Talk:Open Data License/Why You Should Decline

From OpenStreetMap Wiki
Jump to: navigation, search

Dispute tag

The editors of the 'No' page are not lawyers, and through their own "university of life" opinions are spreading fear, uncertainty and doubt about the license change. Much of it is simply wrong. Steve 21:18, 5 December 2009 (UTC)

I do not believe that it is appropriate to flag the whole page with 'disputed'. I suggest that the LWG has the option to add a rebuttal to each comment - people can then make their own mind up about the actual facts. I suggest that we limit each section to one statement and one rebuttal only. I propose that we remove the 'disputed' tag in 24 hours. PeterIto 22:17, 5 December 2009 (UTC)
I propose that the reply (or rebuttal) should be on the main page - I just don't think people will tend to juump backwards and forwards between the article and the talk page. Any objections? PeterIto 22:17, 5 December 2009 (UTC)
+1 I strongly agree. For both Yes and No pages. --firefishy 22:38, 5 December 2009 (UTC)
The dispute flag has been removed, however there is still an 'edit war' going with people adding a removing rebuttals on the page. Personally I still vote for a single 'official' rebuttal from the Licensing Working Group being allowed on the No page. I say this because I think it is unreasonable for the LWG not to have the right of reply after all the work that they have done. If anyone has a problem with the rebuttal then that can be discussed on this talk page or by email. Any comments? PeterIto 05:24, 6 December 2009 (UTC)
I have reinstated an official 'reply' from Matt (which I assume is an official reply from the Licensing Working Group) to one of the statements. I have not reinstated Tom's replies because I don't believe that he is a member of the working group. If people disagree with this appproach then lets discuss it here or on the foundation email list - personaly I would prefer to discuss it here to avoid overloading the email list. PeterIto 05:47, 6 December 2009 (UTC)
And I've removed it. The LWG has prepared their case and presented it in the license change proposal. This page comprises personal statements by those that make a case against. As it states at the top of the page, and in accordance with normal wiki etiquette, discussion should be on this talk page. 80n 10:00, 6 December 2009 (UTC)
80n, you seem to be the only person who is resisting 'replies' on the main page and removing comments by anyone who disagrees with you. It is perfectly normal for wikipedia pages to contain rebuttals of this sort - indeed it is a expected in a balanced article. Note that I am only suggesting that there should only be a single short 'official' response from the LWG and not a general 'open season'. The talk page on a wiki is normally for a 'meta conversation' about the subject, not part of the main subject. Possibly we should change the intro of the article to match the discussion we are having on the talk page that the reply should be on the main page - or possibly reinstate the 'disputed' flag which would be an admission of failure. PeterIto 10:44, 6 December 2009 (UTC)
I'm opposed to anyone editing the sections that are signed by me, yes. Others can police their own sections, I can't speak for them. The Yes and No pages are not discussion pages they were placed there for personal statements. If there are factual inaccuracies then let's hear them and get the personal statements corrected. If it's just argumental comments they belong here on the talk page. 80n 10:58, 6 December 2009 (UTC)
In the interests of moving forward I have moved all comments to the talk page and provided a clear link from the main page to the relevant section of the talk page. I have added a 'disputed' note to the link if there is strong evidence that the claim is not factually accurate. PeterIto 11:32, 6 December 2009 (UTC)
Thank you Peter, this works for me. 80n 11:36, 6 December 2009 (UTC)

Signing claims?

Should these claims that are being made (both for and against the license) be signed by the person making the claim - or indeed by more people if they agree with the statement? PeterIto 21:48, 5 December 2009 (UTC)

+1 --firefishy 21:55, 5 December 2009 (UTC)
We can reverse engineer signatures onto the sections from a quick look at the history tab - I suggest we do that unless there is disagreement, in which case we should discuss it further. 22:17, 5 December 2009 (UTC)
Done --firefishy 22:33, 5 December 2009 (UTC)
Thanks. However I will remove the 'un-signed' phrase from the text to avoid any suggestion that comments to date have deliberately not been signed. PeterIto 05:24, 6 December 2009 (UTC)

Comments on individual claims

Cannot import CC BY-SA licensed data

REPLY: Part of the process to come after the OSMF Membership accepting the ODbL proposal is for the OSM:Licensing Working Group to assist in getting the Community Imports cleared under ODbL. The ODbL retains the key principals of CC-by-SA (Sharealike and Attribution) Disclosure: I am a Member of OSM:Licensing Working Group --firefishy 13:57, 5 December 2009 (UTC)

I have added the 'disputed' tag to this claim purely because of the phrase 'will' - It seems to me that we have a good chance of getting Creative Commons to support the release of data on the appropriate CC0 license in Australia and elsewhere, and as such I think it is likely that we will be able to retain many of the data sources. I am not an expert in this, hence setting up a wiki page where it can be discussed further. PeterIto 11:37, 6 December 2009 (UTC)
I have changed this to 'not disputed' because of the qualifier "without the copyright holders agreeing to those terms". One could believe that the chances are good to get agreement, but unless you get 100% agreement, then not all the data will be able to be imported / migrated / kept. Eric2 12:17, 30 May 2011 (BST)
CC-BY-SA does not allow sublicensing which makes using the data limited to the allowed license, therefore discussion with the source provider is required to make the data available under a less restrictive license. ODbL maintains the Attribution and Sharealike principals. -- Firefishy 01:37, 31 May 2011 (BST)

OSM's Contributor Terms are not compatible with ODbL

I am interested in exploring this one. Clearly the resulting DB is published as ODbL and can't (or I hope won't) contain additional restrictions about what can be done with the date. One version of ODbL allowed notices about further resrictions to be added. PeterIto 11:07, 6 December 2009 (UTC)

The Contributor Terms require that the contributor grants the OSMF some rights. If I find a random ODbL licensed dataset somewhere I can only share that under the same terms that I recieved it, namely ODbL. I cannot add it to OSM because that requires me to grant OSMF rights that I don't have myself. The outcome is that nobody (except the original licensee) can contribute ODbL data to OSM.
Let's take a concrete example. Suppose that ITO, for example, takes a copy of planet.osm, adds some NAPTAN data to it and publishes planet+naptan.osm under the ODbL. If ITO chooses not to agree the the Contributor Terms and chooses not contribute it back to OSM then it cannot be incorporated back into OSM. Nobody else can do it because they can't grant OSMF rights that they do not have. This totally breaks the principle of share alike that is so important to us all.
It's worrying to me that this scenario has only become apparent in the last couple of days, despite considerable scrutiny of the license proposal by quite a few people. I'm worried that there may be other equally undiscovered flaws in the license proposal, particularly in the Contributor Terms.
So far I have not seen any explanation or comment at all from the LWG about this scenario, so I think it is new and unexpected to them as well. 80n 11:34, 6 December 2009 (UTC)
+1 --DiverCTH 06:46, 21 December 2009 (UTC)
+1 Thanks, that is indeed a worrying and probably a common scenario. It is not worse than what e.g. the FSF does (they also require assinging of copyright, therefore making it impossible to import and GNU GPL code by others than the copyright holders), but it should at least be discussed and acknowledged. --spaetz 13:16, 23 August 2010 (BST)
For me, this was the most troubling part of the proposal. But I now realise this is not an issue with the ODbL itself, which remains in my opinion an excellent license for anyone requiring attribution and/or share-alike. So I'm happy for my submissions and the rest of OSM to be ODbL-licensed (though I'd be okay with public domain too). And I have no problem with data sources offering up their data under the ODbL if they feel public domain is too permissive.
The only issue here is with OSM's proposed Contributor Terms. Under them, contributors seem only to be able to import data which is either their own, or essentially public domain. If that's too strict, the terms can always be changed at a later time, and because it doesn't affect already-contributed data it's much simpler than a change of the OSM license.
Anyway, the terms may be quite appropriate for most individual contributors. Any imports of data not in the public domain could perhaps be handled or merely approved by the OSM:Foundation/Import_Support_Working_Group, to permit certain data into OSM without the licensor having to accept the Contributor Terms at all. As long as OSM only distributed its database under the ODbL, I believe any ODbL-licensed databases could be imported in this way without special permissions from licensors. (Could this also be true for CC-BY?)
If the OSM database license must change again in the future, the working group would have the task of reviewing the licenses of, or renegotiating with licensors for, any such data sources, but it should have records of who/what they are. That would avoid (potentially inactive, or unwilling) individual contributors from having that burden, which is a major part of the current predicament. --StevenChamberlain 06:30, 25 December 2009 (UTC)

OSM's Contributor Terms are not compatible with ODbL. This is no longer true since Contributor Terms revision 1.2.4. -- Firefishy 16:33, 1 June 2011 (BST)

The OSM Foundation will have unlimited rights to all your data

The following rebuttal is proposed for the claim that 'The OSM Foundation will own all your data'.

REPLY: Which is why, if you read the contributor terms, it's not only a vote of the OSMF membership, but also "active contributors". This massively raises the amount of effort necessary for any organisation to attempt a take-over of OSM's data. --Matt 19:29, 5 December 2009 (UTC)

It massively raises the number of edit bots that are required. That's all. 80n 11:04, 6 December 2009 (UTC)
I feel that this particular comment is badly phrases - the OSMF will not 'own' your data - it is extremely clear that the contributor will still 'own' their date and are only licensing it. It also claims that it would be 'relatively easy for a large organisation to take over OSMF' - again it would be impossible for a large company to create enough users to change the license as proposed by the Contributor terms. I believe that a rebuttal on the main page is appropriate. See dispute section above for further discussion of the general point of inclusion of a rebuttal on the No page. PeterIto 05:24, 6 December 2009 (UTC)
What rights to the data will the OSMF not have? If you have full rights to something then you own it. 80n 10:03, 6 December 2009 (UTC)
The claim clearly states that the OSMF will 'own' the data and this is just not true because 'ownership' is different from 'rights' and the contributor has only given on-exclusive rights to the OSMF on certain conditions and has retained rights of ownership to do anything else they choose to do with the data as they see fit. PeterIto 10:48, 6 December 2009 (UTC)
The wording of the claim has now been adjusted by the original author to change the word'own' to 'have unlimited rights'. PeterIto 11:26, 6 December 2009 (UTC)
The OSMF will not have the rights to put the data under a closed license, as the contributor terms restrict the relicensing powers to an open and free license. That is a strong enough restriction to say OSMF owns the data or has unlimited rights is incorrect. amm 20:35 8 December 2009 (UTC)

I am having doubts over assigning unlimited rights over contributions to any group of people. This too much like "tyranny by majority" and certainly not the practice with software and CC licenses AFAIK. --TimSC 21:30, 6 December 2009 (UTC)

+1 --DiverCTH 07:21, 21 December 2009 (UTC)
I also think the unlimited rights assignment is unnecessary. If data is contributed under the understanding that the licence is ODbL, then only the rights sufficient for ODbL should be passed on. If implementing the ODbL requires the licensor to have more rights, then this is probably a flaw in the ODbL and should be fixed.
Addressing the next clause: I think new data should be submitted only for distribution under the terms of the ODbL (plus data licence to account for some places allowing rights to the data). There is sufficient provision in the ODbL for upgrade to a future version of the licence. The ability to enact a licence change to something as broad and undefined as a “free and open” licence without the original contributor’s say in the matter just because they haven’t contributed and/or are not contactable for a period of time (and exceedingly short time periods at that) is, very much like the tyranny described above.
Sward 00:35, 28 December 2009 (UTC)

The current license is not broken

Personally I believe it is irrefutably proved that it is 'broken' even by Creative Commons themselves. PeterIto 11:07, 6 December 2009 (UTC)

Maybe not broken, but certainly I'd argue the current license is unsuitable. (In fact I think I argued it at the first OSM mapping party on the Isle of Wight in 2006). In 2007 I spent hundreds of hours tracing Yahoo imagery of Baghdad. I have no interest in Baghdad except that there was no other online maps with good coverage, and I thought it would be good publicity for OSM to have the best maps available. A short while later one of the major UK TV networks was interetsed in using OSM maps, but wanted to do a mashup with some Google data. Due to our licence conditions they dropped the idea of using our maps.[1]Dmgroom 00:03, 8 December 2009 (UTC)
David, you subsequently offered to put the data into the public domain [2]. They still didn't use it, so it's likely that it wasn't the license that was the problem in this case. 80n 20:26, 12 December 2009 (UTC)
True, but I believe by that time the TV company had either lost the will to use OSM maps, or simply the deadline on the relevant news item thay wanted to use it on had passed. Dmgroom 13:34, 22 December 2009 (UTC)
I understand that you're personally involved with this Baghdad discussion, but what about success stories like Haiti? That shows that the licensing in its current form works wonderfully for providing useful services based on volunteer efforts. Just because one TV station didn't produce their mashup (for whatever reasons, we don't know) doesn't mean the CC-by-SA license is broken. Eric2 22:26, 8 June 2011 (BST)

ODbL is unproven

Yes it is, unfortunately there is not a proven solution available and we know that the current license is inappropriate. As such it is the least-worst-license to choose. PeterIto 11:07, 6 December 2009 (UTC)

Is OSM(F) going to be the first major user of ODbL? (I understand it has been quite big in the discussions and changes). Who else is using ODbL and who might start using it after OSM? Someone has to step forward first. - LastGrape/Gregory 00:58, 7 December 2009 (UTC)

ODbL licensed data can be reverse engineered

See original post. Jordan (ODbL Author and Lawyer) + OSM Legal council say that it is protected. Non-lawyer claims that it isn't protected. Therefore question is disputed. -- Firefishy 17:45, 28 August 2010 (BST)

Who owns OSM? You!

I agree with the first two (translated) paragraphs. I commented under OSM:#The OSM Foundation will have unlimited rights to all your data: The ability to enact a licence change to something as broad and undefined as a “free and open” licence without the original contributor’s say in the matter just because they haven’t contributed and/or are not contactable for a period of time (and exceedingly short time periods at that) is, very much like the tyranny described above. —Sward 00:49, 28 December 2009 (UTC)

Add your comments here. PeterIto 11:07, 6 December 2009 (UTC)

Another issue is how to classify a contributor who is still interested in the project, but no longer making large contributions because their area of interest is "complete". Here are some of the ambiguities that I feel need to be addressed:
  1. What constitutes an "active contributor"? Is touching a single node every N days sufficient? What about simply logging into Potlatch?
  2. Will new users be made aware of this controversy? I was accused of WIKI-vandalism when I attempted to add a link from the "Who Owns OSM" portion of the FAQ to this discussion.
  3. What (if any) mechanism will the OSMF use to contact original contributors if they are "inactive".
  4. What (if any) mechanism is in place for determining which bulk imported data sources have to be removed because the new contributor terms & license are incompatible with the original copyright-holder's license?
  5. What (if any) mechanism is in place to ensure that as many contributors are polled as to whether they agree to the license change or not?
  6. How can/will a fair election be conducted when not all contributors are OSMF members?
  7. What (if any) outreach will be made to non-english-speaking contributors? How much effort will go into ensuring their concerns are addressed?
Finally, I feel that the OSMF leadership, and the LWG in particular have gotten into an one-size-fits all mindset when it comes to communication. The decision to push ahead with the transition without polling the existing contributor-base is arrogant at best and borderline-criminal at worse (see point 4 above). Some of the comments to my resignation announcement are reminiscent of the Beware the Leopard sketch in The Hitchhiker's Guide to the Galaxy.

--DiverCTH 04:13, 8 July 2010 (UTC)

Contributor Terms can be changed

Incomplete History for Split Ways

It is possible to reconstruct the original contributor in almost all cases, because when a way is split, it still reuses the old nodes. Thus one has to look at all ways a certain node has been part of during its existence.

A larger problem exists when a user deletes a way and reuploads a new one at the original position. AFAIK there is at least one user who seems to do this frequently for his username to be at the top of the history. --Schuetzm 16:05, 7 December 2009 (UTC)

Good luck with that. 80n 00:17, 8 December 2009 (UTC)
There is no such a problem, we can easily reset every edit of an user, if we also want to remove derivatives we can just reset the ways, nodes and relations to the version number before the given user did edits. --Bürste 12:00, 8 December 2009 (UTC)
We could, in theory, recreate everything from the changesets. There are *only* 3 million or so. It would take a while, but might be doable.
There are other potential issues however; what if user A creates a node, X; some months later user B deletes all of the tags, and moves node X from somewhere else to re-use it as part of a completely separate way, along with several other nodes. Some months later user C adds node X to another way - which intersects with the way B created. User B doesn't choose to re-licence, users A & C do. The node is rolled-back to v1 - which is some miles away - and e.g. tagged as a shop. The way that user B created is removed, but the way that user C created presumably still contains node X. That isn't ideal, nor is removing nodes from the way user C created (could end up with one or zero node ways). Richard B 00:25, 10 December 2009 (UTC)
Recreating the data from changesets will not work. You cannot tell that a way has been split. The changeset will just show that a way previously owned by user x has been edited by user y and now has a few less nodes. At the same time a new way will also appear made up of *some* of those missing nodes and added by user y. Any watertight algorithm would have to be so conservative that you'd probably not be left with much data at the end of it. 80n 20:35, 12 December 2009 (UTC)

Impact on Transitioning to ODbL If Significant Minority "No" Vote

OSM is going to lose data for sure, for a pretty much unclear benefit from a new license. That is the only thing guaranteed by the move. And in addition the whole transition makes people move away from OSM and especially from any OSMF-related project. Trust is important, and now OSMF is gambling with its credibility :-( --amai 21:24, 27 December 2009 (UTC)

I thought it was interesting that at the SotM 2010 session: License Working Group, the only issues which anybody seemed to be raising were these speculative "what if" scenarios about losing data. When given an opportunity to question the people on stage (license working group) nobody actually had anything to say against the license itself. At this stage we don't know how much data would be lost. We have to ask contributors. It's an early step in the process. Later on we can take a look and see how much data is owned by people who are saying "no". The next step then is not set in stone. Maybe the next step will be to drop the whole idea and go back to the drawing board for another couple of years (By the way, that's how long people have spent in meetings with lawyers discussing this, and figuring out the right license for OpenStreetMap) I hope we can have an informed debate, based on people's real opinions of the license. Since ODbL is a positive step for OpenStreetMap, people should say "yes". Some people are going to say "no" for various reasons. We have to face that sad fact. I hope people don't say "no" because they're worried about how many people are going to say "no". That way we'll never really get to assess the situation, and make an informed decision. -- Harry Wood 15:29, 23 August 2010 (BST)

Saying no is anyway the safe bet. Data does not get lost and all options remain valid. And there is dataloss upcoming, but I guess nominating my dozens of ways is the smallest part only... --amai 18:11, 23 August 2010 (BST)
The thing is, the whole "voting no" issue seems to have been "solved" by simply not offering the "No" choice to the user (at least not at this stage, I'm not sure if it is maybe going to be presented to the users at some point in the future). And I agree with Harrys point here (although from somewhat different angle this time) - not offering it will make sure "we'll never really get to assess the situation, and make an informed decision". --mnalis 18:26, 23 August 2010 (BST)
Harry Wood says "I hope people don't say "no" because they're worried about how many people are going to say "no"." I'm more concerned about the people who feel forced to say "yes" because they're worried that otherwise their data will be deleted. And about the people who would say "no" but aren't being given that option because they didn't register before the cut-off date. OSM relies on the volunteer community, without that it's just another database. Deleting the contributions of keen and well-intentioned volunteers is insane. Eric2 22:21, 8 June 2011 (BST)

Fundamental flaw in voting: How to treat non voters

Firstly, there is no voting for general public (OSM contributors). The vote this page talks about was for OSMF members only, and it ended last year with votes in favor of ODbL. What is currently happening is asking new contributors to accept the new licence. Very soon now, existing contributors would be asked to relicense their data too. Those who do not (or are not able to), will have their data removed from current OSM map. While I do not think that those who left or don't care (anymore) are majority as you believe, removing their contributions (and of all who edited same nodes/ways after them) would still probably be too great damage (IMHO). But that call for relicensing is not a vote. However, if the amount of data to be lost is decided to be too big (and what constitutes "too big" is still undefined OSM:http://lists.openstreetmap.org/pipermail/osmf-talk/2009-December/000610.html despite efforts to establish that limit beforehand), the move to ODbL will be aborted OSM:http://lists.openstreetmap.org/pipermail/osmf-talk/2009-December/000615.html or so am I led to believe --mnalis 11:34, 18 July 2010 (UTC)

Fundamental imbalance in voting: New contributors, forced to accept ODbL will eventually outnumber CC-BY-SA contributors

Majorities don't matter for the current license change, as data can only be re-licensed by the contributor who created it. No one, not even a 99% majority, has the right to make this decision in your place. --Tordanik 09:44, 18 July 2010 (UTC)

This argument is simply a lie. New contributors don't have a vote in the current license changeover. --Cartinus 11:29, 18 July 2010 (UTC)

Cartinus, you seem to be misinformed. There is no voting for OSM contributors. The one that was (and ended last year) was for OSMF members only (see here).
Also you seem to misunderstand term "re-licensed" which Tordanik is talking about. You have to have previous contributions under old licence to be able to make re-licensing choice, and new users by definition don't have data licensed under CC-BY-SA. And old users (when time comes soon) will be able to make a choice not to relicense their changes under ODbL. In that case, their contributions will be removed from ODbL dataset, but anybody could still continue to use it under CC-BY-SA (just probably not on infrastructure OSMF is paying for).
While I'm against move to ODbL and new contributor terms (you can see my user page for details), I do not agree with all DiverCTH says here. Firstly, I do not agree that OSMF is trying to hide the change. I mean, OSM:http://www.osmfoundation.org/wiki/License/We_Are_Changing_The_License Find out more about OpenStreetMap's upcoming license change link is on top of EVERY wiki page, it can hardly be missed! Secondly, there is no voting (as mentioned above). Thirdly, there is limit to what (and how) OSMF might change licence (although it really should be improved greatly). The FUD will not help convince users that direction OSMF is taking is wrong. Please don't do that. The only chance we have of convincing users to stop it is by being sincere and pointing out real problems. --mnalis 12:09, 18 July 2010 (UTC)
No I am very well informed, it is just you drawing the wrong conclusions. There _will be_ a question asked about the relicensing of the current data. (You might not want to call that a vote, but it is.) This question will be posed to the people who participated _before_ the _dual_ licensing. So whether the people who have dual licensed their data outnumber the people who joined before is completely irrelevant. Therefor the "Fundamental imbalance" is complete FUD. (And I make this argument even though I don't like the ODbL at all.) --Cartinus 19:27, 19 July 2010 (UTC) P.S. Maybe you should watch the indents: I didn't reply to Tordanik.
Oh, thanks: it really was non-existent indenting I saw :-); I though you were replying to Tordanik, so my post mostly makes sense in that context. But as for your new comment; if you DO decide to call accepting a new contributor terms a "vote" (which I wouldn't, as the point of voting is IMHO that it is binding even for the minority who didn't accept it, which new CT don't do), then your original comment "new contributors don't have a vote" is incorrect (as they currently have a choice either of "voting for ODbL" or "not becoming new contributors", so they always "vote for ODbL" - which is quite different from "they don't have a vote").
So I do agree with your original comment that new OSM contributors don't have a vote, but that is because I do not accept your second comment that accepting or declining new contributor terms is "voting". I mean, yeah, ratio of accepts/declines of new Contributor Terms might influence the final ODbL move decision (or it might not, depending on OSMF Board sanity :-), but so can many other things that are not "voting" (like threating with vandalism, making forks, buying out CloudMade or just bribing LWG members etc) --mnalis 13:33, 20 July 2010 (UTC)

All Share-alike Licences Create Ambiguity

Making Copies of your maps may not be allowed

Thanks for pointing this out. I actually see it as a positive point. If I make render some nice red and pink map images/tiles of Europe they are protected under normal copyright law. Say amazingly my colour scheme (maybe including icons I drew) was considered creative and original, then I own the copyright. You can't render red and pink America map imagess if they're using a very similar style to me, it would be upsetting to all the time I spent dreaming and planning the style.

Interjection: AIUI this is factually incorrect. A copyright protects the expression of an idea, not an idea itself. (That would be a patent, or perhaps trademark or trade dress.) If you have a red and pink colour scheme, I can also have an identical red and pink colour scheme without any copyright problems as long as I did not illegally copy your stylesheet file. That means the rest of this argument does not follow. Gerv 03:53, 7 December 2009 (UTC)

You could however create a red and orange style (providing it wasn't trying to copy my style/shades and was original) and create map images of America and Europe. We both have pretty maps. You know what, I might even be a nice guy and release my images/style under CC-BY-SA(or PD), maybe even provide the stylesheet file. I way to even force style sharing would be if I used online-tool-x to create a style, but in the terms of that software I have to submit exclusive rights to either the tool maker or Public Domain so that you (or anybody) can use online-tool-x to duplicate and modify the style. --LastGrape/Gregory 01:12, 7 December 2009 (UTC)

I consider my self a "nice guy" to publish my maps under a permissive license. I expect you to do the same if you use my maps.
Since we are not in the main discussion I would like to state that I believe, and this is after talking with lawyers, that OSM edits might as well be creative output, we free form a lot to create our maps and that is very much creative. But no one is stupid enough to believe that they can just copy the OSM database so that argument will never have to be used. Erik Johansson 01:54, 7 December 2009 (UTC)

Principle of least damage

I'd disagree as to the worst case with CC-BY-SA. You suggest that it'd be "all of OSM data is effectively public domain". I think the worst case is that OSM data remains in an unclear legal state.

CC-BY-SA is more enforceable in some countries in others. So the situation will never be as unambiguous as "effectively public domain". In the US, yes, lots of it will be PD. In the UK, maybe not. So no-one will use it as PD, because an over-eager OSMer in another jurisdiction could still sue. Meanwhile, the unclear and unspecific definitions of BY and SA, when applied to data, will continue to stop people using OSM data.

CC-BY-SA is hurting OSM already, and has been doing so for many years. A major online map site (one of the two biggest at the time) wanted to include OSM as a "community layer". Their lawyers looked into CC-BY-SA and concluded it was too dangerously ambiguous for them to use it. Another time, a UK TV news programme wanted to use OSM's Baghdad map and were prepared to provide attribution. Again, CC-BY-SA was simply too unclear on what was "derived", and what attribution was required. And so on.

--Richard 00:23, 8 December 2009 (UTC)

Thanks Richard for your opinion. I do not see that "unclear legal state" can be worse than any of the extremes -- it will always be either one of the extremes, or somewhere in between. So you either have people/companies treating it as a most restrictive form of CC-BY-SA imaginable (which makes it as good as ODbL), or as Public Domain (which is a "worst case" if you prefer copyleft license), or as a something in between ("mild" version of CC-BY-SA or whatever). I do not see any way that it can be considered more restrictive than "most restrictive form of CC-BY-SA imaginable". Schrödinger's cat is either dead or alive, it doesn't turn into T.Rex just because you don't know in which state it actually is.
As for your other point, I agree with you that some people/companies will have problems with any license. I do not think that is a fault of license. I could also see the possibilities that some companies would be put off by ODbL complexity or the contract part if it etc. The companies you mention could have assumed the most restrictive (for them) of all possible CC-BY-SA interpretations, and know that they will be 100% safe, no matter which one of possible interpretations is taken. The fact that they choose not to go this fail-safe way, indicates that they either did not want to accept the most restrictive variant (in which case they would reject any other equally restrictive license, even it was 100% clear - such as ODbL), or that they weren't interested enough to put not-too-big effort for finding most restrictive variant of CC-BY-SA (in which case I think ODbL would fare even worse; even if it is much clearer, but also much more complex, involving not one but combination of up to 3 different laws interacting) --mnalis 19:49, 8 December 2009 (UTC)
I think the problem is that the "most restrictive" interpretation of CC-BY-SA makes OSM potentially unusable. For example the most restrictive interpretation of the Attribution clause would be to Attribute all 150.000 users each time you use OSM, which isn't really practicable or inteded by anyone. Also, the idea that any one of the 150.000 contributors could potentially sue you even if their claim is unreasonable isn't directly appealing for using it either. Furthermore, the most restrictive interpretation of the share-alike would make the entire TV program share-alike in which a map is embedded, again not directly practical or doable. Therefore the most restrictive interpretation could simply lead to noone using OSM, which can't be in anyones interest. The ODbL clears much of this up and makes it therefore much easier to use. If you just "use" the OSM, there is no uncertainty with ODbL, It states a clear Attribution clause and produced work can be used in anyway. Share alike won't be a problem in this case either, as you haven't created a derivative DB, but just used OSM data verbatim, so if anything you would have to publish an empty diff. This is in my legaly unskilled opinion a huge improvement in clarity and could make a real difference in OSM adoption. --amm 20:59, 8 December 2009 (UTC)
Thanks amm, that have-to-mention-150k-contributors would really be messy if CC-BY-SA can really be interpreted that way (which I will assume it can, although I'd appreciate a pointer to data on that subject if you have it). I've updated that on the main page. --mnalis 18:33, 9 December 2009 (UTC)
Also, I think if the idea is to get MORE companies to use OSM (and not be scared away by license), than one should move to LESS restrictive license (like CC0), instead of trying to come with more complex and more efficient way of enforcing restrictions (like ODbL). --mnalis 20:50, 8 December 2009 (UTC)
It is a compromise to make OSM as easy to use as possible for the largerst number of uses without giving up (an enforceable) Share-Alike. Independent of personal favour of CC0 vs Share alike, I think one has to acknowledge that the status quo is intended to be share-alike and a substantial fraction of the community (including some of the big data imports) wants share-alike. The chance of a license change succeeding to CC0 is thus by far smaller than to ODbL which in large is very similar to the intent of OSM when using CC-BY-SA . --amm 21:11, 8 December 2009 (UTC)
mnalis - I think you're still under the impression that CC-BY-SA could be declared to be "effectively public domain". It can't. An unclear licence can't be definitively resolved one way or the other - not without several test cases. It would simply remain unclear.
The result of that is that people won't, don't, use OSM. Let's say Zing Maps, for example, want to use OSM. They get their lawyers on the case. The lawyers say "yes, this licence doesn't work for street data in the US, and that's what we want to use, so it's effectively public domain". Fine. By your argument, that's all they need.
But they don't use it, because the next sentence the lawyers say is: "But we cannot unambiguously promise that we won't get sued for using it." Another set of lawyers, or another jurisdiction, may have a different view. Zing do a risk analysis and work out that it's cheaper to pay for Tele Atlas data, rather than use OSM data for free but have a 1 in 10 chance of being sued.
This isn't hypothetical. This is what Zing and Zoogle and so on have said. Even Google, for whom I have very little time, have said in public they'd have used OSM data if OSM had "sorted their licence out" (verbatim quote). --Richard 13:59, 13 December 2009 (UTC)

Staying with CC-BY-SA for now and switching only if it's proven invalid

We've been staying with CC-BY-SA for five years now. If we wait, but then have to change the licence in another five years, the relicensing will be yet more painful still.

As for "proven invalid": the world's largest search engine, the (at the time) second biggest online map site, and a British television company have all said at one time or another they will not use OSM data because of the current licence. The same applies at the other end of the scale - artisan cartographers here in the UK still use Ordnance Survey data in preference to OSM data.

When you say "CC-BY-SA is shown invalid for data, and hence all OSM data is actually public domain and free for grabs", once again, it isn't that simple. More likely, some OSM data in some jurisdictions is public domain (e.g. street networks in the US); other data in other jurisdictions isn't. We need a licence which is capable of applying in most cases, and that's either ODbL or a public domain declaration such as CC0. My personal preference is for the latter, but given how much debate simply changing from one sharealike/attribution licence to another can generate, I think changing from a sharealike/attribution licence to a PD declaration would be verging on impossible.

--Richard 14:16, 13 December 2009 (UTC)

Not really Share Alike

CC-BY-SA has a Collective Work provision, just as ODbL has a Collective Database provision - the difference in wording being because CC-BY-SA is a creative works licence whereas ODbL is a database licence.

OSM has long said that, if you overlay a set of data onto OSM (a "mashup"), this is a Collective Work rather than a Derivative Work. This is known as "Imi's layer clarification" (because the person who first wrote it on the wiki was Imi, original author of JOSM and probably the most fervent sharealike advocate OSM has ever had).

ODbL does not change our approach at all.

There are two significant differences in ODbL's Share-Alike compared to CC-BY-SA's. Firstly, ODbL requires you to share the database source as well as the result, which CC-BY-SA does not. (ODbL is more like the GPL in this regard.) Secondly, ODbL does not require products ("Produced Works") made with ODbL-licensed data to be released in their entirety, but again, it does require the source database behind them to be. But the Derivative/Collective distinction is not significantly different.

--Richard 14:09, 13 December 2009 (UTC)

Creative Commons' view

The page says "Creative Commons have made an excellent case against using the ODbL".

Indeed, CC do not recommend the ODbL. They recommend public domain - no share-alike, no compulsory attribution. They do not recommend CC-BY-SA for data.

Their argument is not a recommendation for voting "no" unless you, too, would rather the entire database was made public domain. Bear in mind the OSM Foundation has already agreed that users will have the option to flag their own work as public domain.

It is also worth noting that CC hasn't made a case against the ODbL when applied to OSM. Rather, they think it's unsuitable for other data sets, and are worried that the prominence which we would lend it would be damaging to their public domain ambitions. To quote John Wilbanks, VP of Science Commons: "If this were the 'Open Street Map License' and not the 'Open Database License' it's unlikely we would have such a strong opinion."

--Richard 11:51, 7 December 2009 (UTC)

Richard, I'd say there is a big difference between allowing individuals flagging PD and the whole project moving to PD. The first would only be a subset that is not actively maintained (e.g. I doubt routing would work on a subset culled from an OSM ODbL database) and the later would be a more complete data set which is actually usable and actively improved. The option to flag individual contributions is a bit of a red herring in this discussion IMHO.
I also fail to see why CC's comments on ODbL don't apply in OSM's case. They might separately be concerned by license proliferation but the issues still apply as far as I can tell. --TimSC 13:08, 7 December 2009 (UTC)
"Richard, I'd say there is a big difference between allowing individuals flagging PD and the whole project moving to PD." Quite. In the interest of pragmatism, I'd probably be willing to release my contributions into the public domain if the whole project moved to PD. I'm not willing to release my contributions into the public domain if the project goes ODbL. In any case, I think the best solution at this point is to continue to use CC-BY-SA, though with a disclaimer that CC-BY-SA doesn't apply in jurisdictions where the data is public domain by law. Then, in jurisdictions where the data is PD, it'll remain PD. In jurisdictions where the data is not PD, it'll remain CC-BY-SA. Anthony 18:47, 13 December 2009 (UTC)
  • The ODbL Fails to Promote Legal Predictability and Certainty Over Use of Databases - a great reason to vote no.
  • The ODbL Is Complex and Difficult for Non-Lawyers to Understand and Apply - a great reason to vote no.
  • The ODbL Imposes Contractual Obligations Even in the Absence of Copyright - IMHO the most important reason to vote no.
  • The ODbL is a bad license for OSM. That's why you should vote "no", regardless of what you think a good license would be. Yes, if you want the database to be public domain, you should vote no. After all, if the entire premise of this change is correct, and the OSM database is not copyrightable, then a no vote gets you exactly what you want. In any case, the purpose of this page is to present the arguments, and to let people decide for themselves. I feel it is quite important for those making this decision to read the entire page of what CC had to say and then decide for themselves if it is relevant. Reading it has helped me personally solidify my decision to oppose this change. Anthony 18:41, 13 December 2009 (UTC)

ODbL data based on CC-BY-SA data

The process is explained on OSM:Open_Data_License/Backup_Plan, especially here:

3. Each element is examined and only those with an unbroken history chain from version 1 to the most recent ODbL'ed version are marked as "OK".

Which means that if user B modifies nodes/ways/relations of user A, that would be considered derivative work and such nodes would be removed (or reverted to last revision before "user A" touched them, but in your example "user A" was first to create them, so they would be removed).

However, the completely new nodes/ways/relations created by user B which do not touch user A data (for example, if user B enters a completely new node such as a POI) would remain on the map.

But any data that was partly or in full based on user A data would be either removed or distorted (depending on amount and nature of data overlapping).

If've tried to warn on interleaving issues here --mnalis 19:31, 20 January 2010 (UTC)

The current license is not broken / ODbL is unproven

Can we turn that statement by 180 degree to say "ODbL is not broken / CC-BY-SA is unproven"? That is about as accurate as the original statement and follows the same logic. To all of my knowledge, there has not been a single case in which ODbL has been found to not hold up in court, so ODbL must be a good and enforceable license. Likewise, CC-BY-SA has so far again to my knowledge not been tested in court, so you could say it is completely unproven, worse even, most lawyers seem to think it won't hold up for databases which isn't the case for ODbL. Now, of cause that argument is pretty flawed, but I don't think the original argument is more sane, particularly the "current license is not broken part". As much as I love OSM, I would say there aren't all that many large chunks of map that are uniformly better than TA or NavTeq data quite yet, so there aren't all that much incentives to steel the data so far, and the current license was just about enough for the hassle of going through courts being higher than the worth of the data. That does however not mean that in future people won't try and succeed as the value of OSM rapidly increases. So please don't use the past as a predictor of the future, or the lack of evidence as evidence of lack, without clearly stating the problems in doing so.

--amm 22:57, 5 December 2009 (UTC)

As far as I know, CC has never been tested in court either?
TomH 00:30, 6 December 2009 (UTC)
I suggest that people should not start changing the meaning of peoples statements. I do suggest (see dispute section above) that the LWG should have the right to add a rebuttal to a claim. The original author may then wish to rephrase their claim and request that the rebuttal is adjusted. If we get into group editing of the claims we have a huge likelihood of a flame war. PeterIto 05:24, 6 December 2009 (UTC)
Sorry, I don't quite understand your objection to my comment. I wasn't suggesting to actually changing that statement on the wiki. That was a rethorical question in response to the line of argument I object to in the two points that I wanted to highlight by exagerating the argumentation into the (unreasonable) extreme. amml 20:41, 8 December 2009 (UTC)
Also CC BY-SA is relatively new and untested in court AFAIK. Jaakl 09:03, 6 December 2009 (UTC)

Users opposing ODbL

Users who oppose the planned ODbL-fork of this project may join OSM::Category:Users_Rejecting_ODbL. --amai 14:37, 21 February 2010 (UTC)

OSMF Statement is no longer relevant

This page is linked to from the osmfoundation website and sets the arguments for why people should not adopt the ODbL. The section provided by the OSMF, titled "OSMF Statement" is no longer relevant as the vote has long since past. I've attempted to clarify this by adding a section at the top that puts the OSMF Statement into context. OSM:user:firefishy disagrees with this change. The following is a transcript of a discussion with firefishy about this matter. As we didn't come to a satisfactory conclusion I'm reverting the page to the October 10th state until it has has been reviewed by LWG this evening.

[14:57:37] 80n: Hi
[14:57:47] firefishy: hey
[14:57:52] firefishy: MOAR hardware?
[14:58:29] 80n: I just wanted to chat about the Why you should vote No page.
[14:58:37] 80n: ...on the wiki.
[14:58:42] firefishy: yup yup
[14:59:16] 80n: I'd like to remove the OSMF Statement, but I see that you reverted it.
[15:00:57] firefishy: Those pages were put there for the OSMF Members vote. The whole voting thing keeps getting distorted.
[15:02:02] 80n: Indeed, but the vote it refers to has passed, however there are still active links to this page from the current license change material.
[15:03:05] firefishy: Which material? Happy to correct.
[15:03:08] 80n: http://www.osmfoundation.org/wiki/License/We_Are_Changing_The_License
[15:03:23] 80n: This page has a direct link to the Vote No page.
[15:03:49] 80n: Which is fine, but we need to be able to keep the content of the Vote No page relevant.
[15:04:00] firefishy: It does: from this section: http://www.osmfoundation.org/wiki/License/We_Are_Changing_The_License#Have_other_options_been_considered.3F
[15:04:34] firefishy: OK but there isn't a vote.
[15:05:04] 80n: Perhaps we need two new pages, rather than distorting the arguments that were applicable at the time of the vote.
[15:06:35] 80n: I added a section to try to clarify the purpose of the page, but it needs to be at the top because the OSMF statement doesn't make sense to people arriving from that page.
[15:08:56] firefishy: I've moved the pages: http://wiki.openstreetmap.org/wiki/Open_Data_License/OSMF_Members_Vote/Why_You_Should_Vote_Yes and http://wiki.openstreetmap.org/wiki/Open_Data_License/OSMF_Members_Vote/Why_You_Should_Vote_No
[15:09:06] firefishy: and added http://wiki.openstreetmap.org/wiki/Open_Data_License/OSMF_Members_Vote
[15:09:58] 80n: What does that achieve?
[15:13:18] 80n: I propose we do one of three things:
1) Update the osmfoundation page to link to two new pages where relevant "what does it mean if you vote yes|no" information can be presented by both camps.
2) Leave everything as is, but remove the OSMF statement from the Vote No page.
3) Leave everything as is, but move the Important section on the No page to the top.
[15:17:49] 80n: Do you have a preference?
[15:20:46] firefishy: I'ved tidied up the page headers and made them match. There is this landing page now: http://wiki.openstreetmap.org/wiki/Open_Data_License/OSMF_Members_Vote
[15:22:48] firefishy: If you want to create a page for why people should not license their data under ODbL then go ahead.
[15:25:33] 80n: Would you please STOP making changes while we discuss this.
[15:26:09] 80n: Can you revert your changes please.
[15:27:09] firefishy: On what grounds?
[15:27:44] 80n: Because I'm trying to reach a consensus with you about what this page should contain.
[15:28:52] 80n: I suggested a choice of three options. You chose to ignore that question and edit the page in some other way instead.
[15:30:12] firefishy: As I have said, those pages were for the OSMF Member Vote, those pages were not clear. You seem to want to distort what those pages are for. I modified the pages to make quite clear why does pages were created. I have encouraged you to create a page on why people should not sign up to ODbL and you seemed to have declined.
[15:32:47] 80n: The problem is that the osmfoundation page about the *current* license change process linke to these pages directly.  Since these pages refer to a vote that happened some time ago they are not very relevant.
[15:33:59] 80n: I think we should have a completely new page as you suggest, but it needs to be linked to from the osmfoundation page as this is where the problem is being caused.  Can you change the links in there please?
[15:34:37] 80n: That way we can keep these Vote Yes / No pages intact as they were at the time of the vote.
[15:43:18] 80n: Are you able to change the osmfoundation page?
[15:46:09] firefishy: Yes. http://www.osmfoundation.org/index.php?title=License%2FWe_Are_Changing_The_License&diff=494&oldid=484
[15:46:19] firefishy: I will discuss further this evening with LWG.
[15:47:28] 80n: That is not what I asked you to do.
[15:48:40] 80n: I asked you to change the links to point to a new page (I understand that you'd probably want to ok that with the LWG), but I also asked you not to change anything until we had reached a consensus.
[15:50:06] firefishy: Feel free to create the page(s). When created ask the LWG or OSMF to link. I'll bring it up with LWG tonight.
[15:50:27] 80n: Could you please revert all your changes in the meantime?
[15:52:07] firefishy: No.
[15:52:57] firefishy: If you have felt I have been dishonest in any of my changes then please complain.
[15:53:12] 80n: Ok, lets talk tomorrow.
The above was a private discussion I had with 80n on Skype chat. I feel quite aggrieved that he posted this private discussion with him without first checking with me if it was ok. Oh well. -- Firefishy 18:04, 30 November 2010 (UTC)