Hi,
It's a particularly garbled tag error caused by my macro keyboard occasionally going faster than JOSM can keep up with!

Thanks for spotting the error tag - both should be fixed by:
changeset/130507290
Happy Mapping,

James

Hi user,
It's a cut-and-paste error into JOSM - thanks for spotting it. Dumb tag removed!
Happy Mapping,

James

Hi Emilius123,
You talked me into it... :-)
changeset/117915598

PS My survey imagery is now processed if you're curious (several new homes have sub-addresses suggesting the plan changed):
https://www.mapillary.com/app/?pKey=656061858932333
Happy mapping,

James

Hi again GinaroZ,
That's a pretty crap bit of copy-pasta from 130m to the SE!
I can remember unpicking a nest of ways joining the footway to the fire_station, so must have got over excited with CTRL-C CTRL-V.

A good spot - thanks!

Fixed:
changeset/117885985

Hi Chris,
It's unfortunate that you've not replied - I'll assume your SPAM folder has eaten the reminder emails.

Open Street Map is a project created by a community, so communication is important to allow us to work collaboratively to improve the overall map.
After checking Mapilliary imagery, I can see no reason for the change in highway type for a whole 68m, so have changed back to highway=trunk:
https://www.mapillary.com/app/?pKey=295053742265197

I'd suggest being careful when accepting editor defaults - they may well be plain wrong, or unnecessary (e.g. adding motor_vehicle=yes to a highway=trunk - not needed, it's a trunk road!).

changeset/116818894

Happy Mapping,

James

Hi Mountain Monkey,
Can I ask that as well as splitting forest relations, that you also check heath relations as well please?

This change removed a chunk of relation/13417952 in the NE corner near Rowantee Hill ( changeset/116734954#map=14/55.3243/-3.0160&layers=N) , meaning a large area disappeared:
relation/13417952/history#map=11/55.2439/-2.8715&layers=N

The relation will now show as continuous as I've created a new way and added it to the heath relation, and JOSM shows the outer role as being a full loop again:
changeset/116734954#map=14/55.3243/-3.0160&layers=N

Unfortunately, given the number of HUGE landcover relations in this area, and the way they are intertwined, this is ridiculously difficult to avoid.
I know - I've fallen foul of similar excessively large relations, and it took me a couple of days to find, split into (slightly) smaller relations, and fix. :-(
Happy Mapping,

James

Hi Chris,
Can I check why the A7 now goes trunk-Primary-Trunk please?

Is there a material difference between these two segments, other than the bus status, that downgraded the second from trunk to primary?
way/33650701/
way/722061473/
Thanks and Happy Mapping,

James

Hi Cleveland steamer,
That's correct - the whole landuse was marked building=fire_station. I removed the building=* tag from the non-building landuse.
Only the building is a building.
Happy mapping,

James

Hi mueschel,
Thanks for an interesting point - I spent time thinking about the namespace for the reference (e.g. maxspeed:type=*), but indeed missed *uppercase* ISO alpha2 country code.
Fixed - way/1003043215/h
Happy Mapping,

James

@kawino85: In the OpenStreetMap community our tagging preferences are flexible, and deliberately NOT set in stone. Often, the best way to understand how to tag a feature is to see how others in the community have modelled similar situations.

I've been contributing to OSM since 2007 so am still learning, so find tools like taginfo useful to look for examples e.g.:
https://taginfo.openstreetmap.org/tags/man_made=wastewater_plant#combinations
The variation in OSM feature tagging may create complexity for data consumers, however the same variation adds richness and detail to our project. Indeed, this is the same reason why "mechanical editing" is discouraged - running what looks like a simple SQL update may erase "ground truth".
Adding tags is usually better than deleting those added by others - e.g. adding utility=sewerage, even a industry:type=* may be useful even if a data consumer may infer from man_made=wastewater_plant.
In addition, "ground truth" is important here as although this facility is a Waste Water Treatment Works (WWTW), the name=* reflects the sign at the gate - which suggests man_made=pumping_station.
Happy Mapping,

James

Hi,

Clean water.

After working as a chartered engineer for a large water company, the process of processing of raw sewage is most definitely industrial. Lift pumps are 2m tall. Screens use chain conveyors to remove detritus up 5m inclines. The whole process is packaged chemical engineering - very much industrial.
And more to the point, I've stood outside the gates of _this_ very plant and listened to the sound of all the automated industrial processes.
Happy Mapping,

James

Hi Dave,
That's a very good question, and frankly, I've only got conjecture.
The imagery shows a rectangular area of concrete/ tarmac along the permanent way of the old St. Boswels & Reston railway, so I decided to tag it as a platform. Shadows suggest it's too low to be a building and it looks old with lots of overgrowth.
The NLS OS 1:10k and the more detailed 25-inch shows two tracks leading into Reston Junction, but nothing else. The NLS OS 6-inch says "Trough" a bit West, suggesting a water supply, but would expect engines to bunker on the main line depot?
It'd take a ground survey to confirm, but what ever it is, it seems too closely aligned to the track formation not to be railway.
The Reston Station construction drone footage looks North away from this area (and even if it were licensed, is a bit oblique to be useful).
Any ideas?
Happy Mapping,

James

Hi GinaroZ,
Wow - that TMS is a _lot_ higher resolution! :-)
The data matches what I saw pretty well, so will add in the connecting pieces of the footpaths.
BTW - the Mapilliary imagery of the bridge is now processed and available here:
https://www.mapillary.com/app/?pKey=1044543182947517

Thanks again, and Happy Mapping!,

James

Hi Matthew,
We've all gone back to edits and wondered how on earth the tags got there!
Thanks for improving the map,
James

Hi MPaine,

Just for information, highway=path regularly confuses as it is for a general use trail (e.g. horses, cycles, just not vehicles), and not just a footpath.

Some of these ways have horse=no, suggesting they could be UK Footpaths, rather than UK Bridleways, so could be better as highway=footway with designation=public_footpath.

More info to help you consider the best option is here:
osm.wiki/Access_provisions_in_the_United_Kingdom#Public_Rights_of_Way

Happy Mapping,

James

Hi MPaine,
Hello, and welcome to OSM!

Can I ask you to review this edit as you have set the _entire_ length of this unclassified highway as a bridge, not just the bridge itself!

You've done the hard part by splitting the bridge into a separate way, so I suggest you need to remove both bridge=yes and layer=1 from ways:

way/112184262/

way/952313225/

Thanks, and Happy Mapping.

James

(Bing offset = 0.97; -2.61)

Hi GinaroZ,
Looks like cadastral (likely RoS) data is available in Duns, but not around the estate example as it's likely never been sold.
Duns centre seems to be within 1-1.5m (OSM buildings - OSMUK Cadastral Parcels) with OSM about 1.2m NNW.
To compare notes, are you seeing the same please?
Thanks,
James

Hi GinzroZ,
Yes - the discussion on talk-gb was very interesting (down to geoid transformations!).
The challenge (as ever) is the Borders have a lot of high-res detail mapped well by ctait to a different baseline. Better aligning one feature rapidly spreads into unpicking unusual building/ landuse/ highway area alignments.
'it's hard' isn't a reason not to do it though - I'll have another look!
Happy Mapping,
James

After reading again recent Talk-GB discussions, I wonder if Mike Collinson may have touched on the root of this changeset and wider debate:
https://lists.openstreetmap.org/pipermail/talk-gb/2021-April/026626.html

Better understanding your desired outcome (here or over on Talk-GB) and level of detail you aspire to would help us all improve the map. Are you aiming for cadastral level, with residential only for occupied plots? What goes in the gaps? (lots of landuse=grass for these pre-war geometric estates?)

As mentioned in my last note, Stakeford was (ISTR!) initially mapped before imagery. One surrounding landuse=residential resulted from the low-resolition geolocated data available (GPSr on a bike and possibly landsat).

Okay - we've got high-res imagery and even land registry parcel data (partially) available now, so more detail is achievable. So how far do we go on this next pass?