Some people were quite excited about recent announcement of built-in OSM editor of navigation app, but now, first version of it is in use for many weeks, bringing up more and more complains about users doing all kinds of unwanted things.

Since pointing on negative facts makes certain people feel offended, I want to clearly explain my position. I agree, that OSM could have more contribution from people, not deeply involved in project. And general idea of enabling them to do such contribution using their favorite navigation app is close to perfect. is the first widely used app helping its users to contribute to OSM. This app does have certain nice features such as opening hours entry. I also realize, that it’s only the first version.

However, there are certain flaws, making its future look not that optimistic (at least - in given current situation). I don’t want to try arranging these flaws by its importance, but I’d like to list some.

One fundamental flaw is that authors of this app don’t seem to care about demotivating effect of “rubbish edits”. There is common mantra: “We need more OSM contributors”. Yes, it’s true, but only in certain ideal case, where new OSM members are at least thorough and responsible. I mean, if some newbie doesn’t know how to do something right, he should learn it after being pointed on documentation or receiving some explanations in changeset comments/private messages. If he doesn’t want to improve his quality of contribution, all responsibility for data quality gets automatically shifted to responsible OSM members overseeing particular territory. They have to fix every mistake to keep data tidy. And they have limited amount of time, good will and energy to work on OSM. Forcing them to take care about unusually large amount of bad data (including investigations using WhoDidIt, achavi and other tools) simply reduces amount of contribution they making. Keeping in mind huge difference between their productivity and productivity of general newbie, it’s not an equivalent exchange, because single amenity, wrongly added by a newbie, steals time, enough for much larger contribution made by experienced person. I call it general fundamental flaw in attitude of authors towards the whole OSM community.

Another flaw is that hosting on GitHub currently seems more like a gesture. I’m not talking about code, I’m talking about interaction with community. @Zverik, as a member of development team, has confirmed, that they have internal bug tracker, completely separate from GitHub, and that developers rarely checking GitHub issue tracker, however, they are reading bug reports, sent to dedicated email address. This situation makes it look like developers are non-responsive. It could be not completely true, but since there is no reaction from their side on GitHub, it’s not unreasonable to conclude that.

Recently, an interview with one of authors was published, and it sounds like major ideologists are going to leave this project in favor of own startup. Which probably means, that less people will actually work on development planning. I don’t believe, that it will improve responsiveness of the whole team.

I know very well, how sensitive authors can be, when it comes to critical feedback from users. But if you can’t treat it as useful feedback, that job isn’t for you - development is not a kind of comfortable place for sensitive unrecognized genius. Indeed, angry OSM members, tired of cleaning after users, could be pretty impolite, but at least certain comments contain useful suggestions such as, for example, separating editor workflow into it’s own portion of UI. However, @Zverik recently called all those comments “non-constructive whining”. It’s typical, and it always leads to greater separation of developers from users. In psychology, it’s called “avoidant behavior”. I don’t know, if it’s just his thing or working style of the whole team, but anyway.

From the point of view of UX, there is an issue with built-in OSM editor. Users of, except ones who already knew about OSM or who learned about it independently, usually have no idea what exactly they doing when editing the map. Automatically, they have no idea about project guidelines and so on. As I’ve already mentioned once, you can’t make people edit OSM without telling them what it is. But now, they are a kind of lured into doing that. Therefore, their ideas of what they actually doing are often quite imaginary. Judging by many different edits, it’s easy to conclude, that people thinking that it’s just their local copy of map, or those edits are some kind of bug reports for professional mapping team, or these are just their own custom POIs. User interface provokes it even more.

Since UI/UX provoking systematic wrong edits of certain specific types, it must be fixed instead of blaming OSM members for their “lack of patience”. Leave aside real programming bugs, such as breaking opening hours in certain complex situations, replacing “№” sign with “N” and many others (including ones, left without any response on GitHub issue tracker).

My personal view on this (and I’ve expressed it long time before the very first version of MapsWithMe/ is that only way to allow people, who have no idea about OSM, to contribute is to limit available actions even more and to introduce more checks (such as duplicate check). It is way more complex problem than just development of an editor, since larger coverage of people, completely unaware of OSM, obviously brings more problems to solve before it will become effective and not demotivating for OSM members.

Comment from Alan Trick on 21 June 2016 at 21:55

I’m hoping in the long run it will be a good thing. Even now, I’ve seen quite a few good changes from users, but also a large number of shitty changeset descriptions (e.g. “Edits via MAPS.ME app, a partial upload”) and strange map notes.

Comment from BushmanK on 21 June 2016 at 22:12

@Alan Trick,

Personally, I’m not into “hoping”, since often we can take in account certain visible tendencies. Here, tendencies are not good.

Changeset descriptions are automated, and “partial upload” means there are many objects changed or added. It doesn’t bother me much, if those edits are okay. But what bothers me, is that generates constant flow of bad edits of similar nature, which means it’s preventable (unlike in case of random mistakes) and nobody actually did anything to assure community, that measures will be taken to address it.

Comment from Zverik on 22 June 2016 at 05:51

We have a release cycle of one month. Therefore we can’t fix out mistaked quickly, except for the big ones.

Alan, “partial upload” means exactly that: doesn’t know in advance which changes are being uploaded, so it updates the changeset comment after it has done. And sometimes user’s internet connection get broken, resulting in such partial uploads with non-descriptive names.

Comment from thermo_nuclear on 22 June 2016 at 06:17

Потлатч-то чем вам не нравится? Гораздо удобнее рисовать прямо в браузере, с удобным и понятном интерфейсе. В Джосме разобраться совершенно невозможно, и там нет никаких других функций, которых нет в Потлатче. Единственное только можно посмотреть, сколько процентов объектов кто-то нарисовал в городе. Новичок в Джосме накосячит с гораздо большей вероятностью, чем в онлайн-редакторах.

Comment from Zverik on 22 June 2016 at 06:45

thermo_nuclear, хорошая попытка троллинга, но нет.

1) Новичку в JOSM накосячить значительно сложнее, так как в нём есть отличный валидатор, и редактор не пытается скрыть модель данных.

2) В этой теме обсуждаем редактор для мобильных устройств.

Comment from SOSM on 22 June 2016 at 08:10

It should be noted that we’ve had apps with similar characteristics in the past and present (with very much the same conflicts), for example OSMand and the various incarnations of the wheelmap editor, and not to forget the original Mapzen POI collector (nothing to do with the current company of the name).

The main difference is that has undoutably marketed the app more agressively and naturally that the potential user base is simply far larger nowdays (because of increased OpenStreetMap quality, density and usefullness of our data). The later point naturally increases the potential for conflict even more.

Were I do agree is your last point, which I’ve pointed out time and time again, writing a “simple” editor that can be used without at least some knowledge of the project and the underlying data models is extremely difficult, so difficult matter of fact that nobody has actually produced one to date.

Comment from saintam1 on 22 June 2016 at 09:29

Hi BushmanK,

I’ve had to clean up after users on a daily basis since the app started allowing edits, but I, for one, welcome this new wave of editors. Yes, you need to go and tidy up after new users, and it does take time, but I think having a more diverse contributor base is worth the effort.

Most of these users will make <10 edits and never come back. Some will get hooked and become active long-term contributors. Along the way they’ll learn the rules (hopefully sooner than later). But every expert user started out with a clumsy first edit.

I’m reminded of Linus Torvalds’ uncharacteristically positive mail about newbie users submitting small, trivial patches to the Linux Kernel, and these patches often turn out to have obvious mistakes in them:

Key quote, adapted to this discussion: “Yes, those [ edits] are a bother. Damn, they are horrible. But at the same time, the devil is in the detail, and they are needed in the long run. Both the [edits] themselves, and the people that grew up on them.”

That’s how I see it anyway.

Comment from Jean-Marc Liotier on 22 June 2016 at 09:52

Love saintam1’s Torvalds quote. Even in Dakar I now see misguided POI additions sprouting but I try to engage them pedagogically - those may well be our future heavy contributors… And at least, a badly described POI (with a wrong position because it was recorded where the user was standing instead of where the POI lies…) is less bothering than the Potlach highway spaghetti I’m used to see… But maybe it is just that I’m not seeing many yet.

Comment from Richard on 22 June 2016 at 12:27

@jml: dude, seven years of knocking Potlatch and you still haven’t learned to spell it correctly ;)

Comment from Jean-Marc Liotier on 22 June 2016 at 13:16

@Richard And on second thought I’m even more ashamed of the awful generalization about its users… Oh well - everyone have one of my nice kula necklaces !

Comment from SomeoneElse on 22 June 2016 at 15:36

“a new evil”?

The news headlines from around the world over the last couple of weeks have been pretty miserable for lots of reasons, but apparently the thing that really matters is some people mislabelling “tourist attractions” on an online map.

Comment from BushmanK on 22 June 2016 at 21:40


We have a release cycle of one month. Therefore we can’t fix out mistakes quickly, except for the big ones.

Isn’t it the first time you telling that (speaking of more or less proper public relations)? And what about tests, are you actually using them?

Comment from BushmanK on 22 June 2016 at 22:00


It wouldn’t be just a figure of speech to say, that people, including OSM project members, are very different. And there are people with views like yours, who are okay with spending their time on “garbage collection” by own reasons. But the reality is that definitely not all OSM contributors are like you and it’s completely natural. Attitude is very individual thing and nobody can demand people to change it, since “not wasting own time in favor of people, who don’t actually care about OSM” is absolutely normal.

I’d also say that quote you mentioned doesn’t fit this situation, obviously. Simply because those contributors Linus was talking about were, at least, aware of what they doing. I mean, they knew, that it’s about Linux and so on. But in case of, people are very often not aware of what’s happening. Reversing your analogy back to Linux development, it would turn into completely ridiculous situation: portions of Linux kernel source are, for example, showing up in people’s favorite opensource text editor with a prompt like “if you feel like something is wrong with it, feel free to edit it”, and the whole bunch of people, who have zero programming knowledge, just playing with that code, and then every edit turns into patch and pull request. Guess, how soon Linux developers will just quit. So, again, be careful with analogies, it’s too easy to transform something originally logical into fancy ideology.

Comment from BushmanK on 22 June 2016 at 22:09


Title refers to some very simple thing: for certain period, Potlatch was known as an editor, most often used for unintentional relation damaging. Now, any changeset is suspicious.

Comment from Omnific on 23 June 2016 at 14:46

I’ve got to disagree with this assessment on several fronts. First of all, I’d say the number of bad changesets coming out of is probably on par with the number coming out of iD, mainly because new editors make mistakes. This is not inherently a problem related to

Second, I think it’s important to get more engagement from the community. That’s what OpenStreetMap is about. It’s not about a walled garden, controlled only by a select few power editors. Sure, people make mistakes when they start, but if they care enough to continue, the quality of the edits increases with time.

Finally, as a power editor, I’ve found the editor to be a fantastic addition to my toolset. I’ve edited well over 3500 items in it, as it makes minor POI edits far more simple and immediate.

There may be room to improve (a duplicate checker would be a good start) but the tool is a great way to bring more participation into OSM.

Comment from BushmanK on 23 June 2016 at 15:04


I have it already explained, but I can repeat it for you. For some experienced mapper, or even for a person, who just barely knows something about OSM, is no worse than any other editor - I didn’t say anything opposite to this statement.

But it does have inherent issue: it allows certain amount of people, who know literally zero about OSM, to edit OSM database directly. They are not a part of community. They don’t even know that OSM community exists. This situation creates several types of wrong edits, barely (or even never) found in changesets made by other OSM editor applications.

Comment from joost schouppe on 23 June 2016 at 15:38

And the same goes for iD. The first time you hit that Edit button on, you know nothing at all about the project. Hell, three years after that, I still feel like I know next to nothing. We even have a whole team needed to keep us up to date of only just the most important stuff!

That said, of course we need to make the tools as good as possible. The editor is obviously a very good idea, even if it has many faults at this time. Let’s not waste time complaining and demotivating the makers of this great tool. Let’s just give them constructive feedback on how they can make it better.

Comment from BushmanK on 24 June 2016 at 16:59

@joost schouppe,

No, situation with is different from one with iD, because provides a tool, way more isolated from OSM community.

First, users not necessarily receiving private messages and changeset comments, since some of them could use one-time emails they never check after completing the registration (in case of iD, you will always see message notification, if you’ll get back to

Second, the whole UX concept is different: isn’t an editor, or a part of “something OSM” - it’s a navigation app. That’s why certain people thinking that they adding their own POIs or sending feedback to professional mapping team.

Third, map in doesn’t change immediately, it has own (relatively long) update cycle, which is also different.

Fourth, iD provides “full immersion” into map, you can (if you want) see all tags and geometry of other objects. Indeed, you still don’t know much about OSM rules, but you can learn from examples way better. And if empirical evidence works for you, I’ve seen all kinds of mistakes, made by using Potlatch and iD, and majority of them were caused by genuine lack of knowledge (except relations, broken by Potlatch users, where it was provoked by its interface), not by completely wrong idea of what people are doing.

Speaking of demotivating developers - I don’t think anything is demotivating them more than they naturally are, since it doesn’t seem like they are interested in feedback at all (read again that part about GitHub issue tracker and another part about several members of development team already leaving for their own project).

I can’t be responsible for other people (and there are cases in Russian community forum section, where people even used obscene words to express their disappointment caused by those massive wrong edits), but I consider my feedback constructive, since I’m not making emotional general statements, but referring to specific real flaws and mistakes in this situation.

Comment from marczoutendijk on 27 June 2016 at 22:23

In the Dutch community we send out a personal welcome message to new mappers after they do their first edit. Mappers working with iD, JOSM or Potlatch are usually aware of the OSM way of communicating with each other by means of mailing lists, forums or personal messages.
But a user is not aware of the fact that I can send him a message through his account page, nor is sending out alerts to users that a message for them was posted on their account page.
Recently we saw an enormous increase of bad edits caused by users, but when trying to get in contact with those users, this seems to be impossible: they never reply.
Hence we stopped sending out the welcome message if that user used as editor.

I do agree that might be a useful addition to the mapping tools, and I do use it myself to do quick and dirty POI adding which I later (at home using JOSM) improve by adding the stuff that you can’t add with

Comment from BushmanK on 27 June 2016 at 22:43


Thank you for your comment. Your observations are consistent with complains of other people, who tried to interact with users. Saying that nothing bad happens is either ignorant or it’s just an expression of pollyannaism.

Comment from Kilkenni on 28 June 2016 at 16:53

@saintam1 > But every expert user started out with a clumsy first edit.

I didn’t. Does that make me a non-expert?

A good description of the problem we’re facing.

Comment from BushmanK on 28 June 2016 at 18:39


I support your point completely. Even if we’d ignore the fact of non-existent “expert” definition, original statement is false. There are people, who prefer to learn first than - to start doing something. I did that before making my first edit (with Potlatch, by the way) and I’m still doing that every time I face some new problem. However, there are people, who prefer doing something first. I’m not judging them while they don’t think that someone else should clean up after them.

Comment from Blackbird27 on 30 June 2016 at 09:07

My suggestion for developers is at least to add a tutorial with extensive explanations (preferably with pictures like on their welcome screen) that will show a user what he about to do and why it is important to add correct data. A wise thing to do is also to add links to the OSM wiki and suggest editors that could be used on PC as well if the goal of developers to involve as much OSM contributors as they can. Some way of communication with OSM community should be implemented, a pop-up message like in JOSM or iD Editor will do. Tutorial may be added on the authorization screen, and only after completion of tutorial user should be able to edit. Even if development cycle is one month, developers should find the way to update the map daily or weekly, that will help with POI duplicates created by users (I deleted 4-6 duplicates in some places created by the same user). For example OsmAnd+ has a paid feature OSM Live where you able to get partial updates of the map data daily or weekly. As a workaround I am using latest beta builds to get latest map data.

Comment from dieterdreist on 30 June 2016 at 09:41

Nice writeup, BushmanK. I agree with your points. The main difference between editor and the ones we’ve seen in the past is the userbase. The amount of edits is unprecedented, and it takes more ressources of the established contributors to fix all the problems that get introduced.

I believe should fix very quickly: - the internationalisation problem (names overwritten or amended with foreign/translated names) - remove whitespace at start and end of values - integrate current data to avoid conflicts - better normalize tags in general (wikipedia, website, phone)

Also the measures suggested by Blackbird27 (more community integration, tutorial/text about osm, links to osm wiki) should be enabled.

From an Italian perspective, it is also problematic that users are encouraged to add housenumbers to buildings, because they only apply to entrances and buildings have many housenumbers.

A very frequent problem I’ve seen lots of times are POIs in the middle of the road, even if the buildings have already been present in OSM.

Login to leave a comment