[OSM-talk] tracklog credibility

Stefan Baebler stefan.baebler at gmail.com
Sat Jul 14 07:51:02 BST 2007


In http://www.openstreetmap.org/user/jcomeau_ictx/diary/73
user jcomeau_ictx wrote:

Can anybody comment on the legality of using Google Maps to get 
approximate lat/lon points, formatting those points into a GPX file, and 
uploading to OSM? I just want to know if I can be open about it, or if I 
have to be sneaky. It's inaccurate enough that nobody could tell, by 
looking at the data, how it was sourced... but if this is really a Bad 
Idea, I'll just have to come up with another method, or save up for a 
GPS unit.
/quote

As far as i understand the legalities this  shouldn't be done.

This brings up an issue how a GPS tracklog could be trusted.
Basic GPX is simple to produce, requiring no more than coordinates, 
which can be artificially "enriched" with timestamps to be technically 
acceptable for OSM.

OSM track importer could check and warn about traces where
-points have same time
-points have non-continuous time
-speed is constant or varies too much
...

In theory, we'd need some signature from the source (satellites in this 
case). Gps units could also add digital timestamps and digitally sign 
the coordinate pair :) But then we'd need to trust satellite owners and 
GPS unit manufacturers.

However, checking satellite signal strength, combining that with 
coordinates and checking those against known satellite orbits (those are 
used for AGPS) could be viable. Unfortunately same orbit data could be 
used to make fake traces if someone would want that very badly.

Stefan




More information about the talk mailing list