Re: [Talk-gb-westmidlands] Undeletion needed

[Alan Mackie]

If it's that recent then finding the changeset will probably be easiest
through osmcha.org filtered by your username and a rough bounding box.

Alternately as you may know a time and narrow bounding box where it
definitely existed you might be able to pick the deleted way from and
Overpass Turbo query along these lines:
https://wiki.openstreetmap.org/wiki/Overpass_API/Overpass_API_by_Example#OSM_data_at_a_certain_date
It may be best to run this from within JOSM as the web interface warns that
there's quite a bit of data and can be awkward about selecting items that
are "under" other areas.

Once you have a changeset or way ID you can use either JOSM's reverter
plugin or its undelete plugin to resurrect the relevant item. With the
reverter plugin it is usually possible to selectively upload specific
deleted ways rather than the whole changeset with the usual upload
selection tool. For a web based tool I think level 0 can be used to restore
old versions, but I've never used it in anger as I tend to prefer launching
a GUI.

- Alan


On Mon, 8 Nov 2021, 10:24 Andy Mabbett,  wrote:

> Mea culpa: a couple of weeks or so ago, I deleted a building (four
> terraced houses) at Perry Barr, around 52.515343, -1.900919, thinking
> it had been demolished; but I misremembered what I had seen as  drove
> past and the building (unlike those to its north and west sides) is
> extant
>
> Can someone kindly talk me through how to find and revert the edit?
>
> --
> Andy Mabbett
> @pigsonthewing
> http://pigsonthewing.org.uk
>
> ___
> Talk-gb-westmidlands mailing list
> Talk-gb-westmidlands@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk-gb-westmidlands
>
___
Talk-gb-westmidlands mailing list
Talk-gb-westmidlands@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-gb-westmidlands

Re: [Talk-GB] UK street addressing

[Alan Mackie]

I struggle with what to call the  in that example.

A recent suggestion for named terraces was to use addr:street=
and addr:parentstreet=, but if the  relates the
whole building to to parentstreet, then reconstructing an address seems
impossible.

The closest existing tag seems to be add:housename=, but I
don't know if that stretches the definition too much.

On Mon, 21 Dec 2020 at 06:41, Peter Neale via Talk-GB <
talk-gb@openstreetmap.org> wrote:

> At the risk of throwing another edge case into the pot (and mixing
> metaphors), can I ask how I should tag our flat?
>
> The Post Office Official postcode checker renders it as:
>
> Flat 
> 
>  
> 
> 
>
> where  refers to the whole block and is common to all the
> flats.
>
> I cannot see what the Post Office is calling the various data fields, but
> I assume OSM would be happy with (taking elements from above)
>
> addr:housenumber=
> addr:street=
> addr:city=
> addr:postcode=
>
> That just leaves me to deal with the "Flat" element.
>
> Consulting the Wiki, I THINK I can cover that with:
>
> add:flats=  (for one specific flat)
>
> ...or addr:flats= (for the whole block)
>
> However, I unsure whether to include the word "Flat" in the value field of
> "addr:flats=*", or not.  The Wiki page for Key:addr includes, as an
> example, "addr:flats=Suite 110A", which seems fine for a single living
> space unit.  It could be called "Flat 110A", "Suite 110A", "Apartment
> 110A", etc., so including the descriptor word could be useful to the data
> consumer.  However, the Wiki page for Key:addr:flats shows only numeric
> values.
>
> TagInfo shows 203.5k uses of "addr:flats", but only 38 uses of
> "addr:flats=*flat*" and 42 uses of "addr:flats=*suite*", again suggesting
> that only the unique value(s) (e.g. "1", "2", "13B", etc.)  are
> sufficiently used to warrant data consumers catering for them.
>
> So, should I omit the word "Flat", "Suite", "Apartment" etc., leaving the
> data consumer to guess (or to default to "Flat...")?
>
>
> Regards,
> Peter
>
>
> On Monday, 21 December 2020, 09:30:37 GMT, Robert Whittaker (OSM lists) <
> robert.whittaker+...@gmail.com> wrote:
>
>
> Like it or not, in the UK addresses are defined by Royal Mail. They're
> introduced the concept of a "postal town", and this is one of the few
> common elements that each address must always have. Once you accept
> that the Post Town is intended to be a nearby significant place (to
> help with delivery routing and identifying the rough location of the
> addressed property) rather than being a place that the address is
> "in", then it's really no more of a fiction than the postcode. (The
> village I grew up in had a GL postcode, despite it being in
> Worcestershire. I've currently got an IP postcode, despite being in
> Norfolk and closer to Norwich (NR) than Ipswich.)
>
> On the basis that it's a required part of each address, I would
> recommend that we do store the post town in OSM addresses. There are
> significant advantages to storing it in a consistent way, and the best
> existing tag to do this would be addr:city. (We wouldn't want to
> invent a new tag (e.g. addr:posttown), since as a UK-only term that
> will simply be ignored by most international data consumers.
>
> We then have a possible hierarchy of named localities between the
> street and the post town to record as part of the address. I would
> suggest using appropriate values from the set {addr:hamlet,
> addr:village, addr:town, addr:suburb}. (I don't see any other
> alternatives to this.) Most of these key names already have a
> reasonable number of uses in OSM (addr:town is the lowest, but that
> still has 59k uses), so it seems that others are doing this too.
>
> Regarding properties (e.g. on named terraces or sub-streets), where
> there are two street names (Thoroughfare and Dependent Throughourfare
> in Rail Mail terminology) then we need a second key to store the other
> street name under. Certainly if there is an addr:housenumber or
> addr:housename, I think we need to use addr:street for the
> street/terrace name on which that name or number applies. Otherwise,
> software that's unaware of the second key name will think it's house
> number n on the main street not the sub-street. There are already
> about 3.5k uses of addr:parentstreet in OSM, so I'd recommend using
> that for the main street, and addr:street for the terrace or
> sub-street name. If any data-users aren't aware of addr:parentstreet
> it's not a major issue, since it will still pick up the correct
> terrace/sub-street name, and the locality, which will probably be
> enough to use as an address.
>
> I would strongly argue against using addr2 in connection with
> sub-streets, as it's not standardised, and is likely to not be picked
> up by any software. There's an abondoned proposal at
> https://wiki.openstreetmap.org/wiki/Proposed_features/addr2 , but that
> was for the case of a single property on a street corner having two
> formal addresses, one on each street, 

Re: [Talk-GB] UK street addressing

[Alan Mackie]

That still doesn't answer the addr: tag question.

I don't think we'd normally expect consumers to need to do such detailed
geometry parsing for address to interpretation.

I think we need a firmer scheme for divisions of privately managed stuff
e.g. business parks apartment 'complexes' and the like would be useful.

On Sun, 20 Dec 2020, 13:45 ipswichmapper--- via Talk-GB, <
talk-gb@openstreetmap.org> wrote:

> What you do is give the outline way "buildong=terrace" and
> "name=" and all the houses with "building:part=house". The
> software can then tell that all those houses are part of the terrace called
> 
>
> --
>
>
> 20 Dec 2020, 17:30 by colin.sm...@xs4all.nl:
>
> On 2020-12-20 18:21, ipswichmapper--- via Talk-GB wrote:
>
> Tag the houses with addr:place maybe?
>
>
> IMHO a house is not a place
>
>
> Or, better method is to use the alternative terrace taggong scheme where
> each house is tagged as building:part=house within a larger
> building=terrace.  (Terracer plugin lets you do this if you check "keep
> outline way")
>
>
>
> That allows the building to be split into parts, but does it tell us how
> to put a distinct address on each part?
>
>
> ___
> Talk-GB mailing list
> Talk-GB@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk-gb
>
___
Talk-GB mailing list
Talk-GB@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-gb

Re: [Talk-GB] UK street addressing

[Alan Mackie]

I'm also unclear how to tag numbered houses in named terraces.

addr:housename doesn't seem appropriate if they are shared along an entire
row and addr:street already has a value.

I've also run into this for blocks of flats. "Block B" doesn't seem like a
housename either? The addr:block tags seems to be for named city blocks.

Do we have some sort of local grouping tag?

On Sun, 20 Dec 2020 at 10:32, ndrw  wrote:

> On 20/12/2020 12:45, Dave Abbott wrote:
> > There is a page at
> > https://wiki.openstreetmap.org/wiki/User:Rjw62/UK_Address_Mapping
> > which mentions "suggested tags" but there is no evidence that this is
> > in use. If correct I would be tagging as -
> >
> > addr:housenumber=99
> > addr:street=Postal Street
> > addr:town=Smalltown
> > addr:city=Largertown
> >
> This is correct, although there is no consensus wrt to the tag used for
> Smalltown. I'm using one of addr:villlage|suburb|town myself. There was
> a proposal to switch to addr:locality only, which I argued against in
> the past, but it would indeed match RM addressing better and often
> classification of the locality is unclear.
>
> This is not the only problem with RM<->OSM address tagging. RM defines
> following address structure:
>
> Dependent thoroughfare
> addr:place (?)
> Thoroughfare
> addr:street
> Double dependent locality
> addr:hamlet|district (?)
> Dependent locality
> addr:town|village|suburb|locality (?)
> Post Town
> addr:city
> Postcode
> addr:postcode
>
>
> This often becomes an issue when mapping business parks,
> hospital/university campuses etc.
>
> ndrw6
>
>
>
> ___
> Talk-GB mailing list
> Talk-GB@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk-gb
>
___
Talk-GB mailing list
Talk-GB@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-gb

Re: [Talk-GB] driveway-becomes-track

[Alan Mackie]

On Sun, 13 Dec 2020 at 11:14, Mateusz Konieczny via Talk-GB <
talk-gb@openstreetmap.org> wrote:

> I run into from time to time and was unsure how to tag this.
>
> On the other hand highway=track is supposed to be used on
> roads used to access fields/forests (often unpaved and of low
> quality, but there are also high quality asphalt
> tracktype=grade1 surface=asphalt ones).
>

I think I have said this previously, but this feels like a very "old world"
viewpoint to me (and yes I do know which list this is). Lexico's first
definition [1] more closely agrees with my feeling on the matter, but over
time the OSM wiki seems to have been edited progressively further away from
this usage.

In countries with a longer history, farming and forestry may be the last
bastions for this sort of "improvised but then improved as necessary" road,
but to my mind it is still counterintuitive to define tracks as "for
farming and forestry work". I would tend to say that the higher grades of
track may be more appropriately tagged as service roads, especially if they
look like someone's gone along and reworked them to be less 'opportunistic'
at some point. Yes, they serve a farm rather than a warehouse, but they are
still largely the same function.

[1]: https://www.lexico.com/definition/track


> So with road that is both access road to single house and
> forest neither highway=track nor highway=service service=driveway
> really matches.
>

I think it's a service road, drop the 'driveway' part for the multi-use
portion.


> Dec 13, 2020, 11:44 by talk-gb@openstreetmap.org:
>
> IMHO, if it leads on to another road, track, etc. it is not a "driveway",
> but could be a track, a bridleway, a service road, or something else.
>
> The Wiki says that a driveway is (with my bold for emphasis),
>
> " ... a minor service road leading to a residential or business property.
> It typically branches from a bigger road and leads toward an entrance to a
> specific destination (building, etc.). It may end at or pass the entrance,
> but either way, it gets close to its destination. *It is rare for a
> driveway to be the way to access another roadway (but see Pipestems
> below)."*
>
> (pipestems allow a driveway to be shared between several properties)
>
> So if, in this case, it leads on to another way (e.g. a bridleway, or a
> track), it is not a driveway.  Does this solve the problem?
>
> Regards,
> Peter
>
> Peter Neale
> t: 01908 309666
> m: 07968 341930
> skype: nealepb
>
>
> On Sunday, 13 December 2020, 10:25:46 GMT, Edward Bainton <
> bainton@gmail.com> wrote:
>
>
> Sorry, I joined this thread late and I see the initial query was, How to
> ensure tracks don't just pop up nowhere'. So driveway first then track
> doesn't solve the problem.
>
> That makes me say track all the way, as someone else has said. The
> different surfaces can be caught in the attributes.
>
> On Sun, 13 Dec 2020 at 10:08, Edward Bainton 
> wrote:
>
> >  https://85a.uk/noverton_farm_1280x800.jpg
> >
> > It seems daft to me that the mud gets rendered but not the hardcore. If
> > I change the "driveway" to "track" that would be the dreaded tagging for
> > the renderer would it not? Generally in this part of the world "track"
> > means mud, rather than a roadway suitable for all vehicles.
>
> I don't know what part of the world you're in, but by my Fenland lights,
> I'd probably call that a track, not a driveway - certainly once it passes
> the farm buildings (since I see a driveway as implying car-worthy access to
> a building).
>
> Would that solve it? Driveway as far as the farm and then track?
>
> I'm going to risk blasphemy and suggest that tagging for the renderer is
> what we all do, all day (or why map?). The problem imo is "fudging it for
> the renderer", or "outright lying for the renderer". In this case, I'd say
> track is a valid choice - I think even for the whole length, if by
> "driveway" we infer something, short, tidy, and suburban.
>
> But I'm still a spring chicken round here, relatively speaking, and I
> await correction by my olders.
>
> On Sun, 13 Dec 2020 at 09:09, Nick Whitelegg via Talk-GB <
> talk-gb@openstreetmap.org> wrote:
>
> >Getting back to this case, this is the farm drive. Beyond the
> >cattle-grid the public bridleway continues left through the farm
> >buildings, and the surface deteriorates to the usual farm mud:
>
>   >https://85a.uk/noverton_farm_1280x800.jpg
>
>
> Apologies for going off topic, but I knew that name (Noverton Farm)
> sounded familiar.
>
> A quick check of where it is would explain why. In 1998 I did a  long
> distance walk from Sussex to the Peak District, following ordinary
> footpaths (planned using OS maps) and went through this area, the Teme
> Valley. It was very nice *but* the footpaths were in an appaling state of
> disrepair, I remember on several occasions that day having to scramble
> through dense shrub cover and attempt to negotiate barbed-wire fences. I
> seem to recall Noverton Farm as being the 

Re: [Talk-GB] Mapping a building that's two connected separate buildings

[Alan Mackie]

On Mon, 12 Oct 2020 at 18:58, Jez Nicholson  wrote:

> Ah yes, a bit like when a hospital or school has a 'corridor room' (for
> lack of a better term) joining two separate buildings. I'd go for three
> joined buildings myself.
>
> And that newer building has been extended a bit more hasn't it? That part
> I would merge with the existing building.
>
>
I'd probably map as three joined building and transfer the hotel tags to a
polygon that surrounds buildings, parking and grounds.

On Mon, Oct 12, 2020 at 5:56 PM Mateusz Konieczny via Talk-GB <
> talk-gb@openstreetmap.org> wrote:
>
>> It sounds like three connected buildings,
>> but one building with three building:part
>> areas also would be acceptable
>>
>>
>> 12 paź 2020, 18:52 od m...@good-stuff.co.uk:
>>
>> I was looking at tidying up a few things around my local area, and came
>> across this:
>>
>> https://www.openstreetmap.org/#map=19/52.08855/-1.94195
>>
>> What you can see there is a building labelled "Evesham Hotel" (which is
>> correct), and, just to the south-west of it, another, unlabelled building.
>>
>> However, look at the aerial view (eg, via the edit feature, although
>> Google Maps will do just as well), and it's clear that there is a link
>> building connecting the two (something which I can confirm from local
>> knowledge):
>>
>> https://www.openstreetmap.org/edit#map=19/52.08855/-1.94195
>>
>> (There's also an unmapped extension to the bottom left building, but
>> that's another matter).
>>
>> That's because, many years ago when the manor house was converted to a
>> hotel, the owners expanded the hotel by building the link to the adjacent
>> building so that it's all one building internally (more of the
>> accommodation is in the bottom left building, the original manor house is
>> mostly reception, function and dining rooms and associated non-public areas
>> such as kitchens and offices).
>>
>> So, how should this be mapped? Should the entire hotel, covering both
>> original buildings and the later link building, be mapped as a single
>> polygon? Or should they be mapped as three adjacent, but separate,
>> polygons? Is there a standard way of approaching situations like this?
>>
>> Mark
>>
>> ___
>> Talk-GB mailing list
>> Talk-GB@openstreetmap.org
>> https://lists.openstreetmap.org/listinfo/talk-gb
>>
>> ___
>> Talk-GB mailing list
>> Talk-GB@openstreetmap.org
>> https://lists.openstreetmap.org/listinfo/talk-gb
>>
> ___
> Talk-GB mailing list
> Talk-GB@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk-gb
>
___
Talk-GB mailing list
Talk-GB@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-gb

Re: [Talk-GB] New Bing Imagery

[Alan Mackie]

On Wed, 19 Aug 2020 at 14:52, Jass Kurn  wrote:

>
> On Wed, 19 Aug 2020 at 13:58, Colin Smale  wrote:
>
>> Possibly even better that StreetView imagery is data that has been
>> imported directly from OS, such as OS Boundary-Line for the admin
>> boundaries. This is probably the closest we can get to cm-level accuracy -
>> even though they don't give us the full resolution, the base points such as
>> tripoints where boundaries meet are likely to be pretty damn accurate. I
>> would recommend using these as a kind of calibration point to sanity-check
>> imagery alignment and other data based on less accurate GPS positioning
>> (e.g. from any consumer-grade GPS kit).
>>
> I've been coincidently wrestling with this issue of offsets for the last
> two days. New Bing imagery is resulting in very detailed and useful mapping
> (e.g. solar panels) but imagery is nearly always out by a problematic
> amount. I also feel the best source for offsets is not Streetview, but the
> vector data from OS OpenData. I've done some experimenting over the last
> two days, and my favourite source for alignment is "OS OpenData Local -
> Vector". Within those downloaded files is a data set called
> "FunctionalSite" which is primarily the boundaries of educational sites.
> They're excellent for alignment because the file is not too big, school
> sites are common, and the boundaries are commonly thin fences which are
> easy to align with Bing imagery, bringing errors to a trivial amount.
>
> I think the secondary useful task is to make offsets available using the
> "The Imagery Offset Database". It's been around for a long time, but is now
> way more useful due to the issues being discussed.
>
>  I think in an ideal world it would be good for the offset to be based on
a record of features detected within the imagery itself. If we were able to
generate these repeatably, then it may be possible to use the same control
points across different imagery and have a chance of self-correcting after
updates. In the real world this is probably too computationally expensive
for OSM editing software, even for the relatively "simple" features used in
panorama stitching software. I also have no idea if the license on the
imagery we have access to would allow this.


> With imagery likely to become more detailed, with more high quality
> tracing of the high quality but misaligned imagery, I think we'll need a
> more formal approach to "Align imagery before tracing guidelines"
>
> Jass
>
>
> ___
> Talk-GB mailing list
> Talk-GB@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk-gb
>
___
Talk-GB mailing list
Talk-GB@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-gb

Re: [Talk-GB] New Bing Imagery

[Alan Mackie]

On Wed, 19 Aug 2020 at 10:51, Alan Mackie  wrote:

>
>
> On Wed, 19 Aug 2020, 10:13 Stephen Colebourne, 
> wrote:
>
>> So, I followed the links below and added an offset. But this simply
>> isn't a viable solution to the problem because it only works for JOSM
>> and not iD.
>>
>> I managed to convince one mapper to type in the offset manually in iD
>> every time, but that is a horrible thing to ask new mappers to do,
>> very offputting. And now I can see Amazon mappers using an iD variant
>> that doesn't have the offset and moving all the roads as a result:
>>
>> https://osmcha.org/changesets/89549551?aoi=758c7f2b-faca-44e5-acd2-0cb8c33034bd
>>  https://www.openstreetmap.org/changeset/89549551
>> This is going to keep happening so long as OSM has multiple image
>> sources and multiple editors. Frankly I'm amazed that this isn't a
>> solved problem.
>>
> Imagery isn't one image. There are carefully blended seams that often
> include jumps if you look closely. There is also no guarantee that
> distortion to "flatten" hills etc. We also tend to receive no metadata
> about where the seams are, or how they were distorted. It is not an easy
> problem to solve.
>
Sorry the above should have read:
  Imagery isn't one image. There are carefully blended seams that often
include jumps if you look closely. There is also no guarantee that
distortion to "flatten" hills etc. *has resulted in a consistent offset*.
We also tend to receive no metadata about where the seams are, or how they
were distorted. It is not an easy problem to solve.

>
>

>> Having done some mapping across the country recently, it seems like
>> Bing is offset to the previous best imagery across the country, but by
>> varying amounts. Is there really no solution that can be applied to
>> the source Bing layer? Or should we all just accept Bing as golden?
>>
> I think we tend to accept OS's StreetView as the best source for position.
>
>>
>> Having added thousands of buildings and fixed roads to align to the
>> previous best imagery, I don't have a good solution to the problem,
>> and it is demotivating to think that others are going to come along
>> and move individual roads/buildings to align without considering the
>> bigger picture.
>>
> Poorly considered hit and run mapping will happen regardless of the
> imagery.
>
> At least talk-gb doesn't have to deal with those with "humanitarian"
> inspiration. These can be prolific and sometime give the impression they
> are too busy "helping people" to read even basic instructions. (Especially
> the ones who go off piste)
>
>>
>> The only solution I can think of is to move all nodes in the area I've
>> worked on to match the new Bing (ie a mass edit). Any other
>> suggestions?
>>
> Readjusting every time a source updates is not a good solution. It assumes
> there is one "authoritative" source and would become and endless task.
>
> I would like it if the iD 'adjust imagery tool' button were more visible.
>
>>
>> Stephen
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>> On Sun, 12 Jul 2020 at 23:36, Mateusz Konieczny via Talk-GB
>>  wrote:
>> >
>> >
>> https://wiki.openstreetmap.org/wiki/JOSM/Plugins/Imagery_Offset_Database/Quick_Start
>> > https://wiki.openstreetmap.org/wiki/Imagery_Offset_Database
>> > (I think that nowadays it is built in - is plugin installation still
>> necessary?)
>> >
>> >
>> > No idea about iD support -
>> https://github.com/openstreetmap/iD/search?q=imagery+offset
>> >
>> > Jul 13, 2020, 00:21 by scolebou...@joda.org:
>> >
>> > Wow, the imagery is really good. But in my area the imagery is about
>> > 3-4m east west and 3-4m north south out of alignment with Esri World
>> > Imagery (Clarity) Beta, which is what I've been using up until now
>> > (for thousands of buildings).
>> > https://www.openstreetmap.org/#map=18/51.39886/-0.24940
>> >
>> > Is there any way to unify the alignments?
>> >
>> > Stephen
>> >
>> >
>> > On Thu, 9 Jul 2020 at 06:41, Gareth L  wrote:
>> >
>> >
>> > I’ve noticed patches of vastly improved bing imagery since December,
>> but it is really patchy.
>> > Gareth
>> >
>> > > On 6 Jul 2020, at 23:21, Cj Malone <
>> me-osm-talk...@keepawayfromfire.co.uk> wrote:
>> > >
>> > > I was splitting houses in Portsmouth/Southsea this morning. The
>> imagery
&

Re: [Talk-GB] New Bing Imagery

[Alan Mackie]

On Wed, 19 Aug 2020, 10:13 Stephen Colebourne,  wrote:

> So, I followed the links below and added an offset. But this simply
> isn't a viable solution to the problem because it only works for JOSM
> and not iD.
>
> I managed to convince one mapper to type in the offset manually in iD
> every time, but that is a horrible thing to ask new mappers to do,
> very offputting. And now I can see Amazon mappers using an iD variant
> that doesn't have the offset and moving all the roads as a result:
>
> https://osmcha.org/changesets/89549551?aoi=758c7f2b-faca-44e5-acd2-0cb8c33034bd
>  https://www.openstreetmap.org/changeset/89549551
> This is going to keep happening so long as OSM has multiple image
> sources and multiple editors. Frankly I'm amazed that this isn't a
> solved problem.
>
Imagery isn't one image. There are carefully blended seams that often
include jumps if you look closely. There is also no guarantee that
distortion to "flatten" hills etc. We also tend to receive no metadata
about where the seams are, or how they were distorted. It is not an easy
problem to solve.


> Having done some mapping across the country recently, it seems like
> Bing is offset to the previous best imagery across the country, but by
> varying amounts. Is there really no solution that can be applied to
> the source Bing layer? Or should we all just accept Bing as golden?
>
I think we tend to accept OS's StreetView as the best source for position.

>
> Having added thousands of buildings and fixed roads to align to the
> previous best imagery, I don't have a good solution to the problem,
> and it is demotivating to think that others are going to come along
> and move individual roads/buildings to align without considering the
> bigger picture.
>
Poorly considered hit and run mapping will happen regardless of the
imagery.

At least talk-gb doesn't have to deal with those with "humanitarian"
inspiration. These can be prolific and sometime give the impression they
are too busy "helping people" to read even basic instructions. (Especially
the ones who go off piste)

>
> The only solution I can think of is to move all nodes in the area I've
> worked on to match the new Bing (ie a mass edit). Any other
> suggestions?
>
Readjusting every time a source updates is not a good solution. It assumes
there is one "authoritative" source and would become and endless task.

I would like it if the iD 'adjust imagery tool' button were more visible.

>
> Stephen
>
>
>
>
>
>
>
>
>
>
> On Sun, 12 Jul 2020 at 23:36, Mateusz Konieczny via Talk-GB
>  wrote:
> >
> >
> https://wiki.openstreetmap.org/wiki/JOSM/Plugins/Imagery_Offset_Database/Quick_Start
> > https://wiki.openstreetmap.org/wiki/Imagery_Offset_Database
> > (I think that nowadays it is built in - is plugin installation still
> necessary?)
> >
> >
> > No idea about iD support -
> https://github.com/openstreetmap/iD/search?q=imagery+offset
> >
> > Jul 13, 2020, 00:21 by scolebou...@joda.org:
> >
> > Wow, the imagery is really good. But in my area the imagery is about
> > 3-4m east west and 3-4m north south out of alignment with Esri World
> > Imagery (Clarity) Beta, which is what I've been using up until now
> > (for thousands of buildings).
> > https://www.openstreetmap.org/#map=18/51.39886/-0.24940
> >
> > Is there any way to unify the alignments?
> >
> > Stephen
> >
> >
> > On Thu, 9 Jul 2020 at 06:41, Gareth L  wrote:
> >
> >
> > I’ve noticed patches of vastly improved bing imagery since December, but
> it is really patchy.
> > Gareth
> >
> > > On 6 Jul 2020, at 23:21, Cj Malone <
> me-osm-talk...@keepawayfromfire.co.uk> wrote:
> > >
> > > I was splitting houses in Portsmouth/Southsea this morning. The
> imagery
> > > is great, I don't know if it was part of this update, or if it's been
> > > like this for a while.
> > >
> > >
> > > ___
> > > Talk-GB mailing list
> > > Talk-GB@openstreetmap.org
> > > https://lists.openstreetmap.org/listinfo/talk-gb
> > ___
> > Talk-GB mailing list
> > Talk-GB@openstreetmap.org
> > https://lists.openstreetmap.org/listinfo/talk-gb
> >
> >
> > ___
> > Talk-GB mailing list
> > Talk-GB@openstreetmap.org
> > https://lists.openstreetmap.org/listinfo/talk-gb
> >
> >
> > ___
> > Talk-GB mailing list
> > Talk-GB@openstreetmap.org
> > https://lists.openstreetmap.org/listinfo/talk-gb
>
> ___
> Talk-GB mailing list
> Talk-GB@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk-gb
>
___
Talk-GB mailing list
Talk-GB@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-gb

Re: [Talk-GB] Proposal: Import EV charging point data

[Alan Mackie]

On Tue, 18 Aug 2020 at 09:09, Rob Nickerson 
wrote:

> For me, once licencing issues have been fully resolved, it comes down to
> accuracy of data.
>
> For example the TfL cycle data is great as it has been collected by ground
> survey and with two photos of everything. Some other third party data has
> been less accurate. At this point it becomes tricky as we've no agreed
> threshold. We all know that OSM is not perfect but less clear on what level
> of accuracy we require for imports.
>

Wouldn't this be partially dependent on the level of mapping in the area in
question? In sparsely mapped regions even relatively coarse data might be
an improvement. In areas where individual gas lamps have been accurately
placed higher precision would be required to avoid misleading entries.

>
> As for tools, consider Ilya's OSM Conflator tool. Works well for node data.
>
> Rob
> ___
> Talk-GB mailing list
> Talk-GB@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk-gb
>
___
Talk-GB mailing list
Talk-GB@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-gb

Re: [Talk-gb-westmidlands] Fingerposts

[Alan Mackie]

Maybe:
tourism=information
information=guidepost
?

https://wiki.openstreetmap.org/wiki/Tag:information%3Dguidepost

I don't know a way of recording the directions they point save via
destination_sign relations.

https://wiki.openstreetmap.org/wiki/Relation:destination_sign

On Thu, 16 Jul 2020, 17:00 Andy Mabbett,  wrote:

> I've just come to map a fingerpost, never having had cause to add one
> previously.
>
> To my surprise there is nothing on the wiki about how to do so, and I
> can find no JSOM preset.
>
> What do folk suggest?
>
> I would like to add the inscription, for each of the three fingers,
> with their comaps points, something like:
>
>inscription:NW=foo
>
> or, using degrees:
>
>inscription:315:=foo
>
> --
> Andy Mabbett
> @pigsonthewing
> http://pigsonthewing.org.uk
>
> ___
> Talk-gb-westmidlands mailing list
> Talk-gb-westmidlands@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk-gb-westmidlands
>
___
Talk-gb-westmidlands mailing list
Talk-gb-westmidlands@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-gb-westmidlands

Re: [Talk-GB] Documenting tagging practice for place nodes in London

[Alan Mackie]

On Thu, 25 Jun 2020 at 13:56, Russ Garrett  wrote:

> On Thu, 25 Jun 2020 at 13:20, Andy Townsend  wrote:
> > Quite a lot of stuff of the placename info on OS StreetView probably
> > _shouldn't_ be in OSM.  Leaving aside farm and house names, the where I
> > used to live in Derbyshire is according to OS StreetView composed of 5
> > different "villages".  It's actually either 1 or 2, depending on who you
> > ask.  It's probably less of an issue in London (less space for
> > extraneous names), though.
>
> I have noticed a few cases, especially in areas of London I know very
> well, where OS shows an archaic name which isn't really in general
> use. This gets a bit tricky because there's not really a way of
> signalling to other mappers that a place name isn't in use based on
> local knowledge. Obviously local knowledge is best here but we
> probably don't have mappers with good local knowledge of all the
> various corners of London, and I'm pretty sure that there are areas of
> London where the area names have never been adequately mapped (which
> is why I started this thread). So I'm not sure how best to solve that
> conundrum.
>
> Out of date names might show up less in addr:? Of course then you need
full addresses rather than street, number and "leave it to the algo".
___
Talk-GB mailing list
Talk-GB@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-gb

Re: [Talk-gb-westmidlands] "Thank you NHS" sculpture in Lightwoods Park, Smethwick

[Alan Mackie]

Not to be callous, but if it's described as temporary and looks that
flimsy, is it worth mapping given the breeze forecast for next week?

On Mon, 29 Jun 2020 at 21:40, Andy Mabbett 
wrote:

> Is anyone kindly able to supply coordinates for this artwork:
>
>
> https://www.expressandstar.com/news/local-hubs/sandwell/2020/05/06/sculpture-in-tribute-nhs-frontline-workers-unveiled-at-sandwell-park/
>
> please?
>
> --
> Andy Mabbett
> @pigsonthewing
> http://pigsonthewing.org.uk
>
> ___
> Talk-gb-westmidlands mailing list
> Talk-gb-westmidlands@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk-gb-westmidlands
>
___
Talk-gb-westmidlands mailing list
Talk-gb-westmidlands@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-gb-westmidlands

Re: [Talk-GB] Still too many universities in Cambridge

[Alan Mackie]

On Tue, 4 Feb 2020 at 15:46, Mateusz Konieczny via Talk-GB <
talk-gb@openstreetmap.org> wrote:

>
>
>
> Feb 4, 2020, 16:37 by talk-gb@openstreetmap.org:
>
> >> (Ironically, the current tagging makes it hard for me to search to see
> >> if there's a "proper" amenity=university in there somewhere, e.g. as a
> >> relation or area covering a large swathe of them.)
> >
> >There isn't, I'm afraid.. it's a right hotchpotch
> IMHO, it would be a waste of time, if you tried to create a single area
> object (do I mean "closed way"?) to be the university.
>
> Or multipolygon, like for https://www.openstreetmap.org/relation/3830877
>
> The University is a collection of colleges, so could be a relation...
>  ...except that each college is probably in several buildings and they may
> not be in a contiguous area, so each college might have to be a relation of
> buildings.  So you would have a hierarchy of relations.
>
> Or areas that belong to multipolygon of college and multipolygon of
> university (?)
>
> We used to enjoy the look of puzzlement on the faces of (mostly American)
> tourists, who stood in the middle of town, surrounded by colleges, mixed in
> with shops, offices and other buildings, and asked which way to go to the
> University.
>
> :) In this case university multipolygon (or closed way) covering most of
> city center
> sounds correct and would help OSM-using tourists.
> 
>

Isn't this what site relations are for? That way POI nodes etc also
belonging to the University can be included.

On a completely unrelated note. Does any software actually support site
relations?
___
Talk-GB mailing list
Talk-GB@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-gb

Re: [Talk-GB] Increase of mail size limit

[Alan Mackie]

>
> It's quite normal to include attachments in line with communication (e.g
> an attachment in an email or an attachment in a chat/instant message
> program).

It is, but those tend to be a little more point-to-point rather than
send-to-all. I think most newsletter type things tend to link rather that
attach images.

On Thu, 16 May 2019 at 21:01, Rob Nickerson 
wrote:

> Personally I don't think that is very user friendly. I can post onto my
> Google drive or upload to the wiki (if it's one of the very few file types
> it supports) but others may not know or be bothered.
>
> It's quite normal to include attachments in line with communication (e.g
> an attachment in an email or an attachment in a chat/instant message
> program). We should allow people the choice so that they are able to do
> what comes naturally to them.
>
> *Rob*
>
>
> On Thu, 16 May 2019 at 13:37, Ed Loach  wrote:
>
>> Rob wrote:
>>
>> > Joking aside, please note I'm not asking for much here. A 40kb limit is
>> tiny and
>> > people do breach the limit occasionally. In this instance it was
>> ~500kb. I ended up
>> > having to put the attachments elsewhere but that delayed the post by 24
>> hours
>> > (it was already late and I decided to delay faffing until the next
>> day).
>>
>> I suspect a small limit is designed to discourage attachments, which
>> could be hosted elsewhere and links included.
>>
>> Ed
>>
>> ___
> Talk-GB mailing list
> Talk-GB@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk-gb
>
___
Talk-GB mailing list
Talk-GB@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-gb