[Ietf-calsify] definitive references for UTC

[Nigel Swinson]

> I'm fine with adding the reference, but it comes with a caveat.
> 
> At the Prague meeting, there was consensus that iCalendar should ignore
> leapseconds, i.e. act in all respects as if they didn't exist.  I still
> think this is both least-surprise for users, and the mode that's almost
> certainly implemented by 100% of current iCalendar implementations.
> 
> The defining characteristic of UTC, however (as opposed to the other
> flavors of Universal Time, or to TAI), is the fact that it has leap
> seconds.
> 
> Getting this right will require some somewhat tricky wordsmithing in the
> document.

In case anyone doesn't know what I researched recently:
http://www.meinberg.de/english/info/leap-second.htm

"The Windows system clock does not know about leap seconds, either. However, there is a precompiled version of ntpd for Windows available which is capable of slewing the Windows system time quickly in order to account for the insertion of a leap second, without affecting the clock synchronization loop provided by ntpd. "

So in the end I had little choice but to ignore leap seconds as the OS didn't make me aware of them, and without the patch didn't even know about them.

Perhaps an option would be to change the phrasing to describe that support for leap seconds is dependant on support from the clock subsystem and therefore the system may never report their existance.  Then go on to describe how to cope with iCalendar data when your OS doesn't report the leap second.

Equally I'm totally happy with iCalendar ignoring leap seconds too.

Nigel