Mapillary images were used to determine the values of speed limit changes & Bing the position. Should have a colon not comma

It should, of course, be source:maxspeed=survey. It means that I have driven this road & noted changes in speed limit (rather than used other sources).

A quick point about the LA building import. There seem to be buildings with low height and small surface area tagged as building=residential + building:unit=2. My suspicion is that these are actually garages & that all buildings in a lot are getting the same tags. Might be worth checking. Some good egs. at junction of Crandell & West 2nd in this changeset.

Was completely mystified as to why the path to S of Rainbow didnt match my recollection, so v. glad about your changeset comment.

Is the diversion permanent?

I've changed B45 in the centre of Erback to highway=secondary. I've assumed that when you changed it to highway=service the former was what you intended.

Can I add again, the standards for admin_level assignment are clear (they were not in 2009 when this was added). You certainly should not have removed the existing relation.

The rule is tags on relation, lowest admin_level. In fact for Nottingham the highest level is 6 as it is a unitary authority without any districts (8) or parishes (10). The fact that boundary line data comes with unparished areas is more an artefact of the underlying GIS model.

Hi Alex, Please dont add (unparished) to names. It breaks all sort of things & Nottingham has never been called N* (unparished).

The fact that it is unparished merely means in OSM terms that there is no admin_level=10 relation within the relevant part of the authority. It is therefore readily captured by simple queries about what is within something else.

You seem to have set a negative width on footpath way/413746723. I imagine you meant something slightly different

I tend to agree with @fkv on all issues: 1) large scale updates often have unpredictable side effects; 2) originally we distinguished between conifers/dicots aka broad-leaved and palms (and possibly one or two others) using a range of tags wood/type etc. Ultimately these have been replaced by leaf_type. However, the palm issue (and also Gingko) highlight that leaf_type=broadleaved doesn't quite do what I think most of us expected. As @fkv nicely shows modern systematics doesn't help either. Whereas I would expect ordinary mappers to not know that Gingkos are gymnosperms; I think most should be able to distinguish palms.

To summarise: the tagging issues involved are rather more complex that initially apparent. Reversion at least would retain mapped information. Further discussion is needed on general values for leaf_type before such wholesale changes are introduced automatically.

As a fellow mapper of trees I wondered if you know of the species or taxon tags. These are more appropriate for tree types than name (especially when the tree also has a proper name). Here are some examples on one's I've done osm.org/#map=19/52.94379/-1.19020&layers=N, and an example of a named tree node/430767743. Either way keep up the good work.

EoghanM: I dont think the usage in Cambridge is drastically different from usage elsewhere. For Nottingham University there are 3 major amenity=university campuses, but also a number of discrete buildings or part of buildings (notably the Medical School & Postgraduate Medical Centre located within NHS Hospital campuses). One of these campuses, Kings Meadow, is purely an administrative site. Nottingham has additional campuses: one 10 miles away (Sutton Bonington), the other two about 6000 miles away, and the latter two are branded University of Nottingham but are actually joint venture enterprises. So a single relation isn't going to work here even to just cover the 6 campuses. The buildings not in or conveniently close enough to be mapped as exclaves of a campus are treated in ways analogous to the Cambridge situation, but buildings within campuses are not so treated; but then Cambridge doesn't really have true campuses.

I would suggest changing the tagging to break a widely used application which can be cited as an example of OSM's real-world utility is not a good idea. Particularly as the University of Cambridge spend good money in developing that app, this would primarily send a message that we care not a fig for people who use the data.

Yes I agree with will_p & SomeoneElse: the solution is to get a fix in Nominatim not mung the data. The existing places have been pretty stable for quite a number of years. Tweaking things in this way may well break other things, such as Garmin Maps based on OSM which many people (SomeoneElse & I, for instance) use all the time.

Hmm, when I visited I needed a permit from SNH or its predecessor (mind you the only way off the ferry was by dinghy). The island is effectively still a private estate (whether in hands of Bulloughs, NCC, SNH or a Community Trust) and should be mapped in a manner consistent with similar estates on the mainland.

See this statement on the IRCT website (http://www.isleofrum.com/contactandlinks.php) which clearly states that the road is private:

Q. Can I take my car to Rum?
A. No. The roads in Kinloch Village are privately owned and managed by the local community. In line with the other Small Isles, visitors to Rum are not permitted to bring vehicles except under exceptional circumstances. People generally leave their cars parked at Mallaig or Arisaig while they visit Rum.

I've changed the road to service as I don't believe it is maintained by local authority. As you have been there more recently than me you may know better, in which case revert my edit.

Is the White House flagpole really called William Estate? I can find no online references to this name.

Welcome to OpenStreetMap. You may be interested to know that the tags which you have added as a single point (node) could equally be added to the existing area marked as landuse=industrial. In this case the name would show up directly. The address information & name you have added is very useful anyway (& will be searchable too), but will not show directly on the map.

A couple of other tags might be of interest too: industrial=scrapyard, amenity=recycling and recycling:scrap_metal=yes. These help separate different kinds of industrial areas.

I wonder if you discussed this import before carrying it out. It appears to have lots of duplicate nodes and areas labelled as forest which are clearly small groups of individual trees.

You never added the name! (done now)

It would be better if both of you respected the judgement of two very experienced OSM bicycle mappers who both regard this as route=mtb. See the mail by Richard Fairhurst to talk-us about this https://lists.openstreetmap.org/pipermail/talk-us/2016-May/016206.html, and the comment by Simon Poole on OSM Help.

Both names seem to be old names of the former CP of Wanstead. I seriously doubt if they have any relevance now http://www.british-history.ac.uk/vch/essex/vol6/pp317-322