02/08/2013, 05:36 AM
Yesterday, I ran my usual test driving home from work, on the latest 0.2.9 "technology preview" version.
It ate a third of my battery in forty five minutes, and that was expected.
Funny behaviour: Remember Jan said that, when rotating map, other elemts, such as markers, will not be rotated? Be amazed because, when map rotation is OFF, markers ARE rotated! It seems that counter-rotation code is called also in not rotation scenarios...
Of course, with map rotation enabled, markers get their counter-rotation pill, and they show up always straight, wich is a great achievement.
Pity that, in the end, I went back to non rotating maps. Just because it's too jumpy: I could easily see the whole thing flash upside down for a split second. Also, it was funny seing the arrow progressing sideways on the road.
Could you thing about some way of ignoring a heading readout, when it's obviously false? For example, if current heading is +/- 30º than the last one, it can't be possible. Of course, that amount should relate to speed. I mean, turning ninety degrees in a second is possible, if you are walking. But faster, it's not, so it must be a spike in the heading readout.
Many systems on other devices give this kind of "filtered" heading readouts, to solve the "jumping compass" issue.