As usual, click on the picture to see a larger version thereof. At first glance, the elevation graph seems to confirm that pressing a button on the 62s causes altitude spikes to be registered. Well, first, this is untouched data, as displayed by TopoFusionPro. At the start, I know I did set the tracking to be on, so I did press buttons and there is a slight, 10ft, loss of altitude whereas I'm in the parking lot. I don't know what caused the next two humps and depression between them but I did touch buttons around the 1.0 mark and the change there is not spectacular. I can't find a reason for any other spikes but the last one, when I was saving my track. 20 ft difference or so, I just eyeballed it, did not look at the actual track elevation numbers.
Do I want Garmin to look into it, perhaps fix it? Of course. But does it really matter to me for Search and Rescue purposes? Frankly, without others talking about it, I would not have noticed it, as I rarely even look at altitude. I rarely, if ever reset the barometer as altitude is of little consequence to me. My legs are good at telling me whether I'm going up or down and I can read the map (on the GPS or on paper).
The GPS graph shows altitudes between 320 to 370 ft, quite acceptable for my use.
As a linguist, I find this amazing. An innovative idea, what else would you expect from Google, and mega computer power making it work. And while the results are not perfect and human touch is often needed, by now, I'd say any translation I would do, will certainly begin on a Google Translate base, then add the human touch to make it better.