APT produced FITS files usually fail Image Solver

[fredvanner]

1. APT already knows the ICRS coordinates when it transforms them to apparent coordinates before sending them to the mount. So we are not asking for something requiring any additional work.

APT may know the ICRS coordinates for a target selected from a catalogue (if I use catalogue look-up, rather than just searching by eye). By the time I have re-centred the field to frame the image I want, APT only knows the current RA, DEC.

2. If the metadata center coordinates are apparent, considerable work is required to perform the (approximate, never accurate) inverse transformation to ICRS coordinates, which are required to perform the necessary database search operations to compute an astrometric solution.

The only way APT can write ICRS coordinates to the header is to do exactly this conversion. If APT writes the proper data to the header (i.e. the RA and DEC, together with the date / time and the location of the observation), then the conversion to ICRS coordinates could (if required) be done at any future date, to any required accuracy, possibly by software more powerful than APT. However, in general this will not be required, since any catalogue data used for image solving / annotation must be forward propagated to the time of the observation in order to handle proper motion (and ephemeris motion of solar system objects). So all the data actually used for solving / annotation will be referenced to the time / date of the observation.
That means that catalogue searches to support plate-solving / annotation do have to do an approximate inverse transformation to the catalogue epoch (which may or may not be ICRS, and may vary from catalogue to catalogue), but only accurately enough to set an appropriate search area - since the resulting searched objects must be forward propagated to the time of the image.