Hi Jerry. “(Unparished)” is how the admin bodies (the councils) themselves refer to their unparished areas, and hence is why I call them that. Go argue with the councils if you disagree.

Please explain - explicitly - some of these many things that adding “(Unparished)” to the name breaks.

Did you notice that I added a Label to the Relation, explicitly so that the “(Unparished)” did not need to show? Please argue with whoever programs that specific feature if it is ignored.

Your statement “no admin_level=10 ... readily captured by simple queries” is complete nonsense in my experience. Months of testing in Carlton was utterly fruitless until I added an Arnold & Carlton (Unparished) relation, when it magically transformed & worked. Have you actually performed such extensive testing yourself, or is this just intellectual off-the-top-of-the-head conjecture?

I appreciate that I'm stomping in my big boots through areas that you & Will have previously mapped, but it is bound to happen, and is the reason that I left Beeston until last, so that I had built up sufficient mapping-mass to be able to cope with it all.

See osm.wiki/User:Csmale/ukboundaries (csmale is the user that has also made all OS .shape files available as .gpx downloads). Looking inside those downloads, the designation for unparished parishes is “designation=non-civil_parish”.

You say: “City of Nottingham (unitary authority) relation could be tagged as admin_level=6;8;10”. That is genius! Leave all other BoundaryLine ‘admin_level’ as single-values, but make UAs triple values. In that case, after undelete on old relation, both new 'City of Nottingham' (level8) plus 'Nottingham (unparished)' (level10) could both be deleted. Sheer genius!
 
You say: “I do recognise you have spent a lot of time tidying the geometry of boundary relations”. Truly, you have no idea. It is very much more than that. I'm trying to do a complete job & make full provision with each one. Each *addition* (many are missing) takes at least a morning/afternoon. Some have taken 2 days. Please, please, please do NOT merge BoundaryLines with other features (roads, rivers, whatever). Maintenance is utter hell if that is so.
 
PS
Is it possible just to undelete a single relation-set, or do you have to unpick the entire changeset? I do not want to lose Oscar, nor the other changes.

Hello Will. Nottingham is unique (as best I can tell) in that the identical area BoundaryLine actually contains 3 admin_levels (see osm.org/user/alexkemp/diary/39062):
1) level=6 Nottinghamshire; County (Nottm is the hole in the county)
2) level=8 "City of Nottingham"; Unitary Authority == District == Borough
3) level=10 "Nottingham (Unparished)"; Civic Parish (even though it is unparished).
Thus the sticking point between us is (2); I was certain at first that you had got it wrong, then thought that I was wrong, and now am not certain.
 
Pro your view:
See osm.wiki/Tag:boundary%3Dadministrative#10_admin_level_values_for_specific_countries
https://en.wikipedia.org/wiki/Unitary_authorities_of_England
The Wiki says UK UAs are level=6. Wikipedia says “District gained county functions 1998”. That is all compelling!
 
Pro my view:
See http://csmale.dev.openstreetmap.org/os_boundaryline/district_borough_unitary_region.html for the list of OS level=8 .shapes.
Ordnance Survey group Districts, Boroughs & UAs together. Thus, all .shape files are together (and thus .gpx are together) on the same page. In addition, all CPs (Civic Parish) are together on a different page, and that actually *includes* the unparished areas (which simply are unnamed).

It may seem perverse, but my insistence for providing a BoundaryLine for the unparished areas is that all of the search & location functions rely on the CPs at base. Thus, wise to include them, even as “Unparished”.

Reverse the deletion if you see it to be wise to do so. You will naturally need to reset the relations that are currently in place. I spent ages & ages researching it all to try to make sure that I got it right, but it looks like I may not have done so (I'm still not certain). Bugger & oops. You will need to change the name of your old relation to “City of Nottingham”. I also think that the ‘unparished’ needs to stay for the sake of search & location.
 
Ah well.

A follow-up having read the link you gave: the key words, I think, are: "it permits simpler searching" with the aspect that 'simpler'=='faster'.'I'm a 10+-year database programmer, and I would far rather have to do look-ups on a boundary-relation rather than have to create on-the-fly (and maintain) an independent in-memory database of northings&eastings-dependant names. No thank you.

In any case, I trust my personal experience better than someone else's intellectual considerations.

Hi [@Rovastar](osm.org/user/Rovastar), thanks for your interest.

My experience suggests the exact opposite, and is the reason for all the deranged additions of those tags!

In brief: my locations made to Diary entries (and also searches) were total pants in Carlton, Nottingham, Notts.. Many, many things were tried until someone made a comment that 'areas' were the issue, not nodes (such as suburbs, etc.). Separately I came across the Unparished parishes & added *Arnold + Carlton (Unparished)*. That made a big difference. Shortly after I added the `is_in` tags. That finally fixed it, big-time, but it took *both* items (admin_level=10 + is_in*).

I sure that the search- + the location-algorithm try more than one thing, but my experience says use a parish area + add a full complement of `is_in:*`.

After many talks with locals in Carlton ever since this conversation it is certain that the current location for Carlton is wrong. The current consensus seems to be the crossroads of Carlton Hill / Cavendish Road / Burton Road / Station Road. However, I will NOT touch anything until I've found an area for Carlton.

In 'Where Am I?' news, the current App appears to ignore Suburb nodes.

Just done another uplift of houses + have also changed both carlton + bakersfield nodes to ‘suburb’.

“the quality of the underlying data is by far the most important consideration” : that is irrefutable. I'm in the middle of adding houses in First Av, etc.. I'll revert the node when done with the houses. The whole thing should be up before midnight.

Does anyone know whether Bakersfield is a suburb or not? If it is, then also changing it's node would neatly fix both problems (although the software glitch needs reporting/fixing as well).

I originally thought that all city boundaries *would* be areas. I was very surprised to find nodes in the map. A quick fix of course, but not very satisfactory in the long run. As we have just found out.

Hi Will. On every recent Diary entry that I made the ‘Location’ said “Bakersfield NG22”, even though I clicked on a location on the map close to the Carlton suburb node (NG4). Checking the Bakersfield node I discovered it to be ‘place:neighbourhood’ with a Note saying “Perhaps place=suburb?”. Looking at the Thorneywood node it was also ‘place:neighbourhood’, whilst the St Anns node is ‘place:suburb’. Much earlier, I had been mystified that clicking in a map location that (to me) was St Anns would show on the Diary entry as “Thorneywood”. I did not know of the theoretical relative importances between different place values, I was basing my decision on the practical implications as demonstrated by the Map App used in Diary entries. Finally, prior to changing the Carlton node my most recent diary said “Bakersfield NG22” with the map-pin placed in Hastings Street. Making the node-change, and then editing & re-placing the pin in Hastings Street again (nothing more), it now says “Carlton, Arnold CP, Gedling...”. I fully understand your comment, but the change was based on practically-demonstrated importance & not theoretical importance. If you choose to change it back I shall try to live with my continual annoyance when the map-pin chooses Bakersfield instead of Carlton!

Hi Will

Thank goodness for that!

PPS
"Saint Bartholomew's Road" was named after (what used to be) the local Anglican church (called, er, 'Saint Bartholomew's Church') which was positioned on the north side of Bluebell Hill Road, where Wickens Walk is now. There is *very* little info on that church, even though it was only torn down in the 1970's. It looked very impressive.

I'll add 'Donkey Hill' as 'loc_name'.

PS
Bottom of Donkey Hill is a mess (junction St Anns Well Road) - bad info from commercial operators. Will be fixing it immediately I finish entering Thorneywood Rise + Gordon Road info, having collected fresh info today.

'local-' or 'alt-' may make better sense, but the fact that "Donkey Hill" shows up on the map is so thrilling (endorsing the local legend & naming) that I would wish it to stay in 'ref'.

Sorry! Other 2 shops are EAST-side of the mini-market (opposite side to town-centre).

Thanks for the changes & info.

How did you add the buildings? Did you simply trace satellite imagery? The buildings seem a long way from the road (Lou's shops butt up onto the pavement with zero gap).
As a small comment, the Mini-Market is on the extreme corner. Lou has 2 other shops next door (west side) to the mini-market, although they are intermittently occupied.

Thanks for the info on the 'name' tag.

The 'highway' tag is accurate as 'road', 'service road' or whatever (I think that the term 'close' should be added as a value to the 'highway' tag, simply because it would quickly become the predominant value chosen for English residential roads). I live close to that close & surveyed it shortly before making the change. It is NOT an alleyway.

The Close with No Name is NOT an alleyway. The modern 'Twells Close' is an 'alleyway' (pedestrian access - it has zero vehicular access). 30 years ago, the Close with No Name was a service road for (part of the modern) Twells Close and was named as such. Today it has no name & provides zero access to anything. It nevertheless is a tarmacked road with twin sidewalks. Even if the only walking that can be done is to walk in & then back out again.

There used to be houses accessed from
the Close with No Name on the West side of the street (the modern Twells Close is on the East side). All that remains of those houses on the West side is empty plots. A sturdy chain-link fence prevents egress from the Close.