Redoing edits that used a prohibited source

If a source, such as Google Maps, is used to make an edit, as evidenced by it having been cited as the source in the changeset information, and that edit is reverted, under what conditions is it acceptable to redo the edit? Is finding another source alone adequate? Google Maps terms of use (and many other sources) broadly talk about “using” it to improve another map as being prohibited, and once a piece of information has been discovered through Google Maps, arguably using that information to find another source that is compatible with OSM constitutes “using” it to improve OSM. Might it be best to allow another mapper to make the edit at some future point relying completely on sources compatible with OSM?

As far as I know, using GMaps to discover something and than do, for example, a survey, is totally legit. Only the direct use is prohibited.
Used in that way I would not mention it as a source. In the case that it is mentioned as a source, it is a strong indicator for a revert.

If you speak of reverting CS, a link to an example might be helpful, for a better answer.


@Robert46798 Thanks for your reply. Google Maps terms of use make no distinction between “use” and “direct use”:

Prohibited conduct… use Google Maps/Google Earth to create or augment any other mapping-related dataset (including a mapping or navigation dataset, business listings database, mailing list, or telemarketing list) for use in a service that is a substitute for, or a substantially similar service to, Google Maps/Google Earth…"

We deal with quite a lot of this sort of thing at the DWG - here’s what I tend to do:

First things first - I’d make sure that Google Maps really was used as a source. It’s more common than you might think that someone familiar with an area will mention “Google Maps” to indicate “it’s not just me that is saying that 23 Railway Cuttings is so-and-so - Google backs me up”. If Google really wasn’t used as a source I’d tend to comment to that effect on the changeset. Sometimes people use “Google” as a synonym for “aerial imagery” - with a DWG hat on I’ve definitely seen data in OSM allegedly from “Google” that didn’t match anything from Google but did match OSM-compatible imagery in iD etc.

If Google was used as a source then it needs to be both reverted and redacted** (just a revert isn’t good enough on its own) so it needs to be reported to the DWG so that the “redaction” bit can be done.

Depending on how much was reverted, it may be feasible for us (or someone else) to have a go at remapping.based on OSM-compatible sources. To be clear, this is done after the initial revert.and redact; we’re not saying “what’s available to OSM is good enough so that it might as well stay”; we’re actually remapping from OSM-compatible sources.

Another fairly common situation is where any Google use is a large number of edits ago - I’ve seen examples where there is literally no Google IP left - for example, the name of a street has been changed, and all nodes making up the geometry and been changed; the only thing left is the trigger’s broom way ID in OSM.

** with a small number of exceptions, including some Google-provided Haiti earthquake imagery.


I can’t imagine a situation where because I have seen (at a copyright source) that something may exist that it is prohibited for me to than survey that in OSM. In that situation I would not be “using Google Maps/Google Earth to create or augment any other mapping-related dataset” - I’d be using my survey.

As a trivial example, I found out that this pub had closed via a couple of non-OSM-compatible sources, but I verified it by looking out of a train window on Monday afternoon.


Just having looked at Google Maps may not prohibit us from a survey, but systematically comparing OSM to Google Maps to generate a list of places to survey (or do other research), would seem to constitute “use” of Google Maps to augment OSM. In the case just described, making use of Google Maps certainly makes the job of the OSM mapper easier, and/or allows them to do more edits.

The cases I have seen are more like someone having added a business (or many businesses) based on Google Maps, when a comment is made on their changeset that Google Maps cannot be used, then goes and finds other sources that support the edit and could have been used to make the edit and then says all is ok. The other case I have seen is where many sources are listed, and the mapper insists that even though they “used” Google, it wasn’t really necessary, since the information is in all of the other sources, and therefore all is “ok.”

I start reporting these (I did report one, but I don’t think it has been acted upon yet). I assume you want to do the revert as well as the redaction, or should I revert and then report?

Either would work - some people do the revert themselves, some people don’t.

I submitted a couple of small changesets to DWG for redaction but was told that a revert (which I had already done), was sufficient. Could the DWG provide guidelines as to when redaction is should be requested (beyond a changeset clearly using a source that is not allowed)?

Do you have a ticket number for that? As I said above, it can vary depending on the individual situation.

Admittedly this is regarding two very simple edits, but it seems very clear that the information was just copied from Yelp, which I don’t think is allowed.

Given that the word “use” is quite vague here, one could conclude that anyone who uses Google Maps at all in their daily life cannot be trusted to contribute license compatible data to OpenStreetMap. How are we to know if some local knowledge they cite actually originated from Google Maps! If I eat at a restaurant I both found and navigated to with Google Maps, and then add it to OSM, I definitely used Google Maps as part of the OSM contribution process. Is this against the terms, or does my local knowledge from having visited and eaten there override the knowledge obtained from Google Maps? How about if I didn’t eat there but I just walked by on the street while navigating with Google Maps? How about if I didn’t visit the restaurant and used Bing Streetside or Mapillary imagery to add it OSM, but the reason I knew to look at street level imagery is that I happened to see it while using Google Maps a week prior?

I consider all the above questions fairly absurd and pointless to worry about. Not to mention quite impossible to prove whether Google Maps was “used” or not. We can just use common sense here: copying from Google Maps is not ok. Copying from other allowed sources is ok. No need to worry about whether the mapper looked at Google Maps first and now somehow has tainted knowledge.


Not a lawyer here - My simple take on what might be actionable:

  1. Violating terms of service
  2. a) Copyright infringement
    b) Unfair competition

I do not think 1) wont ever be acted upon. Some samples mentioned above belong to 2a) e.g. copying names and places of restaurants, shops etc. Is that direct use? This might be different in the US and the EU. Some belong to 2b) e.g. collecting names and places of restaurants to make survey less work. Section 2b also covers using StreetView to trace information from the photos. That might be mere use?

Copyright law is very strict. Mappers must not infringe; Isn’t that also against the terms of use of openstreemap itself? Trade law less so, but still need to be wary of it. Using StreetView certainly not fine. Redoing edits on the base that they were tainted/reverted - perhaps more severe than just list creation - so rather not?

PS: Using google maps or any other map/directory to find a restaurant for dinner, later mapping that into OSM if missing is not using these maps to augment OSM, it is using them to find a restaurant :wink:

I confess that I (years ago, total osm newb) have used a source (by copying from a map to a list and adding details in csv) with unclear license. The map was intended for locating waste containers, operated by the local municipality. I uploaded some waste containers in OSM via JOSM and subsequently checked them by actual survey, taking snapshots of all of them. I now regret this workflow, I should have surveyed first and uploaded later. Since we have waste containers and chutes in almost every street, the survey work would have been the same, and I could have surveyed a million other mappable objects at the same time, within just a few weeks.

However, I am not inclined to reverse those edits, only to add them all again.
This particular source is no longer operational.

Later, for a comparable POI-import, I used a source after explicit permission. Also, import first, then recheck all objects later. Most of these objects already had an OSM entry, so it was mostly a conflation to add a few details (extra details, not present in the source dataset).
Later I found out that the dataset owner had gathered his data from other sources with unclear licenses (ie mostly: no license statement), so in fact he was not in a position to grant permission.
Again, I would nowadays handle the workflow differenty, but I am not inclines to reverse these edits.

Both events were discussed and documented in the local community. In hindsight, I should have paid more attention to the objections, even though the majority approved the edits.

I agree. And terms of services might not be an issue in many cases. E.g., I do not use Google Map or Google Street View. But I do use StartPage sometimes. And it does happen that I study the website of restaurant or real estate developer and there is a map on the website (how to find us). And that map might be based on Google or some other map.

As for Copyright, facts are not protected. However a database of facts can be proctected (In the EU by the Database directive).
So if you see on a map that e.g., a restaurant is in the building north of the lake, that is a single fact.

Sometimes you can find a webpage for e.g., a restaurant that has an embedded map (Google or something else), and if you check the HTML source of the restaurant webpage you can find a link that has the coordinates of the restaurant in the URL. Those coordinates are not the property of whatever map they use, they are published on the restaurants website just like the phone number, address, etc.

First, “use” and “direct use” are not copyright terms. “direct infringement” might be.

Second, if someone contributes to OSM without violating copyright laws, but might be violating some terms of service that they may have agreed to, is that really a problem for OSM? And would reverting the edits change anything at all for the contributor that has agreed to the terms of service, if a violation had indeed already taken place?

I understand that we want to stay extra clear of competing mapping products for many good reasons, but that does not mean that there is a real legal issue.

By only focussing on copyright you unfortunately dismiss trade law. All of a certain, no single OSM mapper will ever get called before a court due to violating terms of service. But OSM consumers might get into trouble by the actions of the OSM contributors.

1 Like

Agreed. And if consumers start getting in trouble, that becomes a deterrent for other potential consumers, reducing the usefulness and viability of OSM overall.

1 Like

This Changeset: 149776734 | OpenStreetMap is by a new mapper (26 changesets) who was trying to mark restaurants as permanently closed. The source given is just “Google” in the first changesets. After @tekim and I commented a changeset the mapper used “Newspaper and Google” as source for this new changeset. We both explained the problem again in the comments, and while the mapper sadly never wrote back, they seem to have stopped adding anything for now.

I’d like to hear the community’s opinion on the 26 changesets. Should they be reverted and even redacted as @SomeoneElse had hinted?

It seems likely that the mapper was using Google to find restaurants over a big area which were marked as permanently closed there. But maybe they did have other knowledge first, like the mentioned newspaper, and only used Google to confirm?

Since I don’t trust Google or old newspapers articles to be accurate about the current situation, I would rather see a revert / redaction of these changes, and maybe add a fixme or a Note to alert local mappers to go and check OTG.

1 Like

@daganzdaanda Thanks for bringing this to the community’s attention. Just to clarify, the mapper in question is citing “Google Maps” specifically on this changeset as one of their sources, not just “Google”

Recently, Friday 19th a Café reopened here after having been closed for seven (7) years. Been there on Saturday for a Pernod or a Ricard, You name it. Just out of curiosity, if Big-G already knows about that, looked it up and got told – Permanently Closed. In OSM data it was open all the time :slight_smile: I got reminded of open-historical-map.

PS: The café does not show a POI in google maps. But searching for it reveals the location and the closed state.