On the DC-10 we had three INS (Inertial Platforms) which were capable of cross-talking. The present position was entered most carefully by the operating pilot and monitored by the other pilot. The alignment process took 17 minutes and during that time the aircraft could not be moved (or even have the tug attached) otherwise the platforms would topple and you had to start all over again.
It was not possible to enter the latitude in error by more than 4 nautical miles for the platforms knew the exact precession rate for a given latitude and would not erect.
Longitude was the big problem. If the present position was entered as East instead of West, the platform had no means of knowing whether that was right or wrong (the distance between the Meridian and 10W, for example, is exactly the same distance between the Meridian and 10E for any given latitude). That is what happened to the Korean 747 that got shot down by the Russians if my memory serves me right.
All waypoints (1-9) were indeed entered by hand and carefully checked by the other pilot and also checked against the Plog and against special charts that we carried for that purpose.
I later flew an ex-Air New Zealand DC-10 which had the facility to automatically load waypoints from a prepared tape which was fed into the nav-computers (Collins AINS 70 system). That was fine as long as the waypoints had been loaded correctly on to the tapes by the planning department in the first place.
Part of the cause of the ANZ DC-10 disaster, when it hit Mt Erebus in Antarctica, was because a chap in the planning department had decided they could save 17 nms by removing one of the waypoints from the tape. This was true but the new track took them straight through the mountain so the flight was cut short by rather more than 17 nms.