> Within the frame of Prototype Fund round 15 (March 2024 to August 2024), the German Federal Ministry of Education and Research sponsored Tobias Zwick to work on StreetComplete for iOS (see progress report)
The current demo doesn't seem to cover many places worldwide, but what is there is really neat! To me this already looks better than modern city sim games, because they never seem to get the scale right, among other things.
Not sure what's stopping them from covering more places, since the idea presumably is to generate these tiles from existing OSM data. And in fact one of the easiest things to do with StreetComplete is add the height and type of a building.
StreetComplete is an easy to use editor of OpenStreetMap data available for Android. It can be used without any OpenStreetMap-specific knowledge. It asks simple questions, with answers directly used to edit and improve OpenStreetMap data. The app is aimed at users who do not know anything about OSM tagging schemes but still want to contribute to OpenStreetMap.
StreetComplete automatically looks for nearby places where a survey is needed and shows them as quest markers on its map. Each of these quests can then be solved on site by answering a simple question. For example, tapping on a marker may show the question "What is the name of this road?", with a text field to answer it. More examples are shown in the screenshots below.
The user's answer is automatically processed and uploaded directly into the OSM database. Edits are done in meaningful changesets using the user's OSM account. Since the app is meant to be used on a survey, it can be used offline and is economic with data usage.
This is great! I wish more knowledge bases realized their barrier to entry. Eg even Wikipedia can be scary to edit the first time, if you want to present things like citations the right way.
Wikipedia is fine. Just make sure the content itself is good and appropriate for Wikipedia (such as no original research) and formatting citations is really least of your worries. When you do it wrong, some other editor will come and fix it.
You accidentally found a good example reason for the rule to exist. It's only relatively recently (2019-2021-ish) that Germany had more pre-slided bread move than unsliced loaves.
Therefore scoping the claim being made with citations actually does improve the wiki's accuracy.
You can say bread is sold pre-sliced accurately in a colloquial sense, but actually making a claim - as an encyclopedia inherently does - should carry information on how general a claim it actually is.
That doesn’t in any way invalidate that bread is sold sliced. That is the assertion, that sliced bread is available for sale. Even in places where bread is typically sold unsliced, the concept of sliced bread exists.
Or are you claiming that you are completely unaware of unsliced bread and need a citation to verify its existence?
Would it be appropriate for the Wikipedia article on "milk" to mention that it's sold frozen, unqualified, simply because that has happened?
Bread is usually sold sliced in some places, and usually sold unsliced in others. For a claim about this to be useful for wikipedia it should be specific and cited.
> rather than simply wording the statement to more accurately reflect the situation?
I'm arguing that this should be done, and yet a citation is still important. You didn't answer my question about milk. In what situation would it be appropriate to mention milk being sold frozen without a citation? For example, human milk banks actually do freeze the milk for transport, but if this is mentioned in a wikipedia article, it should be cited! Same goes for making a claim about whether bread is sold sliced or unsliced (even if both are encountered more commonly than frozen milk)
I love the project idea, unfortunately I had bad experiences with the community. I had lots of fun walking my neighbourhood to perform quests in the app until like one or two other users started reverting my edits. I checked their comments and it was some weird pedantic arguments, like I shouldn't mark a street as not walkable because there is no official sign that prohibits walking; the fact that it's basically a highway with no sidewalk would not matter.
I ended up digging through the OSM wiki for definitions and rules to argue my case but in the end it felt very much like people on a power trip insistent on getting their way. It has ruined the fun in StreetComplete for me and I haven't participated since.
This has been the most controversial quest in StreetComplete for a while. Following several discussions about it, I have now removed it for the upcoming version (i.e. v64.0).
Can you steelman the case of the users who reverted the above commenters' edits?
In most countries, walking on a controlled-access highway[1] is specifically illegal. Why should it not be marked as non-walkable just because there's no signage?
Are they actually checking for said signs? In the US, freeways have explicit no-pedestrian signs at on-ramps from other roads and streets. But those can be miles apart while potential places for a pedestrian to approach the freeway might be more frequent.
Just tag POI, they get stale so quickly. Everybody wants to tag the roads, but the real gap in many places is stale POI, or missing tags like opening hours. You'll get much less attention from these gatekeepers there too
I actually feel better about making edits if I know people may revert them. I don't have the time or interest to dig in deeply to what exactly they want, so I normally wouldn't contribute. But if I know people double check and fix mistakes, I'm more likely to try and be helpful with the app.
"I feel more comfortable contributing to huge open source projects through the use of AI because I know that the maintainers will review my code for me and catch the issues I didn't"
I don't want to attribute anything to how you expressed that you feel, but I believe that this is what's known as "Deferral of Responsibility" and it can actually _negatively_ impact the quality of the data you are claiming to want to help curate.
Please don't use quotes in this context, you're putting words in someone else's mouth that the person did not say and did not mean. In this case, the contributor is making a good faith effort to help, providing data that no one else is in a position to provide, and worries less because any issues can be corrected. In the AI case, anyone can run the slop generator and massive changes can be produced that overwhelm the ability of maintainers to check. It's not the same.
Aren't there many places where a highway with no sidewalk is the only walkable option? So if someone saw your not-walkable, they might think not walkable at all.
Seems like you guys had two different perspectives, both arguably right. Maybe it needs to be split into technically-walkable and comfortably-walkable layers.
The trouble is you start getting into subjectivity. Some people are comfortable with vie ferrate, many are not. The only thing you can do is tag with objective features that can be verified on the ground, including physical features and legal restrictions that are indicated by signage.
It’s tricky and maybe locale specific. A lot of roads near me have no pavement. You have to anticipate pedestrians, cyclists and horses on the road. Not at the same frequency as motor vehicles but not completely out of the ordinary.
If you were using StreetComplete I don't think you would have run into that. StreetComplete is very beginner friendly, and I don't think you can use it to mark a road as prohibiting pedestrians (which, as you discovered, is easy to get wrong if you don't understand what you're doing). You can use StreetComplete to mark whether there's a sidewalk or not, but I don't think anyone would object to you marking roads without sidewalks as not having a sidewalk.
Nevermind, I was wrong. Looks like they added a quest for this recently, and they're removing it now because it the wording of the quest didn't really match the tags it was applying when you solved it.
I'd personally never seen that quest myself, but looking at it now I can see why they're removing it.
It's not pedantic, it's just the way it is. For OSM to be valuable it needs to be consistent. There are many tags available but the "you can't walk here" tag is about whether you have a right to walk there. There are many highways you wouldn't want to walk on, but there are other tags to indicate that, like speed limit and presence of footpaths etc.
Saying a path can't be walked has pretty large consequences since routing apps will refuse to use them for walking directions. A highway with no footpath does seem extreme enough to warrant refusing to route for walking directions, but otherwise it's best to not block walking unless it's literally impossible or illegal to walk there.
I wonder if it's by design. TestFlight apps have a hard user install limit, I remember waiting around for someone to leave previous betas to try them out.
StreetComplete is regularly cited whenever OpenStreetMap is mentioned on HN, and with excellent reason (this is a really good introduction to mapping with OSM). Congratulations to the team for the beta!
Indeed, I've been following the issue since it was mentioned here last time and was positively surprised about how quickly they progressed and got the first version shipped.
Every Door has a very confusing UI (lots of unlabeled buttons with unclear icons), unfortunately. I would love to use it, but even the web-based iD editor is so much easier to figure out!
The Street Complete UI is just perfect for mobile. If I'm walking down the street with a phone, I don't want to edit polygons or accidentally move a road. I just want to fill in missing data on existing mapped points.
I agree and I've been thinking about making something similar to Every Door but that takes advantage of AI to let you snap a photo or just send a voice note and then have the appropriate tags generated for review based on some very strict schemas. Would something like that be of interest to you?
I was thinking it would be ideal to have something that lets you walk down a street holding your phone out, capture a series of images and then process that in to changes you approve manually at home. Could even be processed locally on a computer after the fact.
You can choose the type of questions you want to get asked about in StreetComplete. If you want to only enter store names and timings, then go into Settings and pick those.
The name probably just references the street in OpenStreetMap (OSM) as StreetComplete doesn't only collect street data as the sibling comment already noted.
That's Apple's native way of going back across macOS and iOS no? I'm not an Apple user and even I've come across this when handling a family member's device for a few minutes
Maybe Android's visual ←back button is the way to go after all instead of hoping someone will accidentally discover implicit gestures
It’s only intuitive when you see a page slide from the right to the left. Then you know you can go back to the previous page with a left to right slide. I’ve never seen it used to close a modal
Do you already know the one where you swipe alongside the bottom of the screen from left to right, or right to left? If not it's probably one of my most used gestures to swipe between apps.
I have been using StreetComplete on Android thanks to a comment here some weeks ago. This is a really great tool, the barrier to contribute is brought to nearly zero.
The only one thing missing for me is a small history of my past contributions. At some point, I made a mistake and entered the wrong information (nothing terrible, the surface of a walkway, it was 30% one surface, 70% another, I gave the 30% value instead of what I think 70% would be better). I would have been glad to be able to quickly "revert" this contribution.
As someone who has enjoyed playing GeoGuessr-like games in the past, I find it a bit fun to enter/guess a completely random username and see what edits it has made. Interestingly, quite a few of them are concentrated in the Germany/Switzerland/Austria region (the US is popular too)
Edit: I had no idea that https://www.openstreetmap.org/history shows a list of the most recent changes in a sidebar, and there seems to be a similar trend too in terms of where OpenStreetMap is edited most often.
You can actually split the road into multiple segments and answer each separately if you don't mind the extra effort. I think there should be something like "differs along the way" button which then asks you to pick a point where to cut.
I am incredibly excited for this -- while I love doing StreetComplete quests myself, the lack of iOS support has prevented me from grabbing my friends / random strangers and turning mapping my city into a social outing. I'm going to wait until it's officially available, but this is so wonderful... gosh, I can't wait. First style of quest I'm going to get ppl coordinated with is accurate hours on local businesses :^)
StreetComplete is great. For me it made Openstreetmap much more approachable and now thanks to me there are a few more public toilets and one small road in osm. Especially with the road it was really interesting to see how quickly Google maps also took in this path. But they still also have the old road in the app. So it's linked but not 100%
It's a great app, been using it on Android for years. It has a really low barrier of entry (except maybe OSM account creation) and makes improving OpenStreetMap really easy even for people that have no experience with map editing. There's simple questions, e.g. about road surface or accessibility, with illustrated multiple-choice answers.
Overall, my experience has been pretty smooth. Just look at how little platform specific code is found in the streetcomplete repo and how straightforward these connect with the common code.
Or maybe I have been using it too long so that I don't notice the awkward bits anymore.
I've been building https://maplibre.org/maplibre-compose/ in my spare time for the past two years and helped with this port; happy to answer any questions you have about KMP or CMP
Generally, I'm very happy with the platform and it's only gotten better over the past couple years I've been deep in it.
Yes, to be more specific, every tool like this has benefits offered and then taxes to be considered. For example, what's the tax for:
- Maintaining a significant new moving part in the dev process. KMP updates, management, integration, futures awareness, this all takes non-zero time.
- If an iOS dev wants to work on a KMP project what's the ramp up time? How much less efficiently can they contribute compared to a pure native app?
- How much time is spent on problems, bugs, issues, related to KMP. This is guaranteed non-zero for any tool.
- Is there less benefit to KMP with coding agents? Previously with two native apps, a bug fix had to be done twice. Currently, say a fix in done on Android, it's quite easy to say "propose an equivalent fix for the iOS repo". I do not claim this removes all benefits of a single code base, just that some things are not as bad now.
Regarding moving parts in KMP, I have the impression that it is quite stable now, and has been for some time. I feel it is less maintenance work than having an Android project that uses the Android UI framework and system APIs. Google loves to deprecate and invent new APIs every few years.
I feel there is very little in the way of breaking changes that need changes in the code in KMP. Most of that comes from some third party libraries on major new versions. The biggest change recently was a change in the Android Gradle plugin that required to change the project structure. Arguably, that again was more of an Android than a KMP thing.
> iOS developers
Well, the app is Kotlin Multiplatform + Compose Multiplatform as UI framework. So, while it runs on iOS, the parts that would feel native to an iOS dev are near zero: There are about 50 lines of code in Swift in the project and about 1000 dealing with iOS APIs, but written in Kotlin.
As Kotlin and Compose is used in Android development, the project will always feel more accessible to Android developers. That said, Kotlin is somewhat similar to Swift and Compose is also a declarative reactive user interface framework, i.e. similar to SwiftUI.
As for the setup, they need at least IntelliJ/Android Studio to compile all that Kotlin to native code before launching the app from XCode in a simulator. That's my setup at least. Certainly one could also just use gradle (the build system) from the commandline.
> KMP bugs
KMP bugs, none that I remember. About platform-specific issues we will see what will surface during the beta test. So far, one thing surfaced, but it was found easily and necessitated changing 4 lines of code.
Compose (the UI framework) is a different matter. At times, I feel it is quite unwieldy when not solving standard problems. Being designed primarily by Google, I also find that best practices change and deprecations happen fairly frequently. Certain features lacking in Compose Multiplatform I need to implement myself. For example, there are still no scroll bars.
> Coding Agents
I have little experience with that, but what I have seen, the agents have no problem generating working code for the platform-specific parts of the implementation. Some un-slopping is necessary as always.
And, as I wrote, the platform specific code takes up a tiny part of the whole code base. Of the 80k lines of source code in the production code, 2k lines of source code are platform specific.
(for context, I helped with the streetcomplete port, and have build and maintained a KMP / CMP maplibre sdk for the past two years, but otherwise consider myself a polyglot proficient in most common tech stacks at least to a moderate degree)
> Maintaining a significant new moving part in the dev process. KMP updates, management, integration, futures awareness, this all takes non-zero time.
If you're already doing that for Android, any additional tooling/infra churn for KMP is pretty negligible, except for a one-time cost of separating common kotlin code from android-specific code. KMP updates are just Kotlin updates.
> If an iOS dev wants to work on a KMP project what's the ramp up time? How much less efficiently can they contribute compared to a pure native app?
Two answers here:
1. Kotlin, the language, is quite similar to Swift, and can call ObjC APIs (or Swift annotated with @ObjC) directly via cinterop, so you're calling the same iOS SDKs. Compose is a declarative UI framework in the same vein as SwiftUI and React (and IMO with fewer footguns and quirks than SwiftUI, though there are some). It renders more like Flutter (via Skia) but it's not hard to embed as a component of a SwiftUI app.
2. KMP libraries can also export ObjC or Swift APIs, called in Swift. It's fairly common for there to be a shared business logic / service layer within an app, perhaps even shared view models, and leave the UI layer itself to SwiftUI and Compose for each platform.
Compared to something like React Native, I find that KMP is much friendlier and less opinionated about how you use it and how much you integrate it. You can use it to share a small module, or the whole app, or anywhere in between, and the seams at the boundary are not nearly as painful as you'd expect crossing languages.
> How much time is spent on problems, bugs, issues, related to KMP. This is guaranteed non-zero for any tool.
Hard to quantify, but in my time working on MapLibre Compose for the last two years, I've run into a handful of Androidx bugs, and much fewer KMP / CMP bugs. Even the same APIs, on the Compose Multiplatform side tend to have fewer pitfalls and bugs in my experience than on the Android side.
That said, the time I've spent dealing with _IDE_ bugs in JB IDEs has been far too much. These days, I spend much less time in an IDE, so it hasn't mattered as much, but I'd _like_ to spend more time there as ideally, even if agents are generating code, I want to read and navigate it in a nice environment.
> Is there less benefit to KMP with coding agents? ... just that some things are not as bad now.
Probably depends where you are on the vibe coding vs ai-assisted engineering spectrum. I'm a heavy AI user these days, close to 100% since Fable/Sol/Astra, but on code I care about, I don't vibe code, I still spend time to understand and refine at least the architecture, if not the line-by-line code. So still I prefer the single codebase because that's one fewer set of abstractions, of failure modes, of guardrails, of test harnesses, etc to worry about.
But if you care deeply about the native UI, using Compose on Android, SwiftUI on iOS, and then KMP for a shared core is an excellent architecture. The opposite, a shared Swift core, I think is also becoming more possible as Swift targets Android, though I'm unsure what the kotlin-on-android -> swift integration experience is like.
Maybe I don’t understand how it’s supposed to work, but the very first question I was given had three choices and none of them were correct. What are you supposed to do then?
It’s very confusing to me too. I clicked an icon to see what it was, and it was opening times of a supermarket, which I had no idea about. But I couldn’t find a way to just get back to where I was before. My workflow now is to keep force killing the app frequently when I get stuck.
Same here, I can't believe there is no way to go back. I have to s of things around my house that I can just add because I live here, but I don't remember everything, and once I hit a question that I cannot answer from memory, I am stuck.
Same here, but the GitHub thread actually calls that out as a known bug. Swipe from the left. Intention is that clicking sonewhere on the map closes the dialog.
No, there’s only the listed options (street has a sidewalk, or it doesn’t - no option for “sidewalk extends for half the street but not the other half”).
I don’t know if that counts as it does or it doesn’t.
Aha. If you click 'Uh' exactly on its text, it works. If you click even one pixel away from the actual text, you get the chooser with not enough options.
First-time StreetComplete user here. I just answered a few of these "street has a sidewalk" questions myself. Upon first clicking on one of them, I was able to answer this question for each side of the street individually. The visual selection that you mentioned only applies to one side.
I can tell for sure the current version of StreetComplete for Android does allow specifying that only one side of the road has a side walk. I solved a few of these quests just last week. No idea what's the state on iOS though.
That's when you hit "uh..." and leave a note with details. In an effort to keep things simple on a small screen it doesn't include every possible option. You can read the wiki to find the best way to tag it. https://wiki.openstreetmap.org/
I click a quest and answer it but the app seems to be stuck on that quest and I can't figure out how to get back to all of the quests available near me. Is that a common experience?
OSM editor Go Map!! also has a "Quests" overlay that is very similar to Street Complete (even uses the same icons). And you can also define custom quests for data you want to easily gather.
However, it's lacking the nice example pictures for each option - so, with Go Map!! you need to know what you're doing.
I'd like to see native Apple sign-in, as opposed to via a WebView. When I used the WebView, the password text box was obscured and fiddly to fill in.
There should be an easier way to dismiss something I've clicked on, than having to tap "uh..." and then dismiss two dialog boxes.
Things that I've submitted (e.g. house numbers) should be visible on the map, to visually show my contribution progress (even if pending review or whatever). At present my contributions just disappear into the void.
Satellite overlay would be very useful to correlate houses.
Seeing some indication of other recent local contributions would be nice, to make the app feel more alive.
OpenStreetMap is a database, not a map application. The map on the official openstreetmap website is more of a demo to show some of the data. There are many different map apps that use OpenStreetMap data though.
It's a... mix. I can see all the shapes of the roads and areas i mapped in my area, but they seem not to trust OSM naming so they either have no name or the name TomTom gave to the equivalent street.
I can see most POIs i added with the exact name i gave to them, but the reviews and opening hours are pulled from Tripadvisor.
It's explicitly advertised for driving, not walking, so it's optimized to get you to places. CoMaps and even OSMand are much better for walking around.
And there's a simple reason for the search being so bad in most OSM apps: offline search.
The focus of all these apps seems to be a self-imposed "the app should never touch a server, ever, because what if the user is on a 100$/GB roaming plan in Africa", but indexing the maps is not easy.
I run a local Nominatim database (search engine for OSM) for Italy and it's ~57gb of data and indexes for one country only, and it barely supports fuzzy search.
I tried the TomTom app (on iOS) recently but found it incredibly buggy, especially when connected to CarPlay. The speed indicator would often just show 1mph, the dark mode would work about 20% of the time, and searches would show results from other countries ahead of a place 4 miles up the road.
nice to see StreetComplete coming to iOS. Its question-based approach makes contributing to OpenStreetMap feel approachable even if you’re not familiar with the underlying data model
Thank you, German government:
> Within the frame of Prototype Fund round 15 (March 2024 to August 2024), the German Federal Ministry of Education and Research sponsored Tobias Zwick to work on StreetComplete for iOS (see progress report)
And NLnet.
I was curious what they're currently sponsoring; here's another neat OSM project: https://osm2world.org/demo/?lat=52.5239396&lon=13.4104859&ra...
The current demo doesn't seem to cover many places worldwide, but what is there is really neat! To me this already looks better than modern city sim games, because they never seem to get the scale right, among other things.
Not sure what's stopping them from covering more places, since the idea presumably is to generate these tiles from existing OSM data. And in fact one of the easiest things to do with StreetComplete is add the height and type of a building.
More like: thank you German taxpayers. You are very generous for funding stuff like this for the benefit of the entire world.
Finally a government doing something that benefits us rather than blowing up schools.
Just as they did with solar power. Thank you Germany!
Can you explain further? Wasn't China the one to give us cheap solar?
https://energytransition.org/2016/01/how-germany-helped-brin...
From the README:
StreetComplete is an easy to use editor of OpenStreetMap data available for Android. It can be used without any OpenStreetMap-specific knowledge. It asks simple questions, with answers directly used to edit and improve OpenStreetMap data. The app is aimed at users who do not know anything about OSM tagging schemes but still want to contribute to OpenStreetMap.
StreetComplete automatically looks for nearby places where a survey is needed and shows them as quest markers on its map. Each of these quests can then be solved on site by answering a simple question. For example, tapping on a marker may show the question "What is the name of this road?", with a text field to answer it. More examples are shown in the screenshots below.
The user's answer is automatically processed and uploaded directly into the OSM database. Edits are done in meaningful changesets using the user's OSM account. Since the app is meant to be used on a survey, it can be used offline and is economic with data usage.
This is great! I wish more knowledge bases realized their barrier to entry. Eg even Wikipedia can be scary to edit the first time, if you want to present things like citations the right way.
Wikipedia is fine. Just make sure the content itself is good and appropriate for Wikipedia (such as no original research) and formatting citations is really least of your worries. When you do it wrong, some other editor will come and fix it.
What is, and isn't, original research is quite often subjective.
It’s not. There are clear guidelines. https://en.wikipedia.org/wiki/Wikipedia:No_original_research
"I wish more knowledge bases realized their barrier to entry."
The stupidest Wikipedia argument I ever had, was having to cite the fact that bread is sold sliced.
You accidentally found a good example reason for the rule to exist. It's only relatively recently (2019-2021-ish) that Germany had more pre-slided bread move than unsliced loaves.
Therefore scoping the claim being made with citations actually does improve the wiki's accuracy.
You can say bread is sold pre-sliced accurately in a colloquial sense, but actually making a claim - as an encyclopedia inherently does - should carry information on how general a claim it actually is.
Where I come from bread is normally not sold sliced. That is the exception.
That doesn’t in any way invalidate that bread is sold sliced. That is the assertion, that sliced bread is available for sale. Even in places where bread is typically sold unsliced, the concept of sliced bread exists.
Or are you claiming that you are completely unaware of unsliced bread and need a citation to verify its existence?
Would it be appropriate for the Wikipedia article on "milk" to mention that it's sold frozen, unqualified, simply because that has happened?
Bread is usually sold sliced in some places, and usually sold unsliced in others. For a claim about this to be useful for wikipedia it should be specific and cited.
Why is a citation necessary rather than simply wording the statement to more accurately reflect the situation?
Bread is sold both sliced and unsliced, sometimes geographic location impacts which variation is predominant.
> rather than simply wording the statement to more accurately reflect the situation?
I'm arguing that this should be done, and yet a citation is still important. You didn't answer my question about milk. In what situation would it be appropriate to mention milk being sold frozen without a citation? For example, human milk banks actually do freeze the milk for transport, but if this is mentioned in a wikipedia article, it should be cited! Same goes for making a claim about whether bread is sold sliced or unsliced (even if both are encountered more commonly than frozen milk)
I'm going to merge some of this into my new guidance system. Very cool.
I love the project idea, unfortunately I had bad experiences with the community. I had lots of fun walking my neighbourhood to perform quests in the app until like one or two other users started reverting my edits. I checked their comments and it was some weird pedantic arguments, like I shouldn't mark a street as not walkable because there is no official sign that prohibits walking; the fact that it's basically a highway with no sidewalk would not matter.
I ended up digging through the OSM wiki for definitions and rules to argue my case but in the end it felt very much like people on a power trip insistent on getting their way. It has ruined the fun in StreetComplete for me and I haven't participated since.
This has been the most controversial quest in StreetComplete for a while. Following several discussions about it, I have now removed it for the upcoming version (i.e. v64.0).
Can you steelman the case of the users who reverted the above commenters' edits?
In most countries, walking on a controlled-access highway[1] is specifically illegal. Why should it not be marked as non-walkable just because there's no signage?
[1]: https://en.wikipedia.org/wiki/Controlled-access_highway
In some countries it is legal to walk in places where it is dangerous and noone would ever do it.
Maybe it was such case?
Are they actually checking for said signs? In the US, freeways have explicit no-pedestrian signs at on-ramps from other roads and streets. But those can be miles apart while potential places for a pedestrian to approach the freeway might be more frequent.
Just tag POI, they get stale so quickly. Everybody wants to tag the roads, but the real gap in many places is stale POI, or missing tags like opening hours. You'll get much less attention from these gatekeepers there too
I actually feel better about making edits if I know people may revert them. I don't have the time or interest to dig in deeply to what exactly they want, so I normally wouldn't contribute. But if I know people double check and fix mistakes, I'm more likely to try and be helpful with the app.
"I feel more comfortable contributing to huge open source projects through the use of AI because I know that the maintainers will review my code for me and catch the issues I didn't"
I don't want to attribute anything to how you expressed that you feel, but I believe that this is what's known as "Deferral of Responsibility" and it can actually _negatively_ impact the quality of the data you are claiming to want to help curate.
Please don't use quotes in this context, you're putting words in someone else's mouth that the person did not say and did not mean. In this case, the contributor is making a good faith effort to help, providing data that no one else is in a position to provide, and worries less because any issues can be corrected. In the AI case, anyone can run the slop generator and massive changes can be produced that overwhelm the ability of maintainers to check. It's not the same.
Aren't there many places where a highway with no sidewalk is the only walkable option? So if someone saw your not-walkable, they might think not walkable at all.
Seems like you guys had two different perspectives, both arguably right. Maybe it needs to be split into technically-walkable and comfortably-walkable layers.
Or add a "probability of injury or death" attribute for more granularity.
Walkability based on time of day?
The trouble is you start getting into subjectivity. Some people are comfortable with vie ferrate, many are not. The only thing you can do is tag with objective features that can be verified on the ground, including physical features and legal restrictions that are indicated by signage.
It’s tricky and maybe locale specific. A lot of roads near me have no pavement. You have to anticipate pedestrians, cyclists and horses on the road. Not at the same frequency as motor vehicles but not completely out of the ordinary.
You would think photo evidence would be helpful here.
If you were using StreetComplete I don't think you would have run into that. StreetComplete is very beginner friendly, and I don't think you can use it to mark a road as prohibiting pedestrians (which, as you discovered, is easy to get wrong if you don't understand what you're doing). You can use StreetComplete to mark whether there's a sidewalk or not, but I don't think anyone would object to you marking roads without sidewalks as not having a sidewalk.
Nevermind, I was wrong. Looks like they added a quest for this recently, and they're removing it now because it the wording of the quest didn't really match the tags it was applying when you solved it.
I'd personally never seen that quest myself, but looking at it now I can see why they're removing it.
It's not pedantic, it's just the way it is. For OSM to be valuable it needs to be consistent. There are many tags available but the "you can't walk here" tag is about whether you have a right to walk there. There are many highways you wouldn't want to walk on, but there are other tags to indicate that, like speed limit and presence of footpaths etc.
Saying a path can't be walked has pretty large consequences since routing apps will refuse to use them for walking directions. A highway with no footpath does seem extreme enough to warrant refusing to route for walking directions, but otherwise it's best to not block walking unless it's literally impossible or illegal to walk there.
Link to the actual TestFlight beta invite, since it wasn’t easily found on the linked page:
https://testflight.apple.com/join/K1u3eUU5
Thanks. If anyone from StreetComplete is watching - please put the TestFlight link prominently on your page!
I wonder if it's by design. TestFlight apps have a hard user install limit, I remember waiting around for someone to leave previous betas to try them out.
10,000.
I suspect one day at the top of HN's front page, will exhaust that.
That's a big reason I never mention my apps, here.
StreetComplete is regularly cited whenever OpenStreetMap is mentioned on HN, and with excellent reason (this is a really good introduction to mapping with OSM). Congratulations to the team for the beta!
Indeed, I've been following the issue since it was mentioned here last time and was positively surprised about how quickly they progressed and got the first version shipped.
I tried both StreetComplete and EveryDoor on my Android for a while.
I prefer EveryDoor. It's oriented more towards stores and things I'm interested in on the map.
EveryDoor exist on the Apple store too.
Every Door has a very confusing UI (lots of unlabeled buttons with unclear icons), unfortunately. I would love to use it, but even the web-based iD editor is so much easier to figure out!
The Street Complete UI is just perfect for mobile. If I'm walking down the street with a phone, I don't want to edit polygons or accidentally move a road. I just want to fill in missing data on existing mapped points.
I agree and I've been thinking about making something similar to Every Door but that takes advantage of AI to let you snap a photo or just send a voice note and then have the appropriate tags generated for review based on some very strict schemas. Would something like that be of interest to you?
Also open to feedback from anyone.
I was thinking it would be ideal to have something that lets you walk down a street holding your phone out, capture a series of images and then process that in to changes you approve manually at home. Could even be processed locally on a computer after the fact.
Sounds like hosted AI inference would be involved! Who's paying?
Me!
You can choose the type of questions you want to get asked about in StreetComplete. If you want to only enter store names and timings, then go into Settings and pick those.
Have you tried out the Things or Places overlay? Second button from the right at the top, looks like a cake.
StreetComplete, as indicated by it's name, is about street data. Both are for contributing to OSM, but they are for different parts of OSM
StreetComplete asks about things other than street data as well. For example, I get queries about operating hours of businesses.
The name probably just references the street in OpenStreetMap (OSM) as StreetComplete doesn't only collect street data as the sibling comment already noted.
>Known caveats (we are working on it):
>forms can only be canceled with the back gesture rather than clicking anywhere on the map
Well for me not even the back gesture works, I just have to force close the app if I don't want to answer a survey lol
It takes a bit of practice, but I got it to work after a few tries.
I might be dense, but what's the back gesture exactly?
EDIT: Got it, swiping from left to right, starting from the very border of the screen. Not super intuitive but simple enough
That's Apple's native way of going back across macOS and iOS no? I'm not an Apple user and even I've come across this when handling a family member's device for a few minutes
Maybe Android's visual ←back button is the way to go after all instead of hoping someone will accidentally discover implicit gestures
It’s only intuitive when you see a page slide from the right to the left. Then you know you can go back to the previous page with a left to right slide. I’ve never seen it used to close a modal
Android can work either way. I turn off the gestures because I prefer the explicit back button.
I've been on iOS for like seven years and have never done this. But... hey, it works!
Do you already know the one where you swipe alongside the bottom of the screen from left to right, or right to left? If not it's probably one of my most used gestures to swipe between apps.
I have been using StreetComplete on Android thanks to a comment here some weeks ago. This is a really great tool, the barrier to contribute is brought to nearly zero.
The only one thing missing for me is a small history of my past contributions. At some point, I made a mistake and entered the wrong information (nothing terrible, the surface of a walkway, it was 30% one surface, 70% another, I gave the 30% value instead of what I think 70% would be better). I would have been glad to be able to quickly "revert" this contribution.
But else, a really great tool.
There is an undo button on the lower left corner of the screen. Try it!
(It only shows edits made in the last 24 hours IIRC, but this limit is chosen quite arbitrarily. If there is a use case for it, it can be extended.)
If you go to the OSM website, you can sort of get list of all your modification by date in your profile: https://www.openstreetmap.org/user/YOUR_USERNAME
As someone who has enjoyed playing GeoGuessr-like games in the past, I find it a bit fun to enter/guess a completely random username and see what edits it has made. Interestingly, quite a few of them are concentrated in the Germany/Switzerland/Austria region (the US is popular too)
Edit: I had no idea that https://www.openstreetmap.org/history shows a list of the most recent changes in a sidebar, and there seems to be a similar trend too in terms of where OpenStreetMap is edited most often.
> I had no idea that https://www.openstreetmap.org/history shows a list of the most recent changes
Here are some more views:
- https://live.openstreetmap.fr/
- https://osm-in-realtime.jwestman.net/
- https://theplanetthinks.com/map
If you like this kind of thing, https://osmstats.neis-one.org/ has a lot more stats on edits.
you can revert immediately in bottom left corner of StreetComplete, click on little revert arrow
> I think 70% would be better
You can actually split the road into multiple segments and answer each separately if you don't mind the extra effort. I think there should be something like "differs along the way" button which then asks you to pick a point where to cut.
I am incredibly excited for this -- while I love doing StreetComplete quests myself, the lack of iOS support has prevented me from grabbing my friends / random strangers and turning mapping my city into a social outing. I'm going to wait until it's officially available, but this is so wonderful... gosh, I can't wait. First style of quest I'm going to get ppl coordinated with is accurate hours on local businesses :^)
StreetComplete is great. For me it made Openstreetmap much more approachable and now thanks to me there are a few more public toilets and one small road in osm. Especially with the road it was really interesting to see how quickly Google maps also took in this path. But they still also have the old road in the app. So it's linked but not 100%
It's a great app, been using it on Android for years. It has a really low barrier of entry (except maybe OSM account creation) and makes improving OpenStreetMap really easy even for people that have no experience with map editing. There's simple questions, e.g. about road surface or accessibility, with illustrated multiple-choice answers.
I'd like to hear more about people's experiences with Kotlin Multiplatform
Overall, my experience has been pretty smooth. Just look at how little platform specific code is found in the streetcomplete repo and how straightforward these connect with the common code.
Or maybe I have been using it too long so that I don't notice the awkward bits anymore.
Any specific parts you are after?
Have you ever used flutter? If so, what's the relative pain level like?
I can't compare it to Flutter, I haven't used it.
I've been building https://maplibre.org/maplibre-compose/ in my spare time for the past two years and helped with this port; happy to answer any questions you have about KMP or CMP
Generally, I'm very happy with the platform and it's only gotten better over the past couple years I've been deep in it.
Yes, to be more specific, every tool like this has benefits offered and then taxes to be considered. For example, what's the tax for:
- Maintaining a significant new moving part in the dev process. KMP updates, management, integration, futures awareness, this all takes non-zero time.
- If an iOS dev wants to work on a KMP project what's the ramp up time? How much less efficiently can they contribute compared to a pure native app?
- How much time is spent on problems, bugs, issues, related to KMP. This is guaranteed non-zero for any tool.
- Is there less benefit to KMP with coding agents? Previously with two native apps, a bug fix had to be done twice. Currently, say a fix in done on Android, it's quite easy to say "propose an equivalent fix for the iOS repo". I do not claim this removes all benefits of a single code base, just that some things are not as bad now.
Alright, let's see:
> KMP stability
Regarding moving parts in KMP, I have the impression that it is quite stable now, and has been for some time. I feel it is less maintenance work than having an Android project that uses the Android UI framework and system APIs. Google loves to deprecate and invent new APIs every few years. I feel there is very little in the way of breaking changes that need changes in the code in KMP. Most of that comes from some third party libraries on major new versions. The biggest change recently was a change in the Android Gradle plugin that required to change the project structure. Arguably, that again was more of an Android than a KMP thing.
> iOS developers
Well, the app is Kotlin Multiplatform + Compose Multiplatform as UI framework. So, while it runs on iOS, the parts that would feel native to an iOS dev are near zero: There are about 50 lines of code in Swift in the project and about 1000 dealing with iOS APIs, but written in Kotlin.
As Kotlin and Compose is used in Android development, the project will always feel more accessible to Android developers. That said, Kotlin is somewhat similar to Swift and Compose is also a declarative reactive user interface framework, i.e. similar to SwiftUI. As for the setup, they need at least IntelliJ/Android Studio to compile all that Kotlin to native code before launching the app from XCode in a simulator. That's my setup at least. Certainly one could also just use gradle (the build system) from the commandline.
> KMP bugs
KMP bugs, none that I remember. About platform-specific issues we will see what will surface during the beta test. So far, one thing surfaced, but it was found easily and necessitated changing 4 lines of code.
Compose (the UI framework) is a different matter. At times, I feel it is quite unwieldy when not solving standard problems. Being designed primarily by Google, I also find that best practices change and deprecations happen fairly frequently. Certain features lacking in Compose Multiplatform I need to implement myself. For example, there are still no scroll bars.
> Coding Agents
I have little experience with that, but what I have seen, the agents have no problem generating working code for the platform-specific parts of the implementation. Some un-slopping is necessary as always.
And, as I wrote, the platform specific code takes up a tiny part of the whole code base. Of the 80k lines of source code in the production code, 2k lines of source code are platform specific.
(for context, I helped with the streetcomplete port, and have build and maintained a KMP / CMP maplibre sdk for the past two years, but otherwise consider myself a polyglot proficient in most common tech stacks at least to a moderate degree)
> Maintaining a significant new moving part in the dev process. KMP updates, management, integration, futures awareness, this all takes non-zero time.
If you're already doing that for Android, any additional tooling/infra churn for KMP is pretty negligible, except for a one-time cost of separating common kotlin code from android-specific code. KMP updates are just Kotlin updates.
> If an iOS dev wants to work on a KMP project what's the ramp up time? How much less efficiently can they contribute compared to a pure native app?
Two answers here: 1. Kotlin, the language, is quite similar to Swift, and can call ObjC APIs (or Swift annotated with @ObjC) directly via cinterop, so you're calling the same iOS SDKs. Compose is a declarative UI framework in the same vein as SwiftUI and React (and IMO with fewer footguns and quirks than SwiftUI, though there are some). It renders more like Flutter (via Skia) but it's not hard to embed as a component of a SwiftUI app. 2. KMP libraries can also export ObjC or Swift APIs, called in Swift. It's fairly common for there to be a shared business logic / service layer within an app, perhaps even shared view models, and leave the UI layer itself to SwiftUI and Compose for each platform.
Compared to something like React Native, I find that KMP is much friendlier and less opinionated about how you use it and how much you integrate it. You can use it to share a small module, or the whole app, or anywhere in between, and the seams at the boundary are not nearly as painful as you'd expect crossing languages.
> How much time is spent on problems, bugs, issues, related to KMP. This is guaranteed non-zero for any tool.
Hard to quantify, but in my time working on MapLibre Compose for the last two years, I've run into a handful of Androidx bugs, and much fewer KMP / CMP bugs. Even the same APIs, on the Compose Multiplatform side tend to have fewer pitfalls and bugs in my experience than on the Android side.
That said, the time I've spent dealing with _IDE_ bugs in JB IDEs has been far too much. These days, I spend much less time in an IDE, so it hasn't mattered as much, but I'd _like_ to spend more time there as ideally, even if agents are generating code, I want to read and navigate it in a nice environment.
> Is there less benefit to KMP with coding agents? ... just that some things are not as bad now.
Probably depends where you are on the vibe coding vs ai-assisted engineering spectrum. I'm a heavy AI user these days, close to 100% since Fable/Sol/Astra, but on code I care about, I don't vibe code, I still spend time to understand and refine at least the architecture, if not the line-by-line code. So still I prefer the single codebase because that's one fewer set of abstractions, of failure modes, of guardrails, of test harnesses, etc to worry about.
But if you care deeply about the native UI, using Compose on Android, SwiftUI on iOS, and then KMP for a shared core is an excellent architecture. The opposite, a shared Swift core, I think is also becoming more possible as Swift targets Android, though I'm unsure what the kotlin-on-android -> swift integration experience is like.
Maybe I don’t understand how it’s supposed to work, but the very first question I was given had three choices and none of them were correct. What are you supposed to do then?
It’s very confusing to me too. I clicked an icon to see what it was, and it was opening times of a supermarket, which I had no idea about. But I couldn’t find a way to just get back to where I was before. My workflow now is to keep force killing the app frequently when I get stuck.
Same here, I can't believe there is no way to go back. I have to s of things around my house that I can just add because I live here, but I don't remember everything, and once I hit a question that I cannot answer from memory, I am stuck.
Click into open source on the map (on android). I think for Ios it might be back gesture.
Same here, but the GitHub thread actually calls that out as a known bug. Swipe from the left. Intention is that clicking sonewhere on the map closes the dialog.
Click into open source on the map (on android). I think for Ios it might be back gesture.
Most of the time you can leave a note. Do you get an "Uh..." option?
No, there’s only the listed options (street has a sidewalk, or it doesn’t - no option for “sidewalk extends for half the street but not the other half”).
I don’t know if that counts as it does or it doesn’t.
On android there's a "differs along the way" option which let's you split the way.
There is a limit to the type of quest you can fit into the mobile app. It also should have a fall back "leave a note" option.
In this case, the intended flow is that you tap "Uh..." -> "Differs along the way..."
You will then be led to another UI in which you can split the road into several sections.
Aha. If you click 'Uh' exactly on its text, it works. If you click even one pixel away from the actual text, you get the chooser with not enough options.
First-time StreetComplete user here. I just answered a few of these "street has a sidewalk" questions myself. Upon first clicking on one of them, I was able to answer this question for each side of the street individually. The visual selection that you mentioned only applies to one side.
I can tell for sure the current version of StreetComplete for Android does allow specifying that only one side of the road has a side walk. I solved a few of these quests just last week. No idea what's the state on iOS though.
That's when you hit "uh..." and leave a note with details. In an effort to keep things simple on a small screen it doesn't include every possible option. You can read the wiki to find the best way to tag it. https://wiki.openstreetmap.org/
I click a quest and answer it but the app seems to be stuck on that quest and I can't figure out how to get back to all of the quests available near me. Is that a common experience?
This is a known bug, fixed but may be not released yet.
Go mapp!! For ios is free and very good alternative as well. made by bryce cogswell previous developer on sysinternals. https://apps.apple.com/gb/app/go-map/id592990211
I'm trying to get off of Apple Maps and into OSM (using Magic Earth as my navigation app) - seeing more OSM work is exciting.
Now I need to figure out if it's possible to merge my navigation maps with the deflock maps..
OSM editor Go Map!! also has a "Quests" overlay that is very similar to Street Complete (even uses the same icons). And you can also define custom quests for data you want to easily gather.
However, it's lacking the nice example pictures for each option - so, with Go Map!! you need to know what you're doing.
You should probably link to the relevant post
https://github.com/streetcomplete/StreetComplete/issues/5421...
I feel I must warn you. It is a gateway to more heavy map editing using the online editor or tools like Vespucci.
Good app. Some feedback -
I'd like to see native Apple sign-in, as opposed to via a WebView. When I used the WebView, the password text box was obscured and fiddly to fill in.
There should be an easier way to dismiss something I've clicked on, than having to tap "uh..." and then dismiss two dialog boxes.
Things that I've submitted (e.g. house numbers) should be visible on the map, to visually show my contribution progress (even if pending review or whatever). At present my contributions just disappear into the void.
Satellite overlay would be very useful to correlate houses.
Seeing some indication of other recent local contributions would be nice, to make the app feel more alive.
Yeah, dismissing a question I don't see an answer for right now easily is the worst. I think swiping the app closed is the easiest way out.
I wonder what some experienced OSM editors think when they see StreetComplete edits.
There is actually an edit history, the little arrow icon on the map pulls up your recent edits and lets you undo them.
Can you use OpenStreetMap like Google Maps?
I tried to use it to list restaurants but the UI reacted like nobody had ever tried that before
OpenStreetMap is a database, not a map application. The map on the official openstreetmap website is more of a demo to show some of the data. There are many different map apps that use OpenStreetMap data though.
https://cartes.app is trying to do this. OsmAnd is quite usable as well.
For navigation there's Magic Earth (can route around traffic jams), TomTom, and Scenic. OsmAnd can also do routing.
And we're adding new itinerary options soon.
For cars : avoid highways, tolls. For cycling : follow rivers, less noise, less traffic, etc; for walking : no steps.
Partly, with Organic Maps or CoMaps.
However, in my opinion TomTom (https://play.google.com/store/apps/details?id=com.tomtom.spe... or https://apps.apple.com/us/app/tomtom-maps-traffic/id14381065...) is still the best way to use OpenStreetMap for navigation since it combines OSM data with their own real time traffic information without dark patterns or asking for an account like Waze or Maps do.
Is TomTom's basemap entirely OpenStreetMap? Or do they add their own proprietary layer of map data on top?
It's a... mix. I can see all the shapes of the roads and areas i mapped in my area, but they seem not to trust OSM naming so they either have no name or the name TomTom gave to the equivalent street.
I can see most POIs i added with the exact name i gave to them, but the reviews and opening hours are pulled from Tripadvisor.
It's explicitly advertised for driving, not walking, so it's optimized to get you to places. CoMaps and even OSMand are much better for walking around.
And there's a simple reason for the search being so bad in most OSM apps: offline search.
The focus of all these apps seems to be a self-imposed "the app should never touch a server, ever, because what if the user is on a 100$/GB roaming plan in Africa", but indexing the maps is not easy.
I run a local Nominatim database (search engine for OSM) for Italy and it's ~57gb of data and indexes for one country only, and it barely supports fuzzy search.
I tried the TomTom app (on iOS) recently but found it incredibly buggy, especially when connected to CarPlay. The speed indicator would often just show 1mph, the dark mode would work about 20% of the time, and searches would show results from other countries ahead of a place 4 miles up the road.
For the web, I personally like https://osmapp.org
https://cartes.app I can also recommend.
Both projects have a similar goal, to offer a more useful UI around using an OSM based map.
It's https://osmapp.org - as in OsmAPP. Not to be confused with https://osmaps.org, the Ordnance Survey Maps.
There's also OsmAnd for mobile.
That's legitimately a good idea. I just contributed more to open street map in 5 minute than I have in 15 years.
nice to see StreetComplete coming to iOS. Its question-based approach makes contributing to OpenStreetMap feel approachable even if you’re not familiar with the underlying data model
I’ve been using EveryDoor so far, excited about trying this one out!
I like this project
OH long overdue, glad it finally came out
Is there some reason this can't be done in a browser? We need to stop feeding these app stores.
StreetComplete editing works also fully offline (if data was downloaded first). It also deals with fairly large amounts of data.
Not sure is it doable in browser.
The UX has some glitches... like when you click on an item, you HAVE to set the answer, you cannot go back to seeing all the markers...
You can swipe back on the bottom sheet for now. It's a known issue and is being worked on!
Nice !
> The code base of this app is written in 100% Kotlin
weird choice, from 2023, to code an iOS app. but it doesn't matter today in 2026. frontier models can manage weird software stacks
It is a multiplatform app with one codebase, compiling to both iOS or Android.