[fitswcs] FITS WCS Paper V: Time - Draft for Review

[Steve Allen]

On Fri 2011-10-14T10:25:54 -0400, Arnold Rots hath writ:
> Version 0.90 of the FITS WCS Paper V on Time is now available for review:

Section 3.1
We need to be very clear, so
"At the present time"
should become something like
"Using existing FITS conventions"

ISO 8601:2004(E) does define the meaning of calendar dates prior to
1582-10-15 in sections 3.2.1 and 4.1.2.1.
They are defined as proleptic Gregorian calendar, with the proviso
that parties must agree to use such early dates.
Therefore, since we are asserting that we are agreeing that such
dates are proleptic Gregorian those dates are defined by ISO
and consistent with ISO.

Section 4

"If the HDU containd a table, ...
may be replaced by columns ...
with specific values for each row ("Greenbank convention")."

I'm not sure that this is clear.  Under what circumstances can any of
the keywords summarized in Table 4 not be used as columns?

Table 1

I think there may be too many notes regarding the various time scales
to fit well in the table.

TAI should be avoided prior to 1972 because
1) TAI had not been authorized until the 14th CGPM in late 1971
http://www.bipm.org/en/CGPM/db/14/1/
http://www.bipm.org/en/CGPM/db/14/2/
2) TAI had not been available for any contemporary time stamping
   mechanisms prior to 1972-01-01

TAI should be used with caution prior to 1977 because of the 1e-12
change in rate, and for precision work TAI should always be corrected
using TT(BIPMxxxx).

TAI should be avoided now because of document CCTF/09-27
http://www.bipm.org/cc/CCTF/Allowed/18/CCTF_09-27_note_on_UTC-ITU-R.pdf
where the CCTF "stresses that TAI is the uniform time scale underlying
UTC, and that it should not be considered as an alternative time
reference."

UTC should be used with caution prior to 1974 because the meaning of
the name was unknown outside the metrology community.
UTC should be used with extreme caution prior to 1972-01-01 because
different contemporary sources of timestamps were providing different
time scales.
UTC should not be used prior to 1960-01-01 because coordination of
broadcast time did not begin until then, and prior to 1961 only
time sources in the US and UK were providing it.

GPS system time should not be used before its inception date 1980-01-06.

GMT has to be allowed after 1972 because
1) broadcasts of time continued to say "GMT"
2) the recommendation saying not to use GMT came in 1974
3) the IAU Comm 4 & 31 joint resolution saying not to use GMT came in 1976
4) the CCIR Rec. 535 saying not to use GMT came in 1978
and most folks could not know of those recommendations/resolutions.
So I'm not sure what "should/must/may" wording to use, nor what date.

Section 4.2.1

After TIMESYS is defined the clarity will be improved if
"In relevant context it may be overridden by"
is replaced by something like
"In the context of axes in image arrays or table columns TIMESYS
may be overridden by"

Section 4.2.2

"(as long as one does not venture too far from the earth's orbit),"
I am not understanding what this is trying to communicate.
The use of TOPOCENTER anywhere other than the surface of the earth
begs questions of what meaning is intended.

For a trendy example of this, note the recent OPERA superluminal neutrinos.

Table 2

We now have an object giving us extremely precise measurements of the
center of mass of 4 Vesta, and next year that same will apply to 1
Ceres, so we may find ourselves wishing we had defined Reference
Positions for many of the minor planets.
I wish we could get away with recommending "MPnnnnnn", but the advent
of Pan-Starrs and LSST operations may quickly overflow 6 digits.
Perhaps we should simply allow that a purely numeric value
TREFPOS = "nnnnnnnn"
shall be interpreted as the minor planet number?

Table 3

If Table 3 is to have precise meanings then ET cannot be in a column
heading.  ET could appear in the caption saying something like
"To its commonly available precision ET may be used with a Reference
Position anywhere within the solar system."

If Table 3 is to have precise meanings then it seems to me we either
have to give reference to a prescription for the 4-d transform to the
EMBARYCENTER or else we have to omit that row.

Or, we could overload the table meaning further (at the risk of making
it even less clear) by adding entries like "0.002 s" in the column TDB
for TOPOCENTER, GEOCENTER, EMBARYCENTER as an indication that those
are acceptable uses if the precision of the numbers is no better than 2 ms.
But I really don't like that idea much.  I'd rather delete EMBARYCENTER.

Section 4.2.2, terrestrial coordinates.

The last sentence demands the IAU 1976 ellipsoid (a shape from before
satellite geodesy had matured), but the cartesian coordinates are
specified as being ITRS.  ITRS specifies that the length of the meter
is consistent with TCG, not TT, whereas all recent ITRF publications
eschew that and use a length of meter consistent with TT.

Furthermore, any allowance of geodetic latitude or altitude
again implies an ellipsoid.

I will point out that VLBI avoids the issue of ellipsoid by specifying
terrestrial coordinates in a cylindrical system: Rspinax, Longitude, Z

But again I raise the spectre of the OPERA superluminal neutrinos.
If we want precision we had better be consistently explicit, and
I don't think the terrestrial coordinates here are safely explicit.

Section 4.2.3, at its end

The end note is too complex to be clear.
We need a new section with a worked example showing the math for
computing the time values associated with a pixel location along an
axis, and we need to refer to that example here.

Section 4.2.4, at its end

EventRA,EventDec should now be referring to what is in Table 8, not Table 6

Also, those celestial coordinates must necessarily be accompanied by
RADESYS
EQUINOX (if above not ICRS)
and MJD-OBS or DATE-OBS
either in the header keywords or as other columns as is pointed out by
WCS Paper 2 section 3.  I think we should include words saying that.

Section 4.4.1, at its end

As with section 4.2.3, it's just too complexly dense in meaning, so we
really need a new section with worked examples showing the math for
using TIMEOFFS and MJDREF/JDREF/DATEREF, and we need to refer to that
section here.

Section 4.4.2

"but may be overridden, in appropriate context, by"
would be much clearer as something like
"In the context of axes in image arrays or table columns TIMSYER may be
overridden by"

Section 4.3

I am happy to see tropical and besselian years,
but these numbers are not useable.
McCarthy and Seidelmann took their tropical year from Laskar, Table 5 of
1986A+A...157...59L
and that's ancient by comparison to Simon et al., section 5.9.3 of
1994A+A...282..663S
And besides that the calculation for Besselian year is not right.
Arnold Rots and I should go play with polynomials and units.

Section 5.2.1

"time scales TDB and TCB may"
could become
"time scales TDB and TCB (or ET, to its precision) may"

Section 5.2.3

CRVALia does not have italic "ia"

Appendix A interacts heavily with the things I wrote above about Table 1.
We'll have to figure out how to harmonize which goes where.
Also, the URLs for IAU documents have proved notoriously unreliable,
so we will have to cite the IAU proceedings themselves as primary with
the URLs being secondary references.

--
Steve Allen                 <sla at ucolick.org>                WGS-84 (GPS)
UCO/Lick Observatory--ISB   Natural Sciences II, Room 165    Lat  +36.99855
1156 High Street            Voice: +1 831 459 3046           Lng -122.06015
Santa Cruz, CA 95064        http://www.ucolick.org/~sla/     Hgt +250 m