David Hine wrote: > Telstra is the largest telco in Australia and was once fully owned by the > federal government as part of the post office. They used GRASS for > analysis.. as below GRASS only in research institutions.. Not in this case. .. > >From http://www.pcworld.idg.com.au/index.php/id;1787316484
I was expecting to see v.net.*, but from the methodology in the paper they used v.distance. And they mention lat/lon a lot. Perhaps it is time to revisit GForge bug #378: http://wald.intevation.org/tracker/?func=detail&group_id=21&aid=378&atid=204 Apparently they double-checked the results with published values, and a distance given in decimal degrees is obviously wrong, so I don't think they made any mistakes. My guess is that they actually reprojected the lat/lons for the processing or created a v.dist output=line_segments map and used something like d.what.vect (which does use geodetic distance) to extract the distances. Regardless, it would be nice to fix 'v.distance upload=dist,to_along' to use geodetic distance; check the nearest feature calculation isn't biasing to features in the N-S plane at high latitudes; and more generally make the vector modules & libraries more Lat/Lon aware. Or start adding checks into the code of any module which could output a misleading result: if (G_projection () == PROJECTION_LL) G_fatal_error( _("%s works in Euclidean space. Please project your data."), G_program_name() ); Hamish __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com
_______________________________________________ grass-dev mailing list grass-dev@grass.itc.it http://grass.itc.it/mailman/listinfo/grass-dev
