2026年04月04日、現時点での愛知県瀬戸市にある全バス停のマッピングを完了しました。 また、04月01日にダイヤ改正が行われた菱野団地住民バスの情報を追加しました。 今後は、新ダイヤでの情報、バスの走行する経路の情報の追加や、近隣の尾張旭市・長久手市のバス停のマッピングなど行っていきたいです。

انا احب الخريطه لأنها جعلتني اذهب الى اماكن لم أكن اتوقع الذهاب اليها او من الصعب الذهاب اليها أصبحت مساهمات المدينه يحتويها الضمير لجعل الخريطه اكثر دقه ووضوح

Strado – Neighborhood Livability Scores from OpenStreetMap

I’ve been working on Strado, a free tool that scores every neighborhood across 50 European cities using data from OpenStreetMap. I wanted to share the project with the OSM community since it’s built entirely on your work.

What it does

Strado analyzes 22 categories of POIs from OSM – restaurants, healthcare, transit stops, parks, schools, nightlife, grocery stores, and more – and computes livability scores at the street level using an H3 hexagonal grid (resolution 9, ~174m edges).

The idea is simple: if you’re moving to a new city, you should be able to compare neighborhoods by what’s actually within walking distance. Not opinions, not sponsored listings – just data.

How it works

1. I imported the full Europe PBF into PostgreSQL/PostGIS using osm2pgsql with a custom flex style
2. POIs are extracted across 22 categories based on OSM tags (amenity, shop, leisure, healthcare, etc.)
3. Each H3 hex cell gets a count of nearby POIs with k=1 neighbor expansion
4. Livability scores are computed from essential categories (grocery, healthcare, transit, parks, education, safety)
5. Activity scores come from lifestyle categories (dining, nightlife, cafes, culture, shopping)
6. Everything is served as PMTiles vector tiles from Cloudflare R2 – no backend server needed

The scoring engine runs client-side in the browser using MapLibre GL JS. The entire infrastructure costs $0/month.

The data

• ~78 million POIs from OpenStreetMap
• 50 cities from London to Athens
• 22 scoring categories
• 20.5 million hex cells scored

Try it

• Interactive map – click any hexagon to see the score breakdown
• City pages – browse all 50 cities with neighborhood rankings
• Rome example – top neighborhoods, category breakdown, FAQ

All data is attributed to OpenStreetMap contributors under ODbL. Every page includes the attribution.

What I learned about OSM data quality

The coverage varies significantly across Europe. Western European cities (Paris, London, Berlin) have incredibly detailed mapping – you can score neighborhoods with high confidence. Eastern and Southern European cities are patchier, especially for healthcare facilities and transit stops. The Nordics have excellent cycling infrastructure data but fewer tagged restaurants.

Some categories work better than others with OSM data: - Excellent: dining, cafes, nightlife, shopping, parks - Good: healthcare, education, transit, grocery - Inconsistent: safety (police stations), financial (banks/ATMs), early education

I’d love to hear from mappers in cities where the data feels thin – what’s most worth mapping for tools like this?

Embeddable badges

I also created free embeddable SVG badges for all 50 cities. If you write about European cities or neighborhoods, you can embed a livability score badge on your site with a simple image tag.

Feedback welcome

This is a side project and I’m the only developer. If you notice data issues, missing categories, or have ideas for improvements, I’d appreciate hearing about it. The project is at strado.info.

Thanks to everyone who contributes to OpenStreetMap – none of this would be possible without your work.

Following up on osm.org/user/lhirlimann/diary/404921, I discovered today that newer/better aerial imagery for some parts of Mongolia were available, hence doing me doing some edits.

 

My book of suggested rides rates today's loop as "Easy". 

The Easter weekend is coming up and there was quite a bit of traffic near the coast, but almost all of the ride was on very quiet back roads. Indeed one section is hardly a road at all - more of a farm track. I thought I must have had drifted off my intended route, and double checked. I don't think I had made a mistake though.

As far as gradients go, it was mostly gently rolling countryside. Rolling enough to give the impression that I was practicing some gentle climbs without becoming much of a challenge. The only exception was between Ellingham and Chathill. Near Preston there's a drop to cross Long Nanny, then a climb back up again. But that was just a quirk. 

At Spindlestone the bridge across Waren Burn is currently closed to traffic. For a moment I thought I'd made a mistake when I ignored the Diversion signs. But there's a temporary footbridge, so it wasn't a problem on a bike. 

It would be perverse not to mention that the route passes Bamburgh Castle. But a series of less famous landmarks also deserve a mention. It's worth looking out for Spindlestone Ducket (pictured), St Hilda's Church at Lucker, St John the Baptist at Ellingham and the Corrugated Iron building at Chatton (which I think was a Reading Room). I deviated from the recommended route for a coffee stop at Beadnell. So I can also include St Ebba's Church and the old AA box on the B1340. There's probably more.

The day was dry and sunny. The wind was cold, and quite strong near the coast, but inland it was never much of a problem. 

"Easy" really doesn't do justice to this route. I'd be doing it again. And I don't object to paying for a Car Park. But the charges for a few hours at Seahouses are pretty steep. Next  time I will try to find an alternative starting point.

 

 

As reported on the Irish OpenStreetMap website, we recently “finished” mapping all the buildings in Co. Down. Due to my goal to get high up in the UK statistics (I made it to #1), I apparently mapped 3,283 tasks, if I’m reading the statistics right. Thanks to the grid system, that led to a fairly systematic coverage under my watchful eyes, resulting in the spotting of 13 potential unrecorded archaeological sites. It took me until today to write the reports to the department in Northern Ireland, because it’s not as much fun as mapping.

Link to overpass-turbo query

I usually add a note=might be a something site, discovered by b-unicycling YYYY-MM-DD to the way, so that I can look for them in overpass-turbo, once I get around to writing the reports to the respective government department. I then add reported by b-unicycling YYYY-MM-DD or something along the lines to the note, so that I know I have reported that one already. This is really only to help me keeping track of what I have discovered and what I have reported.

The way to spot these sites are usually crop marks or what they call on Time Team (see Wikipedia) “lumps and bumps” in the fields (or something suspicious in a lake which might be a crannóg, but those didn’t occur in Co. Down, not on the tiles I mapped anyways (okay, Jonny?). They appear as usually darker marks in the vegetation of a field, because the archaeological monument has a different soil composition than the surrounding soil, so that plants grow differently. The best crop, apparently, to spot them, is wheat. And there was a lot of wheat being grown in Ireland in 2022, when all that crisp Bing imagery was taken. The untrained eye might confuse them with tractor marks, especially when the imagery isn’t crisp enough and they cannot see whether there are two lines or one. Also, so called “fairy rings” (see Wikipedia) might throw you off which are often perfect circles or parts of perfect circles caused by fungi. They are completely natural and not of archaeological value. It would help to compare different years of aerial imagery for these, but we are not exactly spoilt for crisp aerial imagery. A useful website of different generations of Esri imagery is this. Sometimes, what looks like a very obvious cropmark of a ringfort will turn out to be an abandoned horse riding track. Tough luck. (And true story.)

Very obvious example on Bing of a cropmark of a “moated site” which I discovered in Co. Kilkenny recently, since confirmed by an archaeologist

I think it was danieldegroot2 who started adding maybe:historic=archaeological_site to them, and I usually still do that. I would like to mark the more or less obvious crop marks independent of that, though, because they are visible (if your eyes are tuned in on them, as mine are by now). Until the sites get explored, which most of them will never, because there are just too many, and there’s no money in archaeology, we will never know what they are and if they are indeed anything. So, the maybe:historic=archaeological_site isn’t the worst solution. The heritage department in Northern Ireland was very pleased with my reports, at least that’s what they wrote, and they said that those sites will be added to their database (https://experience.arcgis.com/experience/8bb16b64f0994385a5c141027ae9d33e/) within 15 days. That is A LOT faster than how things are done in the Republic of Ireland where I’m still waiting for verdicts for sites I reported since 2022. One of the reasons is that the people in Northern Ireland have a “monument class” called “crop mark” which is of course in itself not a monument type, but it allows them to record these without even a site visit. This is quite important, when people try to get planning permission to build on that site, and the heritage department can then send an archaeologist along to monitor the construction work and record possible finds. A pain in the behind for the landowner, but it’s important. In the Republic, they sometimes trust the crop marks, and I have had sites confirmed within weeks, but in most cases, it takes a lot longer.

The possibly most exciting find in Co. Down is what might (MIGHT!) be a banjo enclosure. The name is really self-explanatory (Wikipedia). It’s only exciting to me, because I haven’t found one before.

I had thought about starting a proposal for archaeological_site=cropmark, but luckily, I checked taginfo, and it seems that there is a (not documented) tag used 875 times which is historic=crop_mark. But does that apply to “lumps and bumps”? From an archaeological standpoint, it does not, but I think we could use it as an umbrella term, since most of us are not archaeologists. So I think I’ll go with that in the future, and maybe retrospectively for Co. Down.

And as a reminder: Metal detecting without an archaeological license in the Republic of Ireland is illegal and wrong, so don’t get any ideas. Having just watched a licensed detectorist at a licensed site - it is a highly skilled job. A lot goes on after the beep.

Introduction

Welcome to The #questionable-edits OSM Iceberg!

Over the past 5 months I have been collating some of the strangest, funniest and most excruciating examples of vandalism, mistakes and creative mapping. This Iceberg takes its name from the #questionable-edits channel on the OSM World Discord where we share the weird, wild and wonderful things we’ve seen while mapping. It should be made clear, the intent of #questionable-edits (and my Iceberg) is not to mock, demean or dissuade novice mappers, but to educate about common mapping pitfalls and share some of the entertaining things that have been found. I myself am guilty of a number of things listed below! I’ve included over 80 items in my Iceberg, ranging from the well-known to the downright unhinged. This Iceberg is also available on the OSM forums.

It shouldn’t have to be said, but please do not harass or otherwise antagonise any users whose edits may be visible in these examples. The screenshots and changesets linked below are intended as illustrative examples, and aren’t intended to target individual mappers.

Happy April Fools everybody! (Yes I might be a day late, but I was last year too). Enjoy!

The Iceberg

The Explainer

What’s an Iceberg without a breakdown of all the lore?

First Level

descriptive names Not every map element needs a name, but some mappers treat the name field as a challenge to come up with a descriptive name, for example ‘Morrisons customer car park’, ‘McDonald’s drive through’, ‘private driveway’ or ‘tarmac where a lovely lawn once stood’. Another curious example is this American football pitch with all the painted lines named.

landuse glued to roads Mostly an older mapping style, where landuse is glued to road centrelines instead of ending at the edge of road areas. It can make it really hard to edit roads and landuse, especially in iD. Some landuse areas are even mapped as multipolygons, with roads serving as outer members.

large bboxes One edit in Hanoi, another in Hammersmith. Large bboxes have caused a lot of frustration for users of OSMCha over the years, as they cover areas that weren’t edited. Often, this is a beginner who doesn’t realise they should upload their changes before moving on to a new area, but can also be a result of editing a large relation such as France or the United States.

tagging for the renderer Not everything renders on Carto, which is ok. Some mappers have employed creative tagging to show off particular features, such as these sports field markings.

bad changeset comments We’ve all seen them. ‘Edits’, ‘Update’, ‘.’, ‘asgsdfdgf’ or even a bunch of unreadable hashtags aren’t helpful for conveying a changeset’s purpose at a glance. Adding a bit of humour or personality to changeset comments is fine, but don’t take it too far.

not aligned to aerial imagery Many ways and polygons are misaligned when compared to current aerial imagery, although this is usually down to each imagery provider having a different offset. Other times, features may be mapped inaccurately or at an angle, such as this town filled with triangular buildings.

copying from Google Maps Despite being rule #1, ‘Don’t copy from other maps’ is occasionally broken. At least some mappers are kind enough to mark their changesets with source=Google Maps or tag ways source=Google Street View

unsquared buildings Not all buildings have 90 degree corners, but most do. Not using ‘Q’ to square up buildings is a common beginner error, but is easily fixed. Here’s a particularly perplexing example.

Second Level

dragged nodes Sometimes vandalism, sometimes accidental, dragged nodes can cause really funky rendering.

SEO spam All businesses want to promote their brand, but names in all-caps, pois in incorrect locations and paragraphs of self-promotional material (in notes, tags or even diary entries) isn’t the way to do it in OSM. If you’re a SEO spammer, read this wiki page.

renaming cities rude things A common form of vandalism, which usually gets reverted quickly. What did Los Angeles do to deserve this?

Gulf of America edit war On the 20th of January 2025, the Gulf of Mexico node was on version 20. It’s now on version 140. All that fuss for something that doesn’t even render on Carto. Of course, there’s an accompanying 300 post long forum thread.

Pokémon Go edits Pokémon Go uses OSM data to decide where to spawn certain types of Pokémon, but unfortunately that led to some players adding fake parks or mapping natural=beach in their garden to have easy access to water-type Pokémon. This doesn’t seem to happen much anymore.

typos in tags Very common, even for experienced contributors. Some typos are funnier than others, for example operator:type=pubic, social_facility:for=pregrant, building=semi_permanenet_house and hundreds of roof:shape values, which gave me a good laugh. Or how about a Wikidata Qid pasted in name=.

individual trees as natural=wood Isolated trees should be tagged as a natural=tree node, but some mappers like to trace a natural=wood way around the crown of each tree. This was also a technique used by golf mappers, as their software would render natural=wood areas but not natural=tree nodes.

broken multipolygons Multipolygons are hard :/ . Outers crossing outers, inners outside outers, duplicate multipolygons and incomplete outer rings are just some of the ways multipolygons can be broken.

Third Level

golf mappers There are golf simulators which can import courses from OSM, but many incorrect edits were made to golf courses as a result of incorrect tutorials and forum posts, including mapping individual trees as natural=wood areas, avoiding using multipolygons at all costs, drawing highway=service over main roads, tagging all buildings as office=yes and even tagging a natural=bay as golf=water_hazard. Thankfully, improvements to the import software means these kinds of workarounds aren’t required anymore.

Microsoft AI building slop Particularly in the UK, buildings imported en-masse from Microsoft’s computer generated building footprint database are of poor quality, not matching reality well and clumping semi-detached and terraced buildings together. I even made a whole forum thread expressing my concerns about the huge amount of poor quality buildings being imported.

blindly fixing validator errors While iD’s validator is very useful for finding errors in edits, some mappers feel they have to ‘resolve’ each issue before uploading, even if they haven’t checked the tag upgrade or geometry change is correct. Some examples include tagging crossing footways as bridge=yes instead of connecting them, and roads overlapping buildings with layer=-1 when the building should really be repositioned.

funny real names Sometimes, poi or place names sound so ridiculous you think they must be vandalism, only to discover no, it really is called that. MTB trails often have crazy names such as ‘Sponge Bob’, ‘Lumpy Gravy’ or ‘Grannies’.

unnecessary multipolygons In most areas, it’s agreed that multipolygons should only be used where absolutely necessary. However, some mappers choose to make multipolygons out of what could otherwise be closed ways, just because they share nodes with other ways. As as example, this school car park where every individual parking space was a multipolygon, or the country of Sweden.

deletion of Tel Aviv A major wave of vandalism hit Israel in 2023, with large parts of Tel Aviv being deleted. The DWG undertook a complex revert process, and the situation ended up in the media.

overnoding What’s the minimum distance allowed between two nodes in a way? 1 metre? 10 centimetres? 1 centimetre? Answers on a postcard please.

extremely large landuse Sometimes a relic of old landuse imports, lots of landuse areas could really do with splitting along roads, railways or other natural features. Romania has farmland the size of cities, and Uganda has residential landuse covering 260 square kilometres.

farmland instead of meadow Farmland is for arable crops, meadow is for grazing animals and growing hay (at least according to the wiki). However, large swathes of Ireland, the Isle of Man and the UK are tagged as farmland instead of meadow. We have a project ongoing to turn our islands green.

2.5D buildings A classic, using buildings and other tags to cause apartment blocks to appear in 2.5D on Carto. It’s a shame it had to be deleted, because it’s a real work of art.

fake islands and cities Perhaps someone mistook OSM for OpenGeoFiction, and mapped fake cities in South Sudan and Antarctica.

Fourth Level

not replying to changeset comments All too common is users ignoring changeset comments asking why they’ve mapped a certain way or providing advice how to improve their mapping. Communication and discussion is an important part of OSM, so remember to always keep changeset comments polite and respectful, even if you get ignored!

naughty landuse shapes Not all suspiciously shaped geometry is vandalism. Some islands were just made that way.

poor quality HOT edits HOT has a bit of a mixed reputation on the OSM World Discord, sadly as a result of poor training or validation in some cases. Like, what is this? Changesets tagged with HOT projects have also been used to add a secret bunker on Little Saint James, an animal shelter for fish and Bikini Bottom.

name:ru edit war A major wave of vandalism was visible worldwide in 2024, when a group of activists attempted to delete name:ru en-masse from the database, and proceeded to abuse users who reverted the vandalism. Roads were dragged across oceans and renamed offensive names. The forum thread is here.

razed railways Despite OpenHistoricalMap existing, many mappers still continue to add railways that were removed decades ago and have no trace on the ground. These long-gone railways are either mapped through open fields, or tagged on existing motorways. The prefix razed: implies that no physical presence remains, but nearly 40,000 razed:railways exist in the database. There’s a 370-post forum thread dissecting the issue.

‘secret’ MTB trails A vandalism campaign that’s been ongoing for a while now, where some mountain bikers think their ‘secret’ informal trails shouldn’t be mapped. Vandalism includes deleting trails, renaming trails, and abusing users who revert the vandalism. I call them ‘secret’ trails, because they are visible on the Strava heatmap. One user even posted a link to a photo as proof a trail was renamed, but that link was an ip logger, and another made stickers insulting an OSM editor and stuck them on signposts!

separate sidewalks without crossings Drawing separated sidewalks along streets, but not connecting the footway at any crossing points, particularly in neighbourhoods where only unmarked crossings are present. This completely ruins foot routing.

Farming Simulator edits Various Farming Simulator games allow importing real-world farms from OSM, but this has resulted in users adding nonexistent forests and roads or adding nonsense names to farmland, to alter the farm that ends up in their game.

Maps.me personal bookmarks Many users who wanted to create personal bookmarks in Maps.me unknowingly uploaded their notes to OSM, resulting in hundreds of shop=books or shop=bookmaker being created across the world. Many pois named ‘My house’ or similar exist for this same reason.

Fifth Level

anonymous notes Spam, abuse, personal notes and unclear information can all be found in OSM notes, but some users believe anonymous notes pose a major part of the problem, with no accountability for vandals or way to contact the original poster for more context. Here’s the obligatory 170-post forum thread.

DWG member slander The DWG do a lot of hard work cleaning up the map, resolving disputes and banning vandals, but some people respond with personal attacks, be that in notes, diary entries or edits with sockpuppet accounts. I’m not going to republish any, but I’ve seen plenty.

deleting sensitive sites in South Korea Last year, a wave of vandals attempted to delete sensitive facilities from OSM in South Korea, namely military bases, power plants and intelligence sites. Part of the vandalism could have originated from media reports that OSM was publishing locations of military bases, which are not visible on domestic South Korean maps.

theme park mapping Theme parks have a lot of decorative features which mimic real world features, making tagging challenging. Often there’s a lot of tagging for the renderer, such as natural=cave_entrance being tagged on rollercoasters as they enter buildings, tagging decorative flames as leisure=firepit or splitting pedestrian areas based on the colour of paving stones.

disconnected paths/roads The ends of routable ways being extremely close to the nodes they should be connected to, breaking routing but appearing connected on Carto. The result of ignoring validator warnings.

vibecoded editors breaking the map A relatively new phenomenon, users creating their own editors with AI, then proceeding to break the map. One case involved a vibecoded editor adding tags with no values, and confusing the ids of nodes and ways, dragging a way across the ocean. Disclaimer: There’s way to tell if an editor is vibecoded, but we suspect some of them must be.

undiscussed mass edits The Import Guidelines are strictly enforced; I’ve lost count of the amount of reverted imports I’ve seen now. This also extends to mass tagging changes which while well-meaning, still need to be discussed somewhere first.

stale data Occasionally, someone posts some mapping that is totally wrong, only to realise it matches historic imagery perfectly. Keeping OSM up-to-date with real-world changes is a big job; many places have shops mapped that have been closed for 10 years or more.

tagging for the (3D) renderer 3D building parts can be used to create some impressive rendering of landmarks, but some users stretch the definition of what can be called a building, to be able to use building parts to map them. This might include a roundabout median, aircraft on display, compost bin (view on F4map) or advertising signs and walls.

Sixth Level

warnings in name= We don’t map subjective opinions, but some mappers feel they need to warn map users of certain hazards, for example dangerous cliffs, bad smells, insects, scammers or litter.

TIGER cringe TIGER data, imported in 2007, provided the backbone for all mapping in the USA. However, it’s out of date and imprecise by modern standards, and despite a large effort to clean it up, lots of original data still remains.

impossible tag combos Tags that don’t belong on the same object, like natural=water and shop=, or building=residential and highway=path, or shop=hairdresser and maxspeed=20 mph. Or alternatively, invalid semicolon separated values, like building:levels=2;1.

physical separation Road lanes should only be drawn separately when separated by a physical barrier, but complex junctions or places where lanes are divided by painted medians can cause disagreements over the best way to map roads. This thread has some cursed examples.

deleting things I don’t understand With the great depth of tagging, there’s always going to be tags that are confusing, especially to new users. But not understanding what addr:interpolation ways are doesn’t justify deleting them.

name= instead of addr:housename= More of a personal gripe is houses and apartments in the UK which are tagged with name= instead of addr:housename= . Or equally annoying, houses tagged with both a name= and addr:housename=.

ending landuse at task boundaries If you’ve ever participated in a project that uses a tasking manager, you’re sure to have seen the message ‘Do not map outside of task boundary’. Unfortunately, that advice produces some pretty poor results when applied to landuse, resulting in the map looking more like a patchwork quilt than a cohesive landscape.

don’t map my private property We map private driveways, trails, paths and other features that are visible from aerial imagery or street level (with appropriate access tags). Many landowners have objected to this over the years, deleting trails on their property or adding name=Private (No Entry). See ‘Why can’t I delete this trail’.

Seventh Level

name=My House Whether by accident or ignorance, numerous users have named their own house on OSM, or the house of a friend or relative. Obviously, we don’t map private information in OSM, so report any you find to the DWG for redaction!

watsan:toilet_status= My personal least favourite OSM tag, this tag is used to record the cleanliness of toilets.

insulting my enemy Having a disagreement with your neighbour? Why not rename their house something insulting? Surprisingly, this vandalism wasn’t noticed for three-and-a-half years, at least until the original user made a Reddit post (archived version) bragging about it. Or alternatively, tell the world what you think about your teacher.

vandalism unnoticed for years Similar to Wikipedia, vandalism on OSM can go undetected for years unless someone specifically looks for a tag. I’d be curious to know what the oldest vandalism people have found was.

source=AI generated image Related to the ‘secret’ MTB trails, when vandals were told to provide proof for trails being renamed they responded with AI generated photos of trail signage. (And then they vandalised the sign).

not a heath natural=heath is intended for a very specific dwarf-shrub habitat, mainly covered by heather. However, Wales is covered with huge areas tagged natural=heath, used to represent named upland areas or nature reserves, despite the fact they’re mainly grassland. Grassland is not a heath!

private info pasted in tags Mapping while at work is a widespread practice, or so I’ve heard. Just be careful not to paste confidential information from work documents into OSM tags…

email=female It’s apparently a language barrier thing, but French mappers like mixing up email= and female=

not using lifecycle prefixes When mapping the construction of a new project, it’s common to use proposed: and construction: prefixes to indicate that the physical feature isn’t finished yet. I found a proposed airport mapped as if it were completed.

Eighth Level

every key has 1000s of nonsense values If you go to the final page of any key on taginfo, you’re bound to find dozens of nonsense tags, for example cuisine=Rolex, or this ‘bar’. There’s too many to list, so see what you can find.

mapping clouds Clouds turn up on aerial imagery from time-to-time, but tracing the shape and tagging them as natural=bare_rock, natural=glacier or even natural=water is clearly wrong.

abbreviated road names Names in OSM should be recorded unabbreviated, so don’t type Road as Rd, Drive as Dr, or Street as St. Abbreviation can be performed by data consumers, if they wish.

water=ocean Technically, oceans are the only thing not mapped in OSM. Or they would be, if not for a handful of natural=water + water=ocean.

source=ChatGPT Yes, people do really rely on the output of AI chatbots to inform their mapping. Who knows what kind of hallucinated information might end up in OSM tags.

pedestrian area misuse It is generally agreed that highway=footway + area=yes should be used for pedestrian plazas, and area:highway=footway for linear, routable streets and paths. That doesn’t stop many linear areas being mapped with area=yes, presumably because it renders on carto.

skibidi Ohio rizz sigma amogus Every new fad, phrase and meme causes a bit of vandalism from (presumably) children. Thankfully, it’s quite easy to query for name=amogus. I wonder if there’s a 67 Church somewhere.

flooding of Great Britain A vandal added natural=water to the Great Britain relation, resulting on tiles in England, Wales and Scotland rendering with a blue background for a few days.

Ninth Level

maxspeed:beast Clearly signed, this bridge has both a maximum speed and capacity which applies to any beasts you may be leading.

McDonald’s with 56 names It’s a McDonald’s with its name tagged in 56 different languages.

Kim Jong Un’s secret base Who knew, Kim Jong Un has a secret base on the Canada-US border. (Is it secret if it’s on OSM?) It’s rumoured that he also had a secret base mapped in Greenland, but I can’t find any evidence for this.

source=brain Used when fixing an error so blatant no survey or aerial imagery is required, for example building:levels=2000.

Republic of Pandora Vandalism creating a new country just before an Organic Maps data snapshot was taken made it appear the Republic of Pandora had seceded from the United Kingdom for about two months.

Church of (your name here) Many fake churches with novel new religions have been created, but Mattism might be my favourite.

Federated States of Barbie and Oppenheimer Barbenheimer in OSM, who would have thought it.

Who did this? Me?! If it hasn’t happened to you yet, it will. You stumble across something ridiculous, incorrect or confusing and think, “What idiot did this?” Then, you check the history and realise, it was in fact, you. 🤭

Tenth Level

Chinese oil refinery This entry is the original questionable edit that inspired the Iceberg. An oil refinery, mapped in such an incorrect way, it was impossible to view the area in iD. Mapping so perplexing, it had to be shared with the world. Well, not that it exists anymore (#RIPChineseOilRefinery). The imagery quoted as a source in many of the changesets was extremely high resolution drone imagery, and clearly was some kind of commercial project to map the refinery for internal purposes. However, drawing dense grids of footways (over 20,000 highway=footway in one refinery) to represent pedestrian areas was a choice, and the user’s diary entry didn’t provide much explanation either, declaring “Refine the Fire Pit” almost 200 times. It even featured a building that looked shellshocked by the mapping going on around it, so we made it our unofficial mascot. I even 3D mapped it.

Cat Ba bridge One of the more determined fantasy mappers, who unsuccessfully tried to prove a bridge that doesn’t exist, does. A video of the bridge mapped in iD is in fact not proof that it exists.

dogging=likely This tag was added to a car park by an OSM administrator, so it must be reliable. There wasn’t any source given.

#Living_credits This entry has no explanation. Do not read the lore unless you want to go down a very, very deep rabbithole.

OSM Rule 34 This entry is a JOKE, and is a saying I came up with: “OSM Rule 34: If you can think of a new tag, it already exists”

Free Centrist State of Derryounce What’s the number one priority for a group of friends trying to establish a micronation in an Irish peat bog? Map it on OSM of course, to advertise your fledgling new nation to the world.

Adolf Hitler Mapping Adolf Hitler did make it to Moscow, albeit only in OSM.

JeffreyEpstein blocked by DWG Not much I can say really. He was blocked after one edit for vandalising MTB trails. Probably for the best.

Acknowledgements

My Iceberg was inspired by Xvtn’s OSM Iceberg, created in 2024. I’d like to thank everyone who’s posted their finds in #questionable-edits over the last few years, for providing so much joy (and pain). Thanks also go to everyone who’s spent time cleaning up the map, particularly the DWG who feature in about half of these situations.

And if you’ve made it to the end, thank you for reading. I hope you enjoyed :)

LordGarySugar

الأماكن التي وضعها احد المساهمين في المدينه تحتوي على مساهمات غير موجوده لذالك احاول حذفها من بيانات الخريطه لجعل الخريطه ذات مصداقيه ودقه

………………………………………………………………………………………………………………………………………………………………………………………………………………………………………………………………………………………………………………………………………

I really love JOSM, and the more I use it, the more I discover that it offers some very useful features… that some of you may not be aware of ! Even if some are less useful, simply knowing they exist opens up the possibility that you might find them useful!

I’d like to show them to you here.

This diary post is a follow-up to my previous one on JOSM: Building drawing tips, a perfect guide.

1st manipulation: Use JOSM

It is essential to use the JOSM editor, which enables a wide range of manipulations, particularly when combined with other plugins, especially the “buildings_tools” and “utilsplugin2” plugins!

See JOSM shortcuts.

2nd manipulation: F to follow (F)

You can follow another way by pressing F, but you must first place a node along the other shape to indicate where to follow it.

3rd manipulation: Angle tracing (A)

By pressing A (Draw nodes) and then A again (to activate the “angle” mode of the Draw nodes tool), you can create angles of 30°, 45°, 60°, 90°… press again A to disable the angle mode.

Very precise angles can be useful for creating accurate geometric shapes.

4th manipulation: Transfer history (Ctrl+Shift+G)

To transfer the history, select both objects – the old one and the new one – and press Ctrl+Shift+G. Both objects must be of the same type (two ways, two nodes or two relations).

Transferring history makes it easy to follow the recommendation to Keep the history.

5th manipulation: Drawing buildings parallel to other buildings

To quickly draw a building parallel to other objects, simply select another building (or two reference points) and create the new building.

6th manipulation: Parallel (Shift+P)

Using Shift+P you can create a shape parallel to another that maintains the proportions (smaller or larger). Using Ctrl in the same time, allow you to adjust the distance of the parallel way more precisely.

Combined with other tools, this can be very effective, for example when mapping a grass area around a footway.

7th manipulation: Distribute nodes (Shift+B)

Distribute Nodes, Shift+B, allows you to distribute nodes at equal distances between the first and last nodes. It makes it very easy, for example, to create a perfect row of individual trees.

8th manipulation: Object Intersection (Shift+I)

Shift+I allows you to create an intersection between two objects, creating a shared node without altering the straight lines of the objects, resulting in flawless intersections.

9th manipulation: Set buildings size (Ctrl+Alt+B)

Setting a building width (Ctrl+Alt+B) is useful for creating shapes of the same width, which can also be merged for other purposes. You must reset the value to “0” to use the tool without width constraints.

10th manipulation: Symmetry (Alt+Shift+S)

The symmetry tool (Alt+Shift+S) allows you to move objects symmetrically relative to another.

It is quite similar to the Mirror tool (Shift+M).

This is very useful in combination with Ctrl+C followed by Ctrl+Alt+V.

11th manipulation: Shrinkwrap (Ctrl+Alt+W)

The Shrinkwrap plugin allows you to create a new shape that encompasses the selected shapes (Ctrl+Alt+W). Very useful for creating a single building=yes shape that encompasses all building:part=yes shapes.

En Málaga ya se han realizado contribuciones significativas a OpenStreetMap gracias al esfuerzo conjunto de pequeños grupos de personas. Sin embargo, a menudo resulta difícil encontrar un espacio donde estas iniciativas de colaboración local se coordinen de forma organizada y participativa.

Para cubrir este vacío, he creado un grupo de trabajo dedicado al mapeo colaborativo del municipio. Su objetivo es servir como punto de encuentro para mapeadores locales interesados en trabajar juntos en proyectos concretos, mejorando la organización y facilitando la participación.

Este grupo se centra en iniciativas locales específicas y no funciona como un foro de conversación general. Por ello, el acceso es mediante invitación, asegurando un entorno enfocado, respetuoso y productivo.

Logo de OSM Málaga. Fuente: trabajo propio (CC BY-SA 4.0) derivado de un trabajo previo de Aury88, disponible en el Wiki de OSM.

¿Por qué usar Zulip para el grupo?

Zulip es una herramienta especialmente orientada al trabajo colaborativo. Sus conversaciones se organizan por canales y temas, lo que permite mantener el enfoque y facilitar el seguimiento de cada proyecto.

Ofrece funcionalidades similares a Slack, la popular aplicación para el trabajo en equipo, pero con mayor control y sin depender de servicios privativos. Además, Zulip es software libre, en línea con la filosofía de OpenStreetMap.

Modelo inicial de gestión

Se propone un modelo de gestión sencillo. Los administradores se encargarán inicialmente de revisar solicitudes, enviar invitaciones y asignar roles.

Existen dos roles principales: invitados y miembros. Los invitados pueden participar en determinados canales y demostrar así su afinidad con el enfoque de trabajo del grupo. Los miembros, en cambio, participan plenamente tanto en la actividad del grupo como en la toma de decisiones.

La transición de invitado a miembro se decide entre administradores y miembros activos. Se valorará especialmente la actitud colaborativa, la calidad de las aportaciones y el respeto por las normas del grupo y de OSM.

De forma puntual, los invitados podrán acceder a canales de miembros cuando exista interés mutuo en colaborar. Este acceso será propuesto por miembros activos del canal y se concederá de manera consensuada y provisional. No implicará automáticamente la condición de miembro pleno, sino que formará parte de un proceso progresivo basado en la implicación, la calidad de las contribuciones y la afinidad con los valores del grupo.

Evolución del grupo

El grupo se mantendrá relativamente reducido, para preservar un entorno respetuoso y productivo, y garantizar la calidad de las contribuciones. No se necesita un modelo de gobernanza complejo, ya que los miembros compartirán valores y formas de trabajo afines.

Si el grupo crece, se podrán incorporar moderadores para apoyar a los administradores. Además, el grupo podrá evolucionar hacia un modelo más abierto y participativo, con normas claras y procesos transparentes que fomenten la colaboración colectiva.

Primeras invitaciones

Las primeras invitaciones al grupo se enviarán pronto a colaboradores que ya han demostrado su compromiso con la calidad, el respeto y la colaboración. Esto les permitirá incorporarse desde el inicio y participar activamente en la configuración del grupo, así como en el arranque de discusiones y proyectos locales de forma organizada.

Si compartes este enfoque de trabajo y los valores del grupo, puedes solicitar el acceso como invitado a través del formulario habilitado. Todas las solicitudes serán revisadas para garantizar un buen encaje con la dinámica del grupo y que la experiencia sea productiva para todos.

The Hub, Edinburgh rendered with Beakerboy’s OSM Building Viewer https://beakerboy.github.io/OSMBuilding/index.html?id=42997989

I’m working on a project using OSM building data so I thought I’d familiarise myself with the building:part and roof tags and see what the editing process is like.

The Hub only had height data and a basic outline tagged for the spire, so as the highest man made point in Edinburgh, it made for a good candidate to get stuck in.

I used height data from the OGL licensed Scottish LiDAR Phase 5 DSM that covers Edinburgh, and photos of the building from Wikimedia Commons

The elevation profile tool in QGIS proved useful for exploring the LiDAR data (after I worked out how to right-click to confirm the path). I normalised the elevation to zero with the raster calculator and set a custom colour ramp to aid visualisation:

Definitely fiddly to get right but interesting to see the level of 3D detail achievable with just a handful of tags and open data. There are more details to sort out, but I think I made a decent start.

Visualised changeset: https://resultmaps.neis-one.org/osm-change-viz?c=180674112#20/55.94878/-3.19521

Here’s what it looked like in F4Map before my edits:

https://demo.f4map.com/#lat=55.9489462&lon=-3.1951278&zoom=19&camera.phi=56.15

And after:

 

The NCN1 cycle path along the Northumberland coast is fairly level. I'm not keen on hills, so for me that's part of the appeal. But the off-road section between Lesbury and Warkworth is 30m or so above sea level. That means a short climb from the River Aln at Lesbury. Then at Warkworth the route drops to cross the River Coquet, rises to pass the Castle, then drops again to river level on the way to Amble.

This is a regular route for me, and on a return jouney, that means there are four climbs to negotiate. These days I expect to manage the climb at Hipsburn without getting off and pushing. But until today I've never managed all three at Warkworth. So it was a highlight on today's ride to manage all four climbs on this route without getting off and pushing. That's the first time I've done it.

There was a noticeable wind today, but it was blowing mostly from the west, and quite gusty. So not a great problem in practice. 

At the end of March I've reached my highest monthly mileage this year. Neither my frequency of rides, nor my average speed is quite where I hoped, but both have increased. As well as the successful climbs, today's ride has bumped up my Eddington Number. The sun shone, and the air was warmer than it has been recently. So a very pleasant day.

قمت بمحاولة تنظيف بيانات الخريطه من المساهمات الوهميه والغير موجوده وذالك لجعل الخريطه ذات مصداقيه للمستخدمين

Back “home” with Mom for the weekend, when I picked her up after church, I was advised by everyone that taking the back exit from the church parking lot. What back exit? It wasn’t on OSM or GM. (Makes sense, last time I was in that parking lot, it was all mud and construction; this exit doubles as cover for new water main or something.)

So of course I went out the normal exit, around the block and in and out of the new back entrance/exit to trace both lanes with the GPS. It is indeed now visible in our newest Mapbox imagery.

osm.org/changeset/180651021#map=17/43.834318/-70.432577&layers=VN

Drohnen sind aktuell in aller Munde. Leider allerdings eher im Kontext weltweiter Konflikte und geopolitischer Spannungen. Auch wenn ihr Ursprung in der militärischen Entwicklung liegt, werden sie heute vielseitig eingesetzt: zur Freizeitgestaltung, in der Vermessung sowie bei Hilfsaktionen. Genau hier möchte ich anknüpfen. Durch die Hochschule Mainz habe ich Zugriff auf unterschiedliche Drohnensysteme. Von Fluggeräten für YouTube-Influencer über Selbstbaudrohnen bis hin zu Enterprise-Modellen mit einer Wärmebildkamera. Gerade bei Letzterem stellt sich schnell die Frage: Wie kann diese Technologie greifbar gemacht werden? Die Idee war schnell geboren: Warum nicht genau das mit der Kinderfeuerwehr einmal ausprobieren? Eine kleine Übung, die Technik, Teamarbeit und das Ganze kombiniert in einem Abenteuer!

Meine Planung
Gesagt, getan: Die Begeisterung konnte ich bei den Betreuern der Kinderfeuerwehr schnell wecken. Ein Termin war gefunden und ich begann mir Gedanken zu machen, wie ich das Ganze umsetzen könnte. Der Plan: eine fiktive Vermisstensuche unter Einsatz einer Thermaldrohne. Was ist eine Thermaldrohne? Eine Thermaldrohne ist mit einer speziellen Kamera ausgestattet, die Wärme sichtbar machen kann. So können Menschen oder Tiere gefunden werden, die sich in Not befinden oder Hilfe benötigen, insbesondere wenn es dunkel ist oder sie sich verstecken. Daneben war ich neugierig: Wie gehen Kinder im Alter von 5 bis 9 Jahren eigentlich mit einer Karte, einer Drohne und deren Steuerung um?

Vorstellung und Fragerunde
An einem sonnigen Samstag im März war es dann soweit: 18 Kinder, drei Betreuer und ein paar Zuschauer. Eines war für mich schon im Vorfeld klar: Sobald die Kinder die Drohne sehen, kommt die Frage: „Dürfen wir auch fliegen?“ Meine Antwort: „Lasst uns das gerne auf später verschieben.“ Für mich war es das erste Mal mit Zuhörerinnen und Zuhörern in diesem Alter eine Übung durchzuführen und ich muss sagen: Es war richtig gut. Die Begeisterung war von Anfang an spürbar, und es wurden Fragen gestellt und Dinge von den Kindern gesagt, mit denen ich so nicht gerechnet hätte: „Darf ich dort mit der Drohne fliegen, wenn der Eigentümer es nicht will?“, „Mein Papa hat auch eine Drohne von Temu für 21 Euro.“ „… und deine Drohne kostet ja so viel wie 20 Nintendo Switch 2!“

Vermisstensuche
Dann wurde es ernst, zumindest im Rahmen unserer Übung: Die Vermisstensuche konnte beginnen. Dafür hatten wir im Vorfeld ein Wald- und Wiesengebiet ausgewählt, in dem sich fünf Personen versteckt hatten. Die Aufgabe für die Kinder: Gemeinsam mit der Drohne das Gebiet absuchen und alle „Vermissten“ finden. Und so viel vorweg: Am Ende wurden alle gefunden

Für mich waren beim Einsatz der Thermaldrohne besonders folgende Punkte interessant: Personen mit Rettungsdecken zu finden, ist durchaus herausfordernd. Die Größe und Topografie des Geländes stellen Geduld und Überblick auf die Probe. Nicht nur beim Pilot sondern auch bei den Kindern. Fünf Kinder gleichzeitig an oder eher über der Fernbedienung erfordern gute Koordination … bei allen Beteiligten. Was ich etwas unterschätzt habe: Wir hatten einen herrlichen Frühlingstag mit etwa 10 Grad und Sonnenschein. Für die Suche nach vermissten Personen ist das allerdings nicht optimal, da sich die Umgebung stärker aufheizt und die Temperaturunterschiede im sichtbaren geringer werden.

Fazit
Für mich bleibt vor allem die Erkenntnis, Kinder können schnell ein Gefühl für Technik entwickeln – und es entsteht viel Begeisterung, wenn man ihnen die Möglichkeit gibt, Dinge selbst auszuprobieren. Lieber Torben, lieber Christoph von der Freiwilligen Feuerwehr Hünstetten-Wallbach, vielen Dank für die Möglichkeit, den Nachwuchsfeuerwehrdamen und -herren den Einsatz einer Drohne bei der Vermisstensuche praktisch zu demonstrieren. Und zu guter Letzt: Vielen Dank an all die interessierten Kinder. Es war mir eine große Freude! Dass am Ende jedes Kind noch selbst (natürlich unter Aufsicht) eine kleine Drohne fliegen durfte, hat diesen Tag wohl sehr gut abgerundet.

此文本同時提供 台灣華語版本 This article is also available in Taiwanese Mandarin

---

OpenStreetMap’s collaborative nature is both its biggest strength and a source of persistent data-quality issues. With thousands of contributors independently adding phone tags to shops, restaurants, clinics, and government offices, each person tends to follow their own formatting style. For Taiwan, that means a database where the same country code can show up as +886, +886+, or +886(2), and a single city’s worth of phone numbers might span a dozen different conventions.

This post catalogues what we found when we scanned OSM elements across all six special municipalities and five additional counties — we are working on a normalizer to fix the issue.

---

The Scale of the Problem

Across eleven cached regions — all six special municipalities (臺北市, 新北市, 桃園市, 臺中市, 臺南市, 高雄市) plus 苗栗縣, 新竹市, 臺東縣, 連江縣, 金門縣 — we found 49,260 tags (phone or contact:phone) on 49,229 elements. After splitting multi-value fields on semicolons, that yields 50,643 individual phone number strings to classify.

Format class	Count	Share
E.123 space (+886 2 1234 5678)	41,842	82.6%
RFC 3966 dash (+886-2-1234-5678)	6,655	13.1%
No separator (+886212345678)	1,158	2.3%
Local format, no country code (02-1234-5678)	854	1.7%
Corrupt/typo country code (+866 …, +886(2)…)	92	0.2%
Other (wrong country, junk)	42	0.1%

Roughly 1 in 5 individual values deviates from the most common contributor convention, creating inconsistency that complicates deduplication, display, and machine parsing.

Things went from bad to downright ridiculous.

The format split varies noticeably by region:

Region	Region (ZH)	Tags	Values	E.123	RFC 3966	Other
TPE	臺北市 ★	9,804	10,146	85%	11%	4%
NWT	新北市 ★	13,963	14,395	85%	11%	4%
TAO	桃園市 ★	4,177	4,277	83%	12%	5%
TXG	臺中市 ★	8,065	8,282	73%	24%	4%
TNN	臺南市 ★	4,246	4,322	84%	11%	6%
KNN	高雄市 ★	5,168	5,262	84%	10%	5%
MIA	苗栗縣	1,318	1,338	77%	18%	5%
HSZ	新竹市	1,088	1,094	82%	12%	6%
TTT	臺東縣	1,101	1,178	96%	3%	1%
LIE	連江縣	95	103	98%	0%	2%
KMN	金門縣	235	246	90%	8%	2%

★ Special municipality. Taichung stands out with 24% RFC 3966 usage — roughly double the rate of any other major city — suggesting a dominant local editing pattern or tool default in that contributor community. The outlier island counties (連江縣, 臺東縣) have the highest E.123 consistency, possibly because their smaller contributor pools converge on informal norms more easily.

---

What “Correct” Means: Format Standards in OSM

Before diving into the issues, it’s worth clarifying what “correct” actually means in an OSM context.

The OSM wiki’s Key:phone page does not mandate a single format. It documents E.123 international notation, RFC 3966 (tel: URI dash notation), and NANP-style formatting without expressing a clear preference between them. In practice, E.123 space notation is the most commonly used by Taiwan contributors — which is why we use it as the normalisation target — but RFC 3966 dash notation is a legitimate alternative that the wiki explicitly acknowledges.

So the goal of normalization isn’t strict compliance with any one standard — it’s internal consistency: a dataset where everything follows the same convention is just much easier to work with than one that mixes three formats at random.

What Consistent Looks Like

For Taiwan, the most common contributor convention is E.123, followed by RFC 3966 / NANP (North American +1-style, RFC 3966-like):

ITU E.123
----------------------------------------
+886 2 1234 5678    ← Taipei landline
+886 4 1234 5678    ← Taichung landline
+886 37 123 456     ← Miaoli landline (3-digit area code)
+886 89 123 456     ← Taitung landline (3-digit area code)
+886 9X XXXX XXXX   ← Mobile
+886 800 XXX XXX    ← Toll-free (0800)

NANP
----------------------------------------
+886-2-1234-5678    ← Taipei landline
+886-4-1234-5678    ← Taichung landline
+886-37-123-456     ← Miaoli landline (3-digit area code)
+886-89-123-456     ← Taitung landline (3-digit area code)
+886-9X-XXXX-XXXX   ← Mobile
+886-800-XXX-XXX    ← Toll-free (0800)

Multiple numbers separated by semicolons, no trailing semicolon:

ITU E.123
----------------------------------------
+886 2 8787 8787;+886 2 8787 8765

(or)

NANP
----------------------------------------
+886-2-8787-8787;+886-2-8787-8765

Both are acceptable normalised formats. The open question for the community is agreeing on one and applying it consistently to resolve the current mixing.

Not your average daily struggle

---

Our findings

Issue 1: Inconsistent Separators

The most common deviation is mixing hyphens and spaces. Both of these encode the same number:

+886 2 2181 2345     ← E.123 (space, most common in TW OSM data)
+886-2-2181-2345     ← RFC 3966 dash (legitimate, less common)

The real problem is mixing both within a single value, which belongs to neither convention:

+886 2 2873-6548     ← space after country code, dashes within
+886-2-28358739      ← dashes, then no grouping in subscriber number

We found 1,554 values that contain both spaces and hyphens in a single phone string — the worst of both worlds, unambiguously wrong under either standard.

---

Issue 2: Missing Country Code

Some contributors enter phone numbers the way they would dial them locally — without the +886 prefix:

02-2581-7780
02 8751 3227
0222346763
0921067050

OSM’s phone tag is meant to hold an internationally dialable number. A value like 02-2581-7780 is ambiguous outside Taiwan: consumers have no way to know which country’s area-code conventions apply. We found 854 such values, including mobile numbers entered as bare 09XXXXXXXX strings.

---

Issue 3: No Separator After Country Code

A related variant omits any separator between the country code and the rest of the number:

+886288613257
+886228839850

These are syntactically valid in E.164 (the all-digits form used by telephony APIs) but fail most display validators and are unreadable as stored OSM data. We found 1,158 such values.

---

Issue 4: Corrupt or Malformed Country Codes

A small but non-trivial number of entries contain clear input errors:

+866 2 29126883      ← digits transposed (866 instead of 886)
+886+2 2311 2940     ← extra plus sign
+886(2)28232410      ← parenthesised area code (North American style)
+886.2 2322 3477     ← dot as separator
+8886 2 8780 6278    ← extra digit in country code
+00886-2-23825234    ← international dialling prefix 00 prepended

We found 92 such values. These will silently fail in any phone-number parsing library that enforces ITU-T E.164 syntax.

---

Issue 5: Duplicate Entries in Multi-Value Fields

OSM supports multiple phone numbers for one element using semicolons. We found 1,320 multi-value tags across the dataset. Of those, 24 contain duplicate entries — the same number appearing more than once:

+886 2 2916 0300;+886 2 2916 0300
+886 89 862 326;+886 89 862 326;+886 89 862 326

This suggests copy-paste mistakes during editing. While harmless individually, they can inflate the number of contact options and potentially confusing to machines.

---

Issue 6: Extension Numbers — a Format Wild West

You are the one accountable, Raiden!! (via @M4HCHE3ZY on X (formerly Twitter))

Beyond the main number itself, 635+ values encode an extension, using at least five different conventions found in the data:

Convention	Example	Count
Hash #	+886 2 2536 3001#8653	572
Tilde ~	+886 2 2368 0031~2	26
ext. / ext	+886 2 2741 5991 ext.21	30
Chinese 分機	+886 4 2528 5394分機6000	7
Comma , (iOS)	+886 2 2938 2300,630	~1+

Detecting extensions is tractable

As community members pointed out, a simple rule works: any character that is not a digit, space, or hyphen ([^\s\d-]) can be treated as the start of the extension suffix. This is essentially what our normalizer does — split at the first such character, normalize the base number, then reattach the suffix verbatim.

Encoding extensions is where it breaks down

The OSM wiki’s Key:phone#Extensions page currently documents three different conventions without picking one, which is itself a signal of how unresolved this is.

E.123 specifies ext as the separator. It was standardised in the printed-directory era — ext 8653 is readable on a business card, but apps do not reliably parse it. There is no DTMF interpretation; the extension string is purely informational.

Apple iOS (and macOS Contacts) stores extensions using a comma , as a pause-and-dial separator: +886-2-2938-2300,630. The comma signals the dialler to wait for the call to connect, then send the remaining digits as DTMF tones — so 630 is dialled automatically after the main number picks up. This is practical on-device behaviour, but it creates two distinct problems in OSM data:

1. Ambiguity with multi-value separators. OSM uses ; to separate multiple phone numbers in a single tag. Comma has no such defined role in OSM, so an iOS-style value like +886 2 2938 2300,630 is likely to be misread as a single malformed number rather than a number-plus-extension. We found 16 values with commas in the dataset; most are multi-value numbers incorrectly separated by , instead of ;, but at least one appears to be a genuine iOS-exported extension.
2. Non-portability. A comma-encoded extension is only meaningful to a DTMF-capable dialler. It conveys no human-readable information and is invisible to any parser that does not understand the pause-dial convention.

libphonenumber detects extensions across many separators (#, ext, x, ,, etc.) but emits no canonical output format for the extension part, leaving it to the caller.

RFC 3966 (tel: URI) is the most formally specified option — it uses ;ext=NNN. But RFC 3966 extensions create a structural conflict with OSM’s data model that is worth spelling out in full.

The RFC 3966 semicolon conflict

OSM uses the semicolon ; as the multi-value separator for phone tags:

+886 2 1234 5678;+886 2 8765 4321    ← two phone numbers, standard OSM

RFC 3966’s extension syntax also uses a semicolon as a parameter delimiter:

tel:+886-2-1234-5678;ext=8653        ← RFC 3966 with extension

If a contributor stores this in an OSM tag, any OSM editor or data consumer that naively splits on ; will interpret it as two values: tel:+886-2-1234-5678 and ext=8653. The extension becomes a phantom second phone number — one that is not a phone number at all.

The obvious workaround is to escape the semicolon as \;, a convention some OSM tags use for literal semicolons inside values. But this creates its own problems:

• OSM editors do not consistently honour \; escaping; many will still split on it or display it literally.
• RFC 3966 parsers expect a raw ; as the parameter delimiter — a backslash-escaped \;ext=8653 is not valid RFC 3966 and will not be parsed correctly by any compliant tel: URI parser.
• Machine readability is not improved: a consumer now needs to know both OSM’s backslash-escaping convention and RFC 3966’s parameter syntax, and reconcile the two. It adds encoding complexity without giving any parser a clean path to the extension digits.

The backslash escape is a leaky workaround that satisfies neither standard fully. It is, in effect, a third encoding layered on top of two already-conflicting ones.

The result is that RFC 3966 extension notation is structurally incompatible with OSM’s semicolon-as-multi-value convention, with no clean resolution available today. For this reason, our normalizer preserves extension suffixes as-is rather than attempting to rewrite them into any standard form.

---

A Note on E.123 and Machine Readability

Here’s something worth keeping in mind: even a perfectly normalised E.123 phone tag isn’t as machine-friendly as it looks.

E.123 was standardised by the ITU-T in 1988 — when the primary medium for a phone number was a business card, letterhead, or printed directory. The spaces in +886 2 1234 5678 are visual grouping aids for human readers, not semantic tokens. A parser encountering that string has to strip the spaces, infer the country code, and figure out the area code boundary — all heuristically.

RFC 3966’s tel:+886-2-1234-5678 is marginally more structured (hyphens as explicit separators, a URI scheme that signals “this is a phone number”), but still requires a real parser to interpret the digit groups. The truly machine-readable form is E.164 — +886212345678, all digits, no punctuation — which is what telephony APIs and databases actually want. None of these is what OSM stores by default.

This tension is fundamental: OSM’s phone tag is human-oriented. Normalization to E.123 is about making data consistent and editable by contributors, not about producing a format that apps can blindly ingest without parsing. The downstream app still needs a library like libphonenumber to do the real work — which is exactly why that library’s correctness for Taiwan’s edge-case area codes matters as much as it does.

---

A Note on Unexpected Area Code Grouping by google/libphonenumber

This one is subtle. Google’s libphonenumber — the standard library used by virtually every phone-number parser — groups some Taiwanese area codes differently than how they appear in local usage.

Taiwan assigns 3-digit and 4-digit area codes to several regions. libphonenumber’s metadata appears to represent these as extensions of their 2-digit neighbours, producing a different grouping than what locals would recognise:

Dialled	libphonenumber output	Expected output (E.123)
037-123-456	+886 3 7123 456	+886 37 123 456
049-123-4567	+886 4 9123 4567	+886 49 123 4567
082-123-456	+886 8 2123 456	+886 82 123 456
0826-12345	+886 8 26123 45	+886 826 12345
0836-12345	+886 8 36123 45	+886 836 12345
089-123-456	+886 8 9123 456	+886 89 123 456

Affected regions: Miaoli (037), Nantou (049), Kinmen (082), Wuqiu (0826), Matsu/Lienchiang (0836), and Taitung (089).

This means that even phone numbers already stored in +886 X XXXX XXXX form may carry a different digit grouping if they were entered via a tool backed by libphonenumber. The grouping we use here follows the National Numbering Plan and official government contact listings — though it’s worth noting this may be an intentional design choice in libphonenumber’s metadata.

---

See also:

• Issue Tracker: Unexpected formatting of the TW numbers with 3/4-digit area codes
• 公眾電信網路號碼計畫 (Public Telecommunication Network Numbering Plan, Chinese only) [PDF]
• TG Group Chat

---

About OpenStreetMap Taiwan Community

The OpenStreetMap Taiwan Community (OSMTW) is made up of enthusiastic mappers interested in Taiwan. Since 2010, OSMTW has evolved from a small gathering of individuals into a vibrant local community, welcoming more people to collaborate on mapping projects. OSMTW now co-hosts monthly meetups and occasional expeditions with the local Wikidata community in Taipei. Come by, check it out, and join us!

---

Ends/ Mon, 30 March 2026
Issued at NST 16:15
Last updated at Wed, 1 Apr 2026 NST 12:00

NNNN

The Virtual Institute for Sustainable Development - IVIDES.org® , a virtual research institute on the SD matter, and the IVIDES DATA®, a small Brazilian company on information technology consultancy, are lauching a new partnership with the State University of Campinas (UNICAMP) in order to use OpenStreetMap on the collaborative mapping of disaster risk for coastal communities on the northern coast of the state of São Paulo (Brazil).

---

First meeting on March, 2026. Source: project collection. Map data (c) OpenStreetMap contributors.

---

The Virtual Institute for Sustainable Development - IVIDES.org® and the IVIDES DATA company®, in a partnership with the State University of Campinas - UNICAMP), is starting an effort to provide technical supervision and training for collaborative mapping with OpenStreetMap involving three localities (one of them in indigenous lands) along the Northern Coast of São Paulo (Brazil). These communities are still suffering with the consequences of the latest big disaster, occurred on 2023, and constitute areas with common environmental problems, like inapropriate solid waste, rivers’ pollution and deforestation.

The research focus on the micromapping with OpenStreetMap and uMap in order to develop the “Participatory Environmental Mapping” approach[1] using the collaborative approach.

We hope this project can show better the usability of open data and open software on the disaster risk reduction research, with community participation and improving the personal skills of the collaborators. It is very important, specially on areas with low income levels and literacy. This citizen science approach can opportunize the community participation.

A start meeting was promoved privately on March 27, 2026 and the team is very anxious to use the OSM cartograhic database and related apps.

If you want to contribute and(or) communicate about this effort, please, send a message to this address: ivides [at] ivides.org. We are very happy in account with your participation.

As a virtual institute for research on sustainable development (and sustainability), with an approach based on the visions of Milton Santos and Ignacy Sachs, we are very honored to contribute with this effort, which we consider very legitm and useful for the local people (residents), which were included in the methodology, as mandatory to the co-working initiatives.

---

Reference

[1] The Participatory Environmental Mapping (in Brazilian portuguese: Mapeamento ambiental participativo) is an approach which was proposed by Carpi Junior and other authors over the last two decades or so. This type of mapping takes into account the views of local residents on environmental issues, as well as on people, their health and general well-being.

---

Important note: IVIDES.org® and IVIDES DATA® are registered trademarks. To keep contact: ivides [at] ivides.org | https://ivides.org