visualize last update on mapmarkers

[stephanlindauer]

hi i received the following email (sorry its in german):

Eine Visualisierung, wie alt die Positionsdaten sind, wäre ganz gut. 
Zum Beispiel könnten die Positionen ihre Farbe wechseln, je älter die Daten sind. 
Nicht sekündlich, aber alle 5-10 Sekunden etwas dunkler. 
Zum Beispiel wie beim Lightningmaps.org.

Viele Grüße
Tilo 

i think that if we implement something like that, we should make the difference in color very subtle in order to not confuse people. also it might be tricky to communicate to the user what those different colors actually mean.
what's your opinion on this?

[ligi]

I see ( and had ) the problem - also not really sure if colorizing would really help. Perhaps we should first try attack the root of the problem - the latency in updates. At the moment the data is synced all 12s - that means it might be 24s latency. How the problem could be attacked:

• sync server-sync and location-updates
• leverage gcm ( with fallback to e.g. mqtt for fdroid ) for updates ( might also be good for battery )
• use polyline encoding / optimize the protocol ( at the moment it is really blown up with all the latitude / longitude keywords )
• ..

[cbalster]

I don't think the colouring solution is feasible. Something like that would only make sense if every position was individually colored which would require timestamps, individual updates of the markers etc. and would probably a pretty good hit on performance.
I think the real problem the user has, is that he feels he cannot trust the locations shown. So thinking about it this way might open up some new solutions for fixing it.

[stephanlindauer]

obviously we should make sure the server works as intended. one thing that still baffles me, is, that when @cbalster pulled data with his python script, he saw delays. when i look at what my timelapse rendering, i cant see any significant delays.
lets look into server problems when i have uploaded that code. (sorry for not having done that already)

so if coloring them separately doesn't make too much sense, how about having a badge that counts down to the next update and up from the last successful update?

[JulesDoe]

I just played around with fake GPS on my phone and loked at the web-map what would happen.. in 4 tries the delays in position update i experienced were 3:40min; 4:30min; 20sec; 3:20min; 4:10min.
4 minutes means beeing off more than 600m at 10kmh traveling speed. :/

I very much like the idea of transmitting the timestamp of the actual GPS update. That would allwo a heatmap style diesplay of the Data. Dots could just become smaler and less visible over time till they decay.