Bug report: overall ascend/descend value is wrong

[routeconverter]

I don't know such gps term as NASA SRTM3 and why it is related with date. But I understand it is an issue of algorithm.

Not exactly: NASA SRTM3 is the project name of a mission to create an elevation database for the world. When you use RouteConverter's Complete->Elevation function, it's replacing the rather inaccurate GPS elevation information which more accurate, smoothed elevation data for the NASA SRTM3 database. Just follow the instructions in my previous post to see the effect with your file: If you choose Complete->Elevation for all positions, the Ascend/Descend decreases to 1636/1676.

In my thinking, the total ascend should be sum of ((delta = ph(n) - ph(n-1)) > 0? delta : 0), but now I know it is not so simple.

What RouteConverter does is exactly this simple approach: RouteConverter just naively sums up positive and negative deltas between positions. You can watch this if you make the Ascend and Descend column visible.

I used to upload my track to a track web, so I am surprised when I see the total ascend/descend are so big in RouteConverter. Looks like it is just about different algorithm.

Propably the "track web" is using a more sophisticated algorithm?