improve plot_alignment to show 'dig' points from device coord frames

I'm routinely in the situation were I need to evaluate either the quality of the MEG HPI localizations and/or our external headshape measurements (Polhemus fastscan and fasttrack).

There currently isn't a way to visualize the MEG HPI locations since they are not in head coordinates or if you transform them to head coordinates, they typically end up labelled the same as the points coming from polhemus and you can't tell them apart.

My proposal would be to

1. keep the points in the info['dig'] in the coord_frame they are measured in.
2. have plot_alignment move points from MEG coord frame to the display frame (HEAD or MRI) and plot them using a different glyphs (sphere:head_frame, ??:MEG FRAME)

I'm curious what others think of this or what problems it will cause :)