010, at 12:33, Nick Zitzmann wrote: > On Jan 5, 2010, at 1:23 PM, Quincey Morris wrote: > >> Firstly, you can't *in general* normalize to a specific time, because you >> can't in general know that the time exists on every date. > > Unless you use a specific time zone, and store that time zone information > elsewhere in the model. And if you use GMT, as I mentioned earlier, unless > some cosmic event happens that disturbs the flow of time, there is no such > thing as a GMT time that does not exist, since GMT ignores transforms like > daylight savings.
I'll say this part again, and keep saying it till someone actually tells me I'm wrong: ... Wait, no, I did that already. Assuming that you're going to ultimately present the user with dates in the user interface, it's still a cultural assumption that the dates as the user perceives them map directly onto UTC date/times. Yes, I know I'm standing up for cultural perspectives that likely don't exist, but I think it's important to make the point that dates vs date/times is a subtle matter. The OP wanted to use NSDate objects to avoid the clunkiness of multiple values/variables with NSDateComponents. I'm arguing that avoiding NSDateComponents is a *lot* more work, and that using NSDate for this purpose is *very* easy to get wrong. As to the cosmic event -- well, I already mentioned leap seconds. Depending on which way the leap goes, there *will* be a UTC time that doesn't exist, or one that exists twice*. :) *I think -- I don't really know enough about it. _______________________________________________ Cocoa-dev mailing list (Cocoa-dev@lists.apple.com) Please do not post admin requests or moderator comments to the list. Contact the moderators at cocoa-dev-admins(at)lists.apple.com Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/cocoa-dev/archive%40mail-archive.com This email sent to arch...@mail-archive.com
