[LEAPSECS] A standard for leap second smearing

[Steve Summit]

Tony Finch wrote:
> Steve Summit <scs+ls at eskimo.com> wrote:
> > There seems to be a presumption in several comments in this
> > thread that (3) is necessarily identical to (2), but I think
> > that's a bad idea.
>
> You are completely right up to here, but it's probably unwise to dictate
> how (which part(s) of the system) leap second smear is implemented.

I'm not trying to dictate an implementation, but my point is that
we can't assume one, either.  As Martin Burnicki and Stephen
Colebourne have already pointed out, it's important to understand
whether one system's smeared time "leaks out" to others over NTP,
or not.

> The advantage of a standard is that you could just tell someone your time
> with a flag indicating whether it is smeared or not, and they can fairly
> straightforwardly translate (provided they also know when the leap seconds
> are).

Me, I'd very much rather *not* add this sort of thing to (say)
NTP, because NTP doesn't have a problem with leap seconds.  NTP
communicates leap seconds explicitly, unambiguously, and without
awkward jumps.  Its the (typically Posix-compatible) client
systems that have the problem, so breaking NTP in order to fix
them seems like a step backwards.  (And the conditions that make
an NTP-level smear appropriate for the likes of Google and Amazon
in their data centers don't apply to the open Internet.)