Talk:Legal FAQ

From OpenStreetMap Wiki
Jump to navigation Jump to search

The OSM wiki is not the principal means of communication within the project. If you want to ask a licence question, we suggest you use the legal-talk mailing list, not this page.

Why is it illegal to copy other data

I understand that for works of art, they are protected by copyright, however a map is a collection of facts. Facts are not copyrightable. For discussion see 499 U.S. 340 (A supreme court case where copying records from a phonebook (including bogus marking records) was not illegal. While I am unaware of international law, is it true that facts are not copyrightable internationally? (Is this a bern convention thing?)

"Facts are not copyrightable" is indeed a US viewpoint. It doesn't necessarily apply elsewhere. In addition, in the EU there's the concept of "database right" to consider. I'm currently looking up a load of case law on this and hope to have more to report soon. --Richard 11:45, 18 February 2007 (UTC)

The Wikimedia Foundation does not own the copyright to user contributions. That part is false and should be corrected. --LA2 23:46, 21 Mar 2006 (UTC)

I added that part because someone asked that "we should form a foundation like wikimedia that own the copyright". I know that this is wrong for germany, googled a bit in other wikis and got the impression that it could be true for (by telling from the discussions about licensing). But I did not found any real proof. So I put that "most" in.. :-].
So what is the status of the wikipedia? How can they do any modification in the license? Do they have to (like in germany) ask every contributor whether he agree? --Imi 09:49, 22 Mar 2006 (UTC)
About this, I just want to say that a french court decided recently (November 2007) that Wikipedia has to be considered as a web host and not as editor. He is therefore not responsible for the comments writen by the contributors. Pieren 15:20, 10 Nov 2007

This page is a great idea to summarise and resolve the licence issue, but there are a couple of statements which are (IMHO) closer to opinion than fact: (re: a geodata licence) "CC-By-Sa is very close to this"; and (re: the spirit of the licence) "Unfortunatly, some restrictions has to be accepted to make sure of this".

I'd also be a little wary of using the word "free" - it's a fairly loaded word in licence discussions. A GPL advocate's understanding of "free" differs from a BSD advocate's, and so on.

But again, the page is a great idea, and full credit for trying to nail everything down in one place. --Richard 15:49, 22 Mar 2006 (UTC)

I was just involved in a US lawsuit regarding this. Map data is not copyrightable in the United States, only the style of the map and its contents (fonts, color, coverage, etc) - so a map covering the entire world would be exempt. Tracing satellite photos copyrighted by other companies and using data derived from them is also OK in the United States. TeleAtlas and Navteq have no copyright to the data in their photos, only their actual photographic works. Trap Streets have no bearing on copyright, because they are published as data. See: Feist Publications v. Rural Telephone Service US Supreme Court Case (1991) and Alexandria Drafting Company (ADC) v. Amsterdam, US District Court (June 4, 1997)

ADC placed trap streets in their maps which were copied by Amsterdam. ADC sued and lost. The court ruled their trap streets were not original and were published as fact and thus were not protected.

--Austrini 15:41, 12 June 2009 (UTC)

With respect, we do know all that. Please read up on the legal-talk archives, the licence pages on this wiki, and relevant postings on opengeodata etc. OSM is not just (indeed, not mostly) a US project; Feist v Rural is very useful and definitive on purely factual data, but there's a lot of grey areas - see especially Mason v Montgomery Data for a fuller, post-Feist treatment of map data; and so on and so forth. TeleAtlas/Navteq/whatever data is not just protected (or not!) by copyright, there's also contract, trade secret and EU database right to consider. --Richard 15:58, 12 June 2009 (UTC)
I used commercial satellite photos to map a series of African cities for a non-profit group, was sued in the EU, and won, because the photos were representations of data - the judge mentioned originality in respect to Berne. TeleAtlas sued a company I worked for in Florida two years ago because we used google maps as a source, they lost with the judge citing ADC v. Alexandria. I suppose my point was that OSM seems very narrow in your interpretation of copyright.--Austrini 18:06, 12 June 2009 (UTC)
The point: you were sued and had to spend some resources to defend yourself. Osm has traditionally taken the view that it cannot risk getting sued at all. Alv 22:45, 12 June 2009 (UTC)
Yes, I was. In the first instance I spent no money, and was represented pro bono. In being sued my resources were made stronger and my future products were made better. My winning the lawsuits opened doors for other non-profit groups to use geographic data freely from sources that were previously off-limits to them out of their fear of being litigated against. Much of Africa has been mapped in this way now. Red Cross medical teams can get around obscure places like Okonobi or Bumba because of it now. It's not that geographic "data" should be free, its that it is free. It's the colors, the fonts, and which roads are major and which roads are not and things like that which make it murky. I'd like to see OSM allow for tracing of satellite images, at least, you'd have a stronger information base and the legal stuff has been done before with little ill effect. I wont make any more posts on the topic after this, and i'll go get out my survey GPS unit now. I hope one day a more proactive stance on copyright issues would be addressed here, though. --Austrini 23:31, 12 June 2009 (UTC)
On the question of satellite imagery may I 'umbly refer you towards my blog posting of last week - . cheers --Richard 08:26, 13 June 2009 (UTC)


What do you need to do to "attribute the work in the manner specified by the author or licensor" for OSM? (e.g. I presume we don't need to put the name of everyone whose tracklog appears on the map?)

Do we have an equivalent of

Ojw 15:51, 28 Apr 2006 (UTC)

Yeah I think we should look at how it works on wikipedia. On the wikipedia website itself, users have their work attributed to them on the history tab display, and also the 'my contributions' display. It's a small attribution, but it helps give individual users a sense of "I helped to build this" achievement.
The articles are released with an open licence which requires attribution, but the most widely accepted interpretation of the GFDL license is "You may be able to partially fulfil [attribution obligations] by providing a conspicuous direct link back to the Wikipedia article" (see Wikipedia:Copyrights#Reusers' rights and obligations). The GFDL liscence isn't clear on this point, and presumably the cc-by-sa2 license isn't either, because they're not written with multi-user community contributions in mind, but this is the interpretation which wikipedia people have settled upon by consensus (at least among those who have taken the trouble to look into it). Presumably the exact interpretation the courts would take is still unknown. The point is, people are happy to count a link back to wikipedia as sufficient attribution, after all, anyone following this link will be able to find the history display which attributes all the users and their contributions.
Here on OpenStreetMap we don't have a history display yet. I think this means we are technically in breach of our own license, since we aren't attributing mapping contributions. But I think it has been discussed before. We are gathering the necessary user/timestamp data, but we're just not displaying it yet, due to privacy concerns or something. I think it's something that should be added somehow. This way attribution could work the same way as in wikipedia, couldn't it? i.e we require a prominent link -- Harry Wood 13:10, 8 February 2007 (UTC)
This is a very old discussion. We've since added a 'my edits' display, based on changesets, akin to the "my contributions" page of wikipedia. The attribution requirement has come to be sort-of-semi-agreed-upon by the community too (as documented on this page) -- Harry Wood 00:35, 6 April 2010 (UTC)

With an increasing amount of data being derived or imported from other free datasources that have attribution requirements, I wonder if our attribution and that of sites / applications using OSM data is sufficient to these third party imports? Currently these attributions are on the Contributors wiki page. However, that wikipage is not linked to from (hardly) anywhere. Should for example our attribution recommendation on Legal FAQ state that "and contributors" link to the contributors wikipage, just like "OpenStreetMap" links to "" and CC-BY-SA links to the creative commons page? -- Amm 12:08, 5 April 2010 (UTC)

My personal preference is that we replace the CC-BY-SA image on the front page with a link to, which would: explain the current licence and point towards its text; point towards the Legal FAQ; list large contributors with their preferred boilerplate text (as per Attribution); explain how to find all contributors for a given area ('History' tab). When I get the proverbial spare minute I'll submit a patch for this. --Richard 08:23, 6 April 2010 (UTC)
That would be great. But there are imho two parts to this "problem". One is how do "we", as in the OpenStreetMap webpage, comply with the attribution requirement, which is the part your patch would "solve". And then there is the part of how do we make sure (i.e. recommend appropriate wording) all the (second, third, ... tier) downstream users still comply to the requirements in a sane way without becoming a legal nightmare to use openstreetmap data and having to sort through source tags and the history to figure out what to attribute. The second part was what my suggestion was mainly about, although I did want to initiate a discussion on both parts. -- Amm 09:02, 6 April 2010 (UTC)

Compatibility with which licenses?

I think it would be great to add more info on the compatability issues. For Example can I take Gnu FDL licensed data? Sorry, if that is answered somewhere, but I couldn't find it. I'd appreciate info on this.--spaetz 12:41, 21 March 2007 (UTC)

Copyright on extracted data

  • Moved here from the article page. --inas 06:31, 3 September 2007 (BST)
I'd be interested in legal clarification on the status of extracted data. As far as I know, statistical data (e.g. word usage data) derived from a source like wikipedia is not covered by the copyright on the corpus. It wouldn't surprise me to find there was similar legal protection/exemption for certain kinds of analyses of spatial data. But there might not be. TomC 15:53, 22 Mar 2006 (UTC)

Added disclaimer

Ok, so 21 months on, I can stop being diplomatic (see above).

This page hasn't worked out well, at all. So much of it is people's opinion; there are some opinions presented as facts here which are disputed by a very large percentage of the OSM userbase (who are, after all, the copyright holders). To claim these are definitive answers to FAQs is misleading and potentially very damaging.

I have added a disclaimer for all but the first two entries, which at least are undisputed facts. I strongly suggest that we move anything which could possibly be disputed - probably 90% of the rest - into a "Common interpretations of the licence" page to avoid giving the impression that it is fact.

--Richard 12:16, 19 December 2007 (UTC)

I don't think the stuff you moved gave any hard and fast position. It mostly directed people to look at whether the resulting work was "derived" or "collective", which are the actual words used in the licence we use. I think it could still be fixed, rather than devalued to be surrounded by bold disclaimers. All of what is written on the subject is going to be somebody's opinion. --inas 22:48, 13 January 2008 (UTC)

OSM Data not copyrightable?

It was mentioned in the OpenStreetMap blog that it's possible that the OSM license (CC-BY-SA) is not valid for the reason that geographical data is considered as database data and protected by database rights in Europe, and not by copyright. The idea is, data (i.e., facts) is not copyrightable and since OSM data is data, then it can't be protected by copyright licenses like CC in the United States but could be protected by database rights in Europe, but it is currently not protected by such.

I'd like to argue that the OSM Data is actually copyrightable since there are elements of creative expression in the underlying data. For instance, the selection of which points represents the curvature of a road is arguably a creative process and is thus copyrightable. The exact shape of a road is a fact, and not copyrightable, but we do not capture that fact but instead we "create" a representation of that fact through the judicious placing of OSM points. Am I making sense here?

Unlike an alphabetic listing of phone numbers which is pure data, OSM Data is not just data and there is creative input, thus I think that the OSM Data is still copyrightable and is not in "a curious unlicensed limbo at the moment" like the blog post says.

--Seav 05:04, 14 April 2008 (UTC)

"..and all data created by use of any tools which connect to is to be licensed.."

I feel that the current statement on OSM sign-up page is incredibly wide. I guess the spirit is that everything to be uploaded to OSM is to be licensed, as well as all the derived works using OSM data. But all data by anything which happens to connect, that includes everything done with a web browser that has been used to check out OSM front page. Does anyone else feel that this should be clarified? Why is that "all data" addition needed anyways? -- Jesh 16:17, 6 July 2008 (UTC)

The text on the sign up page ( ? It says "By creating an account, you agree that all work uploaded to and all data created by use of any tools which connect to is to be licensed under this Creative Commons license (by-sa)."
This text needs to be reasonably concise obviously, but I agree that it could be phrased better. -- Harry Wood 16:42, 6 July 2008 (UTC)

How should a company "attribute the work in the manner specified by the author"

Our company is thinking about using OpenStreetMap in our reports to clients. Typically, we would use a screenshot of only a section of a displayed map (not the whole web page) and place it in our report (e.g., to show the location of a site we investigated). Our policy is to comply with legal requirements. The OpenStreetMap license states:

"You must attribute the work in the manner specified by the author or licensor"

Isn't OpenStreetMap created by numerous "authors"? How can we identify who they are (for a given map section we are interested in)? How do we know which manner of attribution they specified? Is attribution needed at all in our case?

This is a key issue for corporate users.

Yes, this is one of the reasons why we're actively considering moving to a different licence. (P.S. Please sign your posts. :) ) --Richard 11:38, 17 September 2008 (UTC)

Wouldn't it be smart to at least "upgrade" to the most recent version of cc-by-sa (3.0) instead of still using 2.0?

My company, a retail store for electronic devices, specifically for Garmin GPS-Devices, is thinking about selling readymade tranflash-cards containing the (weekly updated) most recent version of OSM for Europe, as the professional maps that garmin provides are too expensive for many customers. To be sure to comply with all legal requirements, we want to have a lawyer check the license thoroughly. Too bad the cc-by-sa-2.0 is obsolete and replace4d by versio 3.0 and there seems to be no way to find the legal code of the old version anywhere on Wouldn't it be smart to use the newer version as long as no better license is found for the OSM-Data? Zimbosmurf, September 16th, 2008

As we're thinking of moving to a new licence anyway - and we have found a better licence, the Open Database Licence - the "upgrade" to 3.0 would be significant hassle with no clear benefit. There is a broad consensus in OSM that the CC-BY-SA family of licences is no longer suitable for our needs. In addition, significant doubts have been raised about the
But if you want to distribute OSM data under 3.0, you can do - the licence says "You may distribute, publicly display, publicly perform, or publicly digitally perform a Derivative Work only under the terms of this License, a later version of this License with the same License Elements as this License". A Garmin version would probably count as a Derivative Work, so this gives you express permission to distribute under "a later version" such as 3.0.
The legal code for 2.0 is the very first Google result for CC-BY-SA so I wouldn't say it's that hard to find! --Richard 11:38, 17 September 2008 (UTC)
Sorry, but "my" google just found the legal code of cc-by-sa 3.0 - and still does. The only thing I can find concerning 2.0 is the "you may"- and "you may not"-blahblah. If you say it's OK to distribute under 3.0 everything sems fine to me.
I'll have this checked by our lawyer. zimbosmurf, September 16th, 2008


Is it legal to render not-OpenStreetMap data (without copyright) with Osmarender? --Willem1 21:11, 15 December 2008 (UTC)

Yes. I don't see why not. Data and tools-to-work-with-data are two different things.
Just as you can use an open source web browser to view a copyrighted website, you can use open source tools from openstreetmap to work with any other data. It's legal, but it's not what the tools are designed for of course. -- Harry Wood 13:30, 30 May 2009 (UTC)

Using a map in a scientific work published by a commercial scientific publisher

I intend to publish a scientific work, where adding a map would help the reader understand the matter. As far as I understand the licence stuff and its interpretation, there is no problem with uploading the paper to the arxiv (see [1]).

But publishing at the arxiv is not sufficient for scientific work to be acclaimed. For this the work has to undergo the peer-reviewing process, which still is almost exclusively managed by the major scientific publishing companies like Springer or Elsevier. But when I publish the paper not only at the arxiv, but as well with one of the major publishers the problem starts at the point, where they claim that all content is subject to their copyright. The publisher selling the journal, book, or online access for money is no problem, but it is, when they forbid someone, who has bought it to redistribute it.

Another problem is that they usually want to have an exclusive licence. This - as I understand it - makes it almost impossible to use an OpenStreetMap map. Exclusive licences in science are nonsense, which can be seen by the fact that no publisher risks to sue the arxiv. No publisher I've asked so far on the legal status of publishing as well at the arxiv has ever answered my questions. They know that they do not have the power to enforce their own rules, but still they do not want to allow it explicitly. But while the rule is overwhelmed meanwhile by reality, I like to do things correctly and not rely on "everyone's doing it like this for years now".


  • Is it possible to use one single map image in a scientific paper with a major publisher, who claims the whole copyright for himself? OpenStreetMap woul of course be credited in the References/Literature/Bibliography section of the paper.
  • Is it necessary to state explicitly that the map image is under a different licence than the rest of the article?
  • Is it necessary that the publisher admits that the whole article is under CC-by-sa, thus under a different licence than the rest of the publisher's publications? (I do not hope to find such a publisher.)
  • If the work is part of a conference proceedings book, would it then be necessary to licence the whole book under CC-by-sa?
  • If the work can be downloaded from a homepage (my own or the publisher's), is it then necessary to licence all content under that domain with CC-by-sa? Even if the map is part of a pdf, separate from different html or pdf content?

Unmapped 08:58, 10 April 2009 (UTC)

The wiki is not the best Contact channel for asking this kind of question, hence it has gone unanswered for a year now. You also have some questions which are difficult to answer. These are issues around the definition of "collective work" and "derived work". Some Common licence interpretations try to tackle these thorny issues. You might find those useful.
But if the publisher really claims whole and exclusive copyright, then that doesn't fit with using OpenStreetMap. Actually all it means is, you wont be able to "do things correctly". I reckon you could just give us credit with a nice paragraph explaining the source of your map images, you could probably do this without mentioning copyright. That would be bending the rules of both OpenStreetMap and the publishers... and you know what?... I reckon you'd get away with it ;-) ... [this is not legal advice]
-- Harry Wood 09:28, 6 April 2010 (UTC)

Fair Use

There are certainly ways in which proprietary data could be used without compromising the result. For instance, if we compare maps from independent sources, together with satellite imagery and our own observations on the field, the legal status of the result should be the same as an essay that quotes from references. I don't think we need to be so strict such as to ban completely the use of proprietary geospatial data, but rather, have a well defined fair-use policy. We ought to have as much liberty to use proprietary data as a wikipedian has to use a book as a reference for article.--Josefo14 13:06, 30 May 2009 (UTC)

About box/page for attribution

As I noticed while translating it to German, the FAQ currently suggest that attribution on "browsable electronic maps" usually is - and should be - displayed in a corner of the map. However, an "about" box or page is pretty common, too, particularly in mobile apps (none of the OSM apps I know of puts full attribution directly on top of the map). So I suggest to mention this as an additional possibility. --Tordanik 15:44, 26 September 2012 (BST)

I've mentioned that option now. --Tordanik 12:17, 15 October 2012 (BST)
This contradicts what we request at so I've removed it. If the community consensus is that an About box is enough, that's fine, but we need to discuss that somewhere less obscure than a wiki talk page - the legal-talk list would be the obvious place. At the least, we would need to consider if an About box is ok in all circumstances or solely where screen size is restricted.
As a general rule, we take ODbL's requirement for "reasonably calculated" attribution, and CC's requirement for "credit reasonable to the medium or means You are utilizing", to mean that our attribution requirement should be broadly equivalent to that of the majority webmapping solution. That is of course Google Maps, and we should therefore expect attribution on a comparable level.
I've forwarded this matter to LWG for their consideration. --Richard (talk) 09:01, 16 July 2013 (UTC)
I did think there was consensus that crediting on an 'about' page of an app is acceptable, but I guess there's a difference between what's acceptable, and what forms part of the advice given here.
I'd like it if there was more advice and recommendations. This can't be based solely on the legal text, since there's not enough detail there, but instead looking at industry standard practice and spelling out a much more detailed interpretation within the legal text, covering nuances around different screen sizes etc, and (maybe this is the tricky bit) agreed by community consensus. It should also be describing, not just what's "acceptable" or "required", but how we would like to see people credit us, balancing the desire for giving good credit with the understanding that users don't want to splash overly obtrusive credits on their maps (and we don't want to require or recommend that).
I thought the RFC Attribution Mark discussion took things in this direction, except that it also departs quite radically from advice we currently give.
-- Harry Wood (talk) 09:50, 16 July 2013 (UTC)

Clean up

I guess it would be useful only to list questions which are not (yet) answered on This is affecting the whole section "3. Using" --Jojo4u (talk) 12:12, 15 August 2015 (UTC)

Getting permission

Can someone please clarify seek written permission to licence their data under our licence and contributor terms.

Particularly the word written? Seems in a search, and suggest that, at least in the US email is probably fine.

Considering getting involved towards some data ingests to improve quality (addresses to improve routing, perhaps contacting companies to add their locations so we start to get some types of businesses at least filled out a bit, such as food). Just wanted to know in advance whether I should specifically request written permission or whether email permission, if well explained, will suffice. Probably will move over to the talk message boards to pursue advice on whether messages are complete enough, and on doing the importing, but this would be a useful thing to clarify in this wiki for the community.

Thanks much! To God be all glory, Shane --JeopardyTempest (talk) 05:57, 30 March 2016 (UTC)

Just lumping providers together

Looking at one cannot tell if a given map on Facebook is 0, 1, 99, or 100% OSM data. Jidanni (talk) 02:51, 10 April 2020 (UTC)

Military / Government secrets

Let's say OSM gets a take down orders about some military secret etc. classified information that some mapper has mapped. Should OSM obey them if they are coming from the government of e.g.,:

  • Some dictatorship, saying "erase all content of my country"?
  • Some country where most of OSM's internet connections pass through, for fear of cut off?
  • The country where OSM's headquarters is registered?
  • Any country that we deem the request to be reasonable.
  • etc.?

Jidanni (talk) 09:04, 5 July 2020 (UTC)

OSM is an international project with a single database. If we violate the laws of a nation, that nation might ultimately block access to OSM. So in general we should respect the laws of each nation, but only with regards to data inside that nation's borders. We could decide to ignore the laws of certain nations altogether and risk being blocked until their laws improve to an acceptable level. Given the choice between removing a whole nation from OSM and that nation blocking us, we should choose the block. What we shouldn't do is map wrong data. It's acceptable to map a military zone as landuse=military and otherwise leave it empty. But it's not okay to instead map it as farmland. --Wulf4096 (talk) 10:24, 5 July 2020 (UTC)
In China I am using landuse=military but was requested once by a single user to not tag it like that. One could additionally tag it with something like china_class=military to let military areas still be findable if landuse=military gets removed one day. On Chinese maps (Baidu, Sogou, Amap, Tianditu) I've never ever discovered any military area being outlined... --katpatuka (talk) 10:43, 5 July 2020 (UTC)