[LEAPSECS] the big artillery

Warner Losh imp at bsdimp.com
Wed Nov 5 10:46:47 EST 2014

On Nov 4, 2014, at 6:07 AM, Zefram <zefram at fysh.org> wrote:

> Warner Losh wrote:
>> Users can only get UTC(foo) or a signal derived from UTC(foo) (e.g.,
>> traceable to NIST) and never UTC itself. Of course they can get to a
>> putative TAI(foo) trivially (I say putative, because as far as I know, no
>> lab generates TAI synchronized signals for reasons you go into). However,
>> they cannot get back to TAI(BIPM)
> ...
>> That's the hair I'm trying to split.
> I'm not seeing the split.  I think you've just said exactly the same
> things about TAI and UTC, and hence agreed with me.  Let's see:
> A user can listen to a radio time    :  A user can listen to a radio time
> signal that provides PPS markers     :  signal that provides PPS markers
> synched to UTC(national_lab)         :  synched to TAI(national_lab)
> along with time codes that can       :  along with time codes that can
> be decoded to UTC(national_lab).     :  be decoded to TAI(national_lab).
> UTC(national_lab) is not exactly     :  TAI(national_lab) is not exactly
> the same as UTC(BIPM), but tends     :  the same as TAI(BIPM), but tends
> to be within tens of nanoseconds     :  to be within tens of nanoseconds
> of it.  UTC(BIPM) (the canonical     :  of it.  TAI(BIPM) (the canonical
> UTC) is not precisely available      :  TAI) is not precisely available
> in real time: it is determined       :  in real time: it is determined
> retrospectively by comparing the     :  retrospectively by comparing the
> clocks of the various national       :  clocks of the various national
> labs.  UTC(national_lab) timestamps  :  labs.  TAI(national_lab) timestamps
> can be corrected to UTC(BIPM)        :  can be corrected to TAI(BIPM)
> retrospectively by applying the      :  retrospectively by applying the
> offsets given in Circular T.         :  offsets given in Circular T.
> Which part of this is wrong, for which time scale?
> Of course, I'm referring to the same radio signal in each case, and the
> PPS markers are the same.  The decoding process is slightly different,
> but with ready availability of the TAI-UTC difference it's not appreciably
> more difficult.  The national lab's job of steering clocks and generating
> the signal is the same in both views.  The received signal is traceable
> to the national lab and ultimately to BIPM in both views.  The only
> remaining asymmetry is that the "TAI(national_lab)" notation isn't
> officially approved, but that doesn't diminish the concept.

The markers aren’t the same. TAI doesn’t actually exist until after the fact.
That’s the point I’m trying to make. BIPM maintains TAI and UTC as paper
time scales (see http://tf.nist.gov/general/pdf/1498.pdf). Different national
labs steer all their clocks to what they thing UTC should be, and distribute
that. Although technically a paper clock UTC is very much seen as the
operational side of things. It is realized in real time, everybody exchanges
time stamps based on UTC as seen on the GPS constellation (so UTC(USNO))
etc. TAI is viewed more the source of truth, but finding that truth is done
after the fact.

Could one create, in real time, a TAI signal, sure. But no one does.

> There is no technical difference here, only a political difference.

The difference is also philosophical. TAI is for the BIPM time geeks to tinker with,
and get in closer alignment with TT over time. It’s very much oriented to the
paper clock aspect of things, with much tinkering of data. It is more than just a labeling
of seconds. It connotes all these other things which are published on a regular
schedule, just not a real time schedule. UTC, while also a paper clock, is also more
than just a labeling of seconds. It is the operational time scale. It is the one everybody
is using to coordinate time, to get their notion of civil time, etc.

So while an outside may see “oh! look! TAI is a nice way to label seconds that
doesn’t suffer from the leap second issue, let’s use it!” the insider reacts to that suggestion
with horror because they know all the behind the scenes machinations. They know that
TAI isn’t a time scale produced in real time. It is the connotations of the terms that cause
this reaction.

So maybe I’m again off in the weeds trying to differentiate between philosophy and a
notion of the right thing and politics and ideology.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 842 bytes
Desc: Message signed with OpenPGP using GPGMail
URL: <https://pairlist6.pair.net/pipermail/leapsecs/attachments/20141105/8eb16fae/attachment.pgp>

More information about the LEAPSECS mailing list