All right. I think I have this figured out now.
In Google Earth the location of the Viking 2 lander is given as 47.66° N, 134.28° E. That area is in the middle of an image from HiRISE which matches the one discussed here and the lander icon is pretty close to where that page says the lander was found in the image.
The minimum and maximum latitudes and longitudes of that image in Google Earth match up pretty closely with what the image label says:
MAXIMUM_LATITUDE = 47.783626969238 <DEG> MINIMUM_LATITUDE = 47.563352508918 <DEG> ... EASTERNMOST_LONGITUDE = 134.39963632293 <DEG> WESTERNMOST_LONGITUDE = 134.19901519181 <DEG>
So: within a few hundredths of a degree, the coordinates shown by Google Earth have to be correct.
Viking 2 is at 47.66° N, 134.28° E
In particular, the latitude isn’t 44°, of course, but it’s not 47.97° either — despite the previously mentioned claim associating that latitude with the MRO/HiRISE image.
As far as I know the Mars Science Laboratory (Curiosity lander) coordinates of 4.58° S, 137.45° E are accurate. They agree (within a hundredth of a degree or so) with Google Earth.
Mars Science Laboratory is at 4.58° S, 137.45° E
And putting those into the Mileage Calculator gives a distance of 3095.3 km. Google Earth, on the other hand, gives 3055 km between the same two points. I think the former is more likely to be right:
Distance is (?) 3095 km
though it’s harder to check. The two values differ by only about a percent, but that corresponds to about a week of walking at the rate I’m currently going, so it’d be nice to reconcile them!