Closed jhrrsn closed 10 years ago
Thanks, where did the full definition come from? I have seen the towgs84 param before but haven't looked enough at Leafletjs internals to know if it was required. I guess if Leaflet transforms all points from projected co-ords into wgs84 then into map points then it could be required.
I was using the same proj4 definition via Proj4js and also via Leaflet and getting different results which I blamed Leaflet but this could explain those errors too.
Will have a look later...
My evidence is mainly experiential - it fixed the offset on the points I was plotting. It's an alternative one I found on the proj4 forums, unsure of its provenance...— Sent from Mailbox for iPad
On Fri, Mar 7, 2014 at 6:26 PM, Robert Murray notifications@github.com wrote:
Thanks, where did the full definition come from? I have seen the towgs84 param before but haven't looked enough at Leafletjs internals to know if it was required. I guess if Leaflet transforms all points from projected co-ords into wgs84 then into map points then it could be required. I was using the same proj4 definition via Proj4js and also via Leaflet and getting different results which I blamed Leaflet but this could explain those errors too.
Will have a look later...
Reply to this email directly or view it on GitHub: https://github.com/rob-murray/os-leaflet/pull/5#issuecomment-37052491
I have merged this into a new branch def_towgs84
on this repo for now...
ccb40edc85990462e4f9e078f28ae5be73281f01
Look at the 27700 projection details in QGIS, and I think you'll see the towgs values. Using the datum parameter seems to lead to an offset of ca 200m. I recommend merging this into master.
OK - I haven't had a chance to think about this for a while but the towgs84
transformation param seems valid.
So, I think this change is an improvement and seeing as this repo is a bit experimental and doubt anyone is using it in the real world* (?) - I will merge.
This seems to further improve on Bug #2. Something in the proj4 OSGB36 datum is miscalculating coordinate conversions, and using this alternative definition seems to move points to the correct location with my test cases.