[Openstreetmap] Re: How do we handle large amount of GPS points?

[Joachim Mæland]

Sorry about my threading... (The import was somewhat incomplete)

I'm Petter's Linux user, holding ~7000k GPS points and new on this list... 
Unable to participate directly, I still have some experience in 
collecting and using GPS points for map production...

[Petter]
> I plan to export on time instead, and split the dump into week or
> month bulks.  But that code isn't done yet, so this will do for now.

Go for smaller sections. There is no way you can predict accuracy and 
layout of the points, based on date only. Unless you run the track in a 
systematic manner.

[Petter]
> I believe the exchange format should include enough data to filter it
> using DGPS info, and that the long time storage format should include
> some info on the percieved accuracy.  Not sure which values are most
> useful for this.

IMHO: There are no NMEA information available that can be used for 
trusting GPS points blindly... If it looks accurate, the fix is 3D, PDOP 
reports < 3 and the sun is shining: You just might have the correct 
positions for your street. I have seen (and hopefully deleted from my 
recordset) positions several metric miles from my correct location. At 
the same time; all the relevant accuracy inforation is in place, looking 
great. Inacurate, yet close to accurate are however more dangerous.

My mandatory tools logs GPS positions, timestamp from localhost and fix. 
Limited resources have stopped further enhancements, but my wishlist 
includes: "GPS timestamp, DPOP, HDOP, VDOP and # of active satellites. 
Hopefully this might, together with creative coloring and plotting, be 
used for verifying the accuracy of the dataset.

If you provide a lightweight log_to_file_tool, I'll surely use it and look 
into a plotting system using color to visualice accuracy.

[SteveC]
> > fix - you guys might want that?
[Petter]
> It might be useful, or it might be a complete waste.  I'm not sure
> yet.

The overhead involved in keeping this information is neglectable, compared 
to usefullness in special areas, where the accuracy of your dataset is 
questionable. You can use this information to disregard the points, not 
to trust them.

BTW: I'm now investigating the use of a Asus WL-500G for collecting GPS 
positions for another project. I'll keep you informed.

Regards
Joachim