ADK46erNow Test Day 1

A Little Background

Most of the ADK46erNow’s app features could be tested without needing to actually be in the high peaks region.

The one feature that required on site testing was the 360 View. This is an augmented reality (AR) feature allowing the hiker to look through the camera view within the app and identify the high peak summits around her.

To test this feature I created three maps and three tables of peak locations, compass bearings, and GPS locations. I chose The Meadows (not a peak, in case the name didn’t give it away) which I could get to by car and park with views of some of the high peaks. This location would provide me with a quick and easy way to test the app in a short period of time (not hiking) so I could retreat to my laptop and fix the app if needed.

I also chose Cascade Mountain, a high peak, and Hurricane Mountain, not a high peak, but with great views of the high peaks to the west and south.

And…

Day 1: The Meadows

…and retreat I did. I arrived in Lake Placid around 3pm and immediately headed to The Meadows.

Compass bearings were… let’s call it, “not even close”.
On iOS the 360 degree AR video view through the phone was reversed (mirrored). What the..! How is that even a thing.
Android’s video feed was correct (not mirrored), but the compass bearings swung erratically all over the place.

Take no notice of the hand-written checkmarks or degrees on the map below for the moment. The peaks and bearings the app gave me are scribbled in at the top (bottom?) of the page. In summary the readings were not even close to what I calculated they should be.

Did I have a bad starting GPS lat/lon coordinate for The Meadows which I guessed at from Google maps and used for the calculations in the table? Well, no, actually taking the GPS reading on location gave me very close to my Google maps pin drop.

I’ve highlighted Phelps and Wright in the images below. You can compare the bearing scribbled in on the map page with what they should be in the data table. Yes, I’m a bit disappointed at this point.

And that’s why you test in the field… and then have a beer to try and sooth the depression.

So, late night debugging back at my motel room found a difference between “true north” and “magnetic north”, but that only accounted for small degree differences. The actual bug was an embarrassing variable substitution (calculating a necessary intermediate value and then not using it in the final calculation)! Yeah, kind of a stupid miss there.

And after much searching I figured out how to fix iOS’s reversed video issue while leaving Android’s alone. Apparently this is something iOS does. Don’t ask me why.