[PROJ] Tie breakers used when ordering candidate operations

[Greg Troxel]

"Lesparre, Jochem" <Jochem.Lesparre at kadaster.nl> writes:

> Greg wrote:
>>  People talk about WGS84
>
> Everyone should stop doing that and avoid EPSG:4326 at all times. ITRS
> is the one official international system, recognised by the UN.

That's an entirely different argument, dragging in the UN and hence
poltics :-(

I was trying to make the point that transforming coordinates to WGS84 is
problematic because it is an ensemble -- which is orthogonal to opinions
about which government is preferred.  And, One can view recent WGS84
realizations as esssentially being realizations of ITRS.

> The major difference is that WGS84(G1762) is really messy at
> centimetre-level. It ignores the tracking station velocities during
> the year and uses an annual step-wise update (at a random date in
> January) of tracking station coordinates used to generate ephemeris
> broadcast message.
> https://earth-info.nga.mil/GandG/publications/NGA_STND_0036_1_0_0_WGS84/NGA.STND.0036_1.0.0_WGS84.pdf

Interesting point that I had previously missed, but given that there is
no way to access WGS84(G1762) at anywhere near a cm, that's basically a
rounding error that has been operationally accepted.  If all the people
trying to use WGS84 had as their biggest problem this several cm of
fuzz, we'd be in a far better place.

As I read that, it's not WGS84(G1762) that has a stepwise change, but
rather that GPS itself adopts new station coordinates once a year, and
ignores the difference between reference epcoh coordinates and current
coordinates over the year.

> That is what I am doing all the time, except that I first ask them "Do
> you really want time-dependent international coordinates? Otherwise, I
> will assume you mean ETRF2000".

I don't see why WGS84(G1762) has a different time-dependent story than
ITRF2014.

The problem is that for global datasets (e.g. OSM), and TMS, there is a
need for a single CRS.  I don't see that everyone can be told to use a
regional/national crust-fixed frame, especially as that reduces but does
eliminate the problem -- stations are me have non-zero (albeit much
smaller) velocities even in NAD83.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 194 bytes
Desc: not available
URL: <http://lists.osgeo.org/pipermail/proj/attachments/20210212/9a85b554/attachment.sig>