My thought was to treat the ways as the primary datum (per Connelly) and align the saddle parallel (albeit slightly acute for concave facing) to said datum via a precision square off the freshly ground inverted V-way... and then later align the spindle to the ways as well. I reckon I can check the alignment between saddle and headstock to verify I didn't futz anything up, but my primary objective is to get the saddle to travel parallel to the ways. I saw others align their saddle this way in other rebuilds. Bad plan?
To do that alignment you're trying to take a datum from a surface ground at an angle, perpendicular to the cross slide ways and well below the ways. I'm sure that it *can* be done, just not sure that it *should* when aligning the headstock first eliminates all of those issues and makes the saddle alignment simple.