Is this going to be removed?
1 Like
Thanks for raising this, Iâve reported it to the Data Working Group
3 Likes
Thank you.
Iâve found other inappropriate changesets:
Also some other useless relations have been created with random members, for example, and residential areas with postcodes for names. Additionally, Iâve seen a lot of changesets changing âSt âŠâ to âSaint âŠâ where the short version seems to be correct according to OS OpenData.
1 Like
We donât follow OS OpenData blindly.
1 Like
St is generally correct for placenames, and Saint is wrong, in British English orthography. See OpenStreetMap Help
I would suggest reverting these changesets or posting them here so someone else can do it.
2 Likes
The vast majority of the StâSaint rename has gradually been put back, maybe 100 or so along the south coast. They were well spread/hidden in the main in changesets with titles not referring to this, eg the userâs usual [placename] or â[placename] 3Dâ style changeset comments, 1 to 3 at a time with some other changes. There may be a few that other users have since touched but the level of occurence is down to that everywhere else in the country (ie some everywhere).
The misuse of relations with type=site,name=⊠(with no other tags) wrongly using relation as category is cleared up - a dozen or so groups of houses of a street (was one of lampposts of a street at one point but that was already gone when I looked). Also a couple of selections of local electrical substations (but with ambition to cover the UK?!), one of some bike rental nodes, one of some cctv cameras; all incomplete (and the intended content pretty easily retrieved using a simple Overpass query), a few odd selection errors.
The user is persuaded to move postcode from (residential etc) landuse small areasâ name= tag to postal_code= of that area, I think theyâve probably done a load of this themselves.
There were also a few highly confused uses of multipolygon relations for buildings - eg buildings with voids having building= on the multipoly outer and the multipoly used for a building:part; and vice-versa: solid buildings with building on a multipoly and building:part on the outer. I tidied a modest number of these. There is a level of âexaggeratedâ colours, and I also stumbled across some false information eg churches marked as building:material=brick when they are (centuries old) stone and I think a tendancy to over-estimate height values.
So we have a visually and internally much better and more normal map in Brighton etc than there was. We also have 1 seriously annoyed user, feeling properly victimised, but I can do little about the time they spent on going down rabbit holes of their own choosing.
I havenât got the heart to tell them that boundary=postal_code,postal_code=X with the same post code on more than one way isnât strictly right. They were just invited to shift to this (or at least postal_code= ) to get the data off name. Iâm not intending to do any more change relating to this user.
1 Like
Thanks.
Weâve still got a lot of Peacehavens, but I guess that can wait for a bit too.
Would it make life easier for you if were to avoid doing any UPRN/postcode imports in the BN postcode area for a couple of weeks?
re. postcode imports, Iâll leave that to you to review some samples of the data already present and what information you would need as input or be using for cross-checks. Probably there are a lot of super-accurately drawn and aligned individual houses and very many with postcodes already set. If you would be ignoring any postal_code values set on areas anyway, then thereâs perhaps no interaction with your UPRN match/store and postcode check/set processes.
re. the rather lively exchange on ch. 175071215, I was a little cheeky with my âor copied the name from a databaseâ line. However there is actually great evidence of which substations they actually visited and which not: the (street) name in all uppercase with source:name=signage is clearly visited; the (nearest street) name in mixed case with source:name=by_location is evidently not visited. This distinction is made in the changeset where this particular one was set; 1 was set all uppercase (with a typing error) from âsignageâ and 4 or 5 others had roads set mixed case âby_locationâ including the one on a saint-named private gated road. And there is evidence from an uncertain one (2 possible nearest and â?â in the name), one on an unnamed road (name not added), and one where someone else later did survey and corrected the name that this is the case. Which leaves the question open - where did the ref numbers come from for the ones not surveyed! - must have been a list which was matched to âby locationâ whatever that exactly means.
One other thought, there are a whole bunch of residential service=driveway with (legally dubious in most cases - there wonât be signage or gate on most of them) access=private. This in itself is pretty benign for boring short driveways, but when used with separately mapped sidewalks, this kills the connectivity between the sidewalk and road that would otherwise be delivered by driveways, with a whole load of the length of each as wrongly access=private within the bounds of the public highway. eg plenty around N part of Woodingdean eg around way 1290017223
1 Like
I was interested to see tagging of asylum seeker accommodation mentioned, as I myself tagged a couple (hopefully correctly). Please note that I made these changes solely based on mainstream news reports / council statements, so it canât be said that any private information was revealed. Certainly it wasnât accurate to continue having these locations tagged as tourism=hotel anymore. The tagging I used was amenity=social_facility social_facility=group_home and social:facility:for=migrant. Here for example is how I tagged the Bell Hotel in Epping ( Way: âȘThe Bell Hotel⏠(âȘ209568075âŹ) | OpenStreetMap ). Opinions about this are welcome seeing as @trigpoint reverted similar (speculatively tagged) changes, citing âharmful contentâ ( Changeset: 172038243 | OpenStreetMap ). Additional note for clarity: I donât hate asylum seekers, I just wanted to update these hotels that are long-term closed.
The Bell Hotel is still a hotel, owned and operated by a hotel chain and offering hotel rooms. Yes, all those rooms are currently being rented out by the Home Office, but that is a temporary arrangement.
I think you should revert that change.
EDIT: To provide a bit of rigour on that - from this article:
However, Home Office lawyer James Strachan KC told one of the hearings that block-booking rooms at The Bell Hotel âdoes not change its use as a hotelâ.
I donât think the opinion of a KC necessarily dictates ground truth but this question is at the centre of this legal battle (with Epping Council) and it was ruled that it was still a hotel.
2 Likes
You cannot book a hotel room. Just because the website is still up, doesnât make it a hotel. These are closed to the public and tourists, end employ contracted security guards. Hotel amenities such as pools and gyms are closed. Catering is run by external contractors, not the hotel chainâs previous arrangements. Such contractors that run these sites include Mears Group and Serco.
Well, temporary lasting a few years.
You are correct. Just because itâs still legally referred to as a hotel, does not make it one in the OSM sense. I cannot stress this enough, it is not possible to book a room and stay at any of the hotels used as accommodation for asylum seekers. In fact, if you read Epping council cannot challenge Home Office intervention - BBC News you will see that Somani Hotels even made a planning application in 2023 to Epping Council for a âtemporary change of useâ.
1 Like
The ground truth is that theyâre hotels. Who is staying there, and who is paying for them to stay there, is not something we map.
The ground truth is that they are closed as hotels.
3 Likes
I didnât tag who is paying for them to be there.
100,000 uses of social_facility:for would beg to differ social_facility:for | Keys | OpenStreetMap Taginfo
They are not closed as hotels, they are booked as hotels.
Look, if you take the view that these are functionally now migrant centres for entirely map-brained reasons then fair enough, but youâre also making the arguments that people with a political agenda are making. These people are not interested in the technical points - theyâre re-framing these as migrant centres to âotherâ migrants and distinguish them from normal people, as part of a wider attempt to dehumanise them (and, in the long tail, anyone who isnât white-british). If they can get people to start thinking like they do then they progress with their agenda.
I believe that your argument is flawed - hotels are booked out for lots of reasons, conspicuously mapping this reason as âno longer a hotelâ is inconsistent - but I think in this case it also matters a lot more than just on objective points. Marking these up as âno longer hotelsâ is a political act, regardless of your intent - it re-shapes how people see the world, no different to changing a border or the name of a city.
If itâs the general consensus that these are no longer hotels then fair enough, political framing should yield to our best attempts at objective truth. But I donât think that wrapping this in a 1600-node changeset marked âAdding details from surveyâ was the right way to attempt that and I think it should be changed back at least while itâs being discussed.
1 Like
Thanks for focussing on the important issues at hand
what?
Not if theyâre actually no longer hotels
When I actually see a good reason for that to happen, maybe
Isnât that the whole point of a map? It reflects the real world and helps people interpret it? We canât just ignore real-world changes??
Thanks, this is very helpful to the discussion. Trying to compare my attempt to reflect the truth neutrally and accurately with extremist ideology.
1 Like
As a very general observation, if an attempt to do something objectively and neutrally results in actions that are suitable (not intended - suitable!) to further extremist ideology, it can sometimes be a good idea to pause and ask oneself if maybe (just maybe) the pros of the objective and neutral action are outweighed by the cons.
2 Likes
Focussing on this point only, itâs an interesting point. When should an amenity that is no longer functionally offering the service of that amenity, have its tagging as that amenity changed?
By functionally, my take is that a hotel should have rooms that are bookable by the public. If this is no longer the case, because all rooms are fully booked indefinitely (and weâre obviously not just talking about the fact that itâs a busy time for the hotel and its fully booked for a couple of weeks), then it, instead, becomes private lodging run by a hotel chain.
My take seems similar to the relevant UK law:
Section 1 (3) Hotel Proprietors Act 1956 defines a hotel as an establishment held out by the proprietor as offering food and drink and, if so required, sleeping accommodation, without special contract, to any traveller presenting himself who appears able and willing to pay a reasonable sum for the services and facilities provided and who is in a fit state to be received.
(Note: my understanding is that the above is not a definitive legal definition of a hotel as such, it is more about separating hotels from inns. But, still, a useful starting point to work from.)
NAL: These âhotelsâ donât appear to meet the âwithout special contractâ test. Since they are under contract with the government for all of their rooms indefinitely, it means that âany travellerâ could not stay at the hotel.
From a map userâs perspective, if Iâm searching for hotels and this hotel is functionally never available to book, is it really useful for us to have it stored in our database as a hotel?
1 Like