On 10 Jun 2026, at 17:07, David Woolley <[email protected]> wrote: > > On 09/06/2026 16:01, Graham Whaley wrote: >> The OS trig data says (iirc) that it should still be accurate to 0.1m > > It shouldn't be! Continental drift should have changed it by almost double > that. Even the longitude of Greenwich varies relative to WGS 84, over time, > and latitude also does. GPS uses the WGS 84 reference frame, not the ETSR89 > one, which is specific to the Eurasian tectonic plate, so is gives more > stable coordinates in past and present parts of Europe. > > The national grid references should be stable, but OSM is GPS based, not NGR > based. > > (It is possible that some high precision GPS receivers return coordinates > that are corrected for this, but I don't think those are true GPS > coordinates.)
Heh, indeed. Not long after posting that I was reading about OSM, WGS84 and continental drift… all a bit beyond my current comprehension, but, I’d like to do the right/best thing we can with the OS trigpoint data… if anybody would like to review and chip in ? I was actually reading up a bit as I want to work out if I need to apply a decimal place ‘filter’ to some of the co-ordinates, as I think a bunch of them have ended up as doubles that are not really showing an honest representation of the underlying data iyswim. The OS data description can be found at: https://www.ordnancesurvey.co.uk/geodesy-positioning/legacy-data/trig-search and the passage I was thinking of says (but, I think really needs to be read in context on that page…) “ The archive coordinates supplied for these triangulation stations have not been realised via ETRS89 and OSTN15. They are the original archive coordinates and can therefore no longer be considered as true OSGB36 National Grid Coordinates of the station. It is expected that agreement between ETRS89/OSTN15 derived coordinates and the original archive coordinates of triangulation stations (down to third order) will be at the 0.10 metre root mean square (r.m.s) level. “ (side note - I’ve just identified a bug that is spitting out FB numbers in the reviewnodes.osc file that are not matching what I expect - will work on finding/fixing that. Just posting here in case somebody happens to be running the code. I’ll probably start using github Issues to track such things from now on). Graham
_______________________________________________ Talk-GB mailing list [email protected] https://lists.openstreetmap.org/listinfo/talk-gb
