Option to not rephase the data

[mkuiack]

Is rephasing the data to a central pointing important for the measurement set format?
It would be preferable to have the option for each timestep to have its original pointing.

[tammojan]

It would be possible to have a pointing per time step (in the POINTING table of the measurement set), but neither DPPP nor wsclean support handling such measurement sets in their entirety. A better option would be to write out the pointing in Az/El-coordinates. DPPP is almost capable of handling those.
If you can give me some data I could work on this a bit.

[mkuiack]

This isn't a high priority request though, I was just wondering what the technical reason for the rephasing is.

Because there is no streaming measurement set format or support for such a thing in DPPP and WSClean, and because of computational resource limits observations of several hours need to be broken up into bite-size chunks anyway.