The fence had been removed by late 2019, however there is still a kerb and Seething Lane should not be connected to the cycle track or carriageway of Byward Street, just the sidewalk.

https://www.mapillary.com/app/?pKey=4049675791760653

No problem. I've changed it to bridge:structure=arch.

My concern here was that the word "Dyke" in names appeared to have been treated as potentially obscene in the UK *at all*. It is the more common spelling of the word in British English and although Dike also occurs it does so only 150 times in OS Open Names compared to Dyke's 410.

https://osdatahub.os.uk/downloads/open/OpenNames

I think that bridge:structure=humpback is more appropriate for smaller bridges with a much steeper gradient than an arch-supported bridge like Blackfriars Bridge.

Possibly chosen due to ambiguity in the diagrams shown by StreetComplete?

osm.wiki/Key:bridge:structure

Correct name (Pussy Lane) restored by another user in osm.org/changeset/128547975

Could you explain how Bing was your source for this edit?

It's also on OS 25 inch maps from 1892-1914 (via NLS) and existed before Westerley Way.

https://maps.nls.uk/geo/explore/side-by-side/#zoom=18.0&lat=53.59838&lon=-1.69402&layers=168&right=osm

In the case of Lin Dike, "Dike" is the spelling used on the out of copyright OS 1:25k 1st series 1937-61 maps available to OSM via the National Library of Scotland, however that refers to the (unmapped) watercourse parallel to the Arrow Lane footpath.

Typo clean up, possibly, but not "profanity". Both spelling are common in the UK and are better checked against reliable sources like OS Open Data rather than assuming a commercial sign is authoritative.

OS Open Names contains 410 objects with the spelling "Dyke". If that offends transatlantic prudishness, hard luck.

(to Friday Street junction, not to Cannon Street)

I have also added Three Mills as a place node:
osm.org/node/10304527922

The wikipedia and wikidata tags for Three Mills had already been added to Three Mills Island:
osm.org/way/24224345

I have restored the wikidata tag for Clock Mill, which refers to that (listed) building only.
osm.org/changeset/130861112

I have added a note to objects where I have used permit, explaining that it is undocumented and needs discussion, together with a link to
osm.wiki/Talk:Street_parking

The "correct" one is private. Unfortunately, real resident parking schemes in this area do not fit the description. Permits are readily available.

Please could you check the talk page for street parking as well?

Unlikely in the near future, as other area:highway polygons don't get rendered in OSM Carto. Unfortunately, that results in the areas around some linear ways getting tagged for the renderer as e.g. highway=footway + area=yes, then some routers try to follow the perimeter.

Hi,

I think the wikidata and wikipedia tags for Three Mils might be better suited to something like a place=* node in the middle of the area, rather than replacing the wikidata entry for a single listed building within that area.

Thanks!

Changed to service in osm.org/changeset/130264945

Feel free to change the highway type back to service. I can't find out whether or not it is shared ownership or maintainable at public expense from any OSM-compatible source, so will happily follow OS Local's map style here.

You removed the USRN from Sword Close ( osm.org/way/1026019436 ) and changed it to service, which I have restored. I have no other issue with your edit and had no intention of making any other changes.

I will only apply a USRN as far as it is verifiable, which may involve splitting a way, since I can only add that identifier where it is clearly verifiable. Access to the National Street Gazetteer or OS Master Map would probably answer that, but neither of those datasets is likely to be released under OGLv3. If you would rather both ends of a T remained as residential, I have no problem with that. For a simple cul-de-sac ending in a T, we still have two ways in OSM however they are mapped.

I didn't suggest that it was arbitrary, but that different mappers choose differently. I would be more than happy to see and abide by anything which finds its way into the UK mapping guidelines.

The streets in this area which spent the last decade as highway=living_street, on the other hand, really did appear to have been classified as such rather arbitrarily.

Just to clarify, I only had an issue with the (inadvertent) deletion of a USRN, the rest was context for how I'd used the data.

I produce OpenUSRN extracts by exporting them from QGIS as GeoJSON files which I can then load into JOSM. It's a simplified geometry, so it has to be used alongside aerial imagery and the Land Registry boundaries.

From the data I've incorporated into OSM, generally the T or Y at the end of a cul-de-sac (or other short spurs) doesn't have a USRN. In some cases, that may be due to the simplification of the geometry by OS, but so far it seems to have held where the arms of the Y or T are relatively long. Where roads go from residential > service > driveway is probably down to the personal taste of individual mappers than it should be.