[LEAPSECS] My FOSDEM slides

[Zefram]

Brooks Harris wrote:
>As a practical matter its most convenient use currentUtcOffset = 10s
>at 1970-01-01T00:00:00 (TAI) = "Zero PTP Time".

So your entire justification for claiming this amounts to it being a
convenient fiction.  Your reference to clauses of IEEE 1588 didn't add
anything to this argument.

>No no no! NTP, and this table, do not address Leap Seconds at all.

Right.  That's why you can't infer any particular leap second history
from that table.  Not for 1972 to 1999, and not for 1900 to 1972.

>There *are* zero leap seconds between [ 1900-01-01 *plus 22 forgotten
>Leap Seconds* ] and 1999-12-31.

If the leap seconds have been forgotten, how do you know how many
there were?  Another part of the NTP spec is explicit about the notional
pre-1972 UTC having an unknown leap history.

Applying the view that NTP scalars are actual counts of seconds from
epochs that change at every leap second, the unknown notional leap seconds
between 1900 and 1972 impose an unknown slippage of epoch between the 1900
and 1972 entries in the table.  The 1972 instant is 2272060800 seconds
from one of these epochs; there's no reason to suppose that that's the
same epoch as the one from which the 1900 instant is 0 s removed.

>On 2015-03-08 03:43 PM, Zefram wrote:
>>1900-01-01 00:00:00 UTC is a fictitious timestamp, because UTC doesn't
>>extend back that far.  The NTP epoch is, in that respect, fictitious.
>
>Yes, of course. That's what NTP says, and that's what I'm saying. Its
>proleptic, I think is the correct word.
...
>I guess you could call it "fictitious" if you want. But it seems like
>any numbering system is "fictitious" in that sense.

There's a big difference between "proleptic" and "fictitious".
Applying the Gregorian calendar to the year 1415 is proleptic, because
the Gregorian calendar hadn't been invented at the time.  But it's
not fictitious, because the Gregorian calendar serves perfectly well
to label specific days of that year.  The Battle of Agincourt happened
on a day that the proleptic Gregorian calendar labels as "1415-11-03".
We can firmly link the modern label "1415-11-03" to the day that at the
time was referred to as "25 October 1415" (using the Julian calendar)
or as "Saint Crispin's Day".

The reference to "1900-01-01 00:00:00 UTC" is fictitious because the
mechanisms of UTC do not extend back that far.  UTC cannot be used
to label specific instants in 1900.  The label "1900-01-01 00:00:00
UTC" does not refer to a specific second that we could identify using
contemporaneous terminology, or for which we could identify historical
events.

>>And that's your misunderstanding of the NTP table again.
>
>I don't think so.

You haven't dissuaded me from this position.  I don't quite grasp the
logic you're using, but it's clear that you're being inconsistent in how
you relate leap seconds to the table entries.  Your entire rationale for
this leapless pre-1972 pseudo-UTC is based on your misinterpretation of
this clarifying table in the NTP spec.

-zefram