DanRyanIrish

This sounds intuitive on first thought, but I think it's actually a big undertaking because of the pixel alignment requirement.

Hi @wtbarnes. Thanks for raising this issue. I would class it as a bug. After a quick glance _I think_ this should be fixed simply by extracting user input `coord_params`...

Yes, I think if possible this should be a figure test. I'm not that expert on figure tests when the output is an animation. Perhaps you can compare with any...

@Cadair @sageyu123, let me know what you think of this summary of our discussion

Good point, I completely forgot about that! But it seems ```NDCube``` has not inherited that. It seems the ```write``` method is part of ```NDDataRef```, not ```NDData```. See http://docs.astropy.org/en/stable/nddata/ Anyway, ```NDDataRef.write```...

Hi @yashrsharma44. I'm really glad to hear you are interested in addressing this issue and would love to see you write a PR. I'd be very interested to hear more...

OK. Well it's good to have that confirmed. Would you still be interested in working on a PR to do this?

OK, great. In that case I can write up an outline of how the write method should work and then we can discuss.

Hi @yashrsharma44. I think it's best to discuss it here so anyone who's interested can see. Then we can update the description at the top of the issue with decisions...

So an ```NDCube``` is comprised of ```data```, ```uncertainty```, ```mask```, ```meta```, ```wcs```, ```unit```, ```missing_axis```, and ```extra_coords```. Let's ignore the ```extra_coords``` to start with as they may pose some difficulty. One structure...