Talk:Open Data License/What is clean?

From OpenStreetMap Wiki
Jump to: navigation, search

Purpose of Page

Shoudn't this page limit itself to discussion of what we believe is untainted data that can be reused? The page starts off ok but then drifts in to tips and methods for remapping, those should really be on the Remapping page. SimonPoole 11:50, 6 December 2011 (UTC)

Safe approach discussion

An object is not necessarily dirty if any mapper in its history is anonymous

That is a safe assumption in that OSM data would be safer if data contributed by the anonymous account were removed and remapped. That does not tell the whole story, and this safe assumption is likely to change. Rw 22:37, 2 December 2011 (UTC)

  • Currently we have no way to distinguish anonymous mappers who have accepted CT/ODbL from anonymous mappers who have declined CT/ODbL from anonymous mappers who have not responded. Rw 22:37, 2 December 2011 (UTC)
  • Now, a list of changesets by anon-contributors who have agreed to CT/ODbL is available This list will only be updated periodically as a way to continue to offer anonymity to these grandfathered anonymous accounts. Rw 13:53, 14 December 2011 (UTC)
    • Is the anonymity issue in any way related to licensing or License Change, or are they two separate issues? (Balrog 19:52, 16 December 2011 (UTC))
  • I don't understand why the OSMF is protecting the copyright of anonymous authors. I think the position should be that anonymous users may choose to identify their edits and keep their copyright protection. The principle here is the same as that which requires that property owners who wish to exclude trespassers need to post their property and identify themselves on the posters. I can't think of any context in which property owners are allowed to remain anonymous. RussNelson 15:56, 21 January 2012 (UTC)
    • Further, a valid copyright notice says "Copyright", gives the date of first publication, and names the copyright holder. If the copyright holder declines to be named, I suggest that they are declining to publish a copyright notice, which means they are abandoning their copyright. RussNelson 14:39, 28 January 2012 (UTC)
      Copyright exists regardless of any notices. The US required a notice, but that changed like decades ago. Alv 09:11, 10 February 2012 (UTC)

Deleted Object/Tag discussion

Currently the document assumes in multiple places that deletion is not a creative act and does not warrant protection. While I would tend to agree in the context of typical editing in OSM, I do offer the example of a sculptor or carving where the whole process of creating the work of art is removing unwanted material, the resulting work is naturally clearly protected. SimonPoole 17:44, 3 December 2011 (UTC)

If we say that naming a point is copyrighted, then removing is even copyrighted and have to been restored. --Quedel 17:10, 20 January 2012 (UTC)

Clean-ness can't be decided based on object history only

The page should probably refrain from giving statements on "what is clean" and instead say "what is dirty".

Safe approach says

An object is clean if all mappers in its history have either accepted the licence change or joined the project after acceptance was required at signup

This isn't a safe approach, for instance because users have been massively copy-pasting objects in JOSM in order to clean the object history. There's nothing in OSM contributor terms or guidelines that prevents doing that. Splitting a way in an editor can have a similar effect.

Moving a node cleans its geometry

Edge cases says Non-agreeing mapper touched a geometry-only node (no significant tags), a later agreeing mapper changed its position node None of the facts provided by the non-agreeing mapper live on, we invest our usual trust in the remapper to cleanly derive his facts from other sources.

Mappers adjust the position of nodes all the time only with reference to the existing OSM map. If the map has a phone-booth just north of a road and I adjust the road 50m and move the phone booth at the same time so it remains north of the road I might have no source for that phone booth other than OSM. My edit of the phone booth doesn't make the geometry of the phone booth ODBL clean if node was placed north of the original road by a decliner and if I have no other source for the phone booth OSM. Stevens 03:41, 13 December 2011 (UTC)

In your example, the phone booth would be a single-node POI, would it not? If it was created by a non-agreeing account, it would have to be removed or remapped; simply moving it would not clean the phone booth. If it was created by an agreeing user, then moved to the north side of the road by the non-agreeing user, then only the position is tainted. You then adjust the road and phone booth because you have updated imagery or a GPS trace or some other source that drives you to improve the map. The road is fine in the example. You move the phone booth to a better position according to your judgment. That's good mapping. It also removes all trace of the non-agreeing mapper involvement in that node. Rw 14:04, 14 December 2011 (UTC)

Deletion Paradox?

What is that all about? You could just as easily say:

User A creates a node.
User B deletes that node.
Both user A and user B are non-agreeing.  
To clean the node work backwards through the versions.
To clean the deletion of the node by User B the node must be un-deleted.
To clean the creation of the node by User A the node must be removed from the database.
Both original actions have been reversed leaving a clean database.

No paradox. Having said that, if everything currently marked as deleted is removed from the database then there will be no risk of someone undeleting a previously deleted but otherwise dirty object. --EdLoach 15:59, 14 December 2011 (UTC)

Yes, that's easy. But what do you do if A is agreeing and B isn't? Will you have to restore the node because you have to remove B's contribution? --Dieterdreist 20:51, 18 January 2012 (UTC)
That's the problem with all other things deleted by non-agreeing users. You have to do the same as you would go on if non-agreeing user deleted values or tags or nodes or ways. Even i think there is only one paradox: if you undo edits by A and B you have it clean, even with same status as not clean it. --Quedel 16:50, 20 January 2012 (UTC)

Deleted History

All history prior to 7th October 2007 was lost when the API was upgraded from 0.4 to 0.5.

An email from that date confirms this [1]:

   4. History cleared. History will continue to be written as before,
   but we have removed past history data from the database today. When
   accessing existing objects you can still see the person who last  
   modified them (even if that modification was before the switch), but
   no details about any previous modifications.

Everything that was not version 1 on 7th October 2007 has an incomplete history and ought to be considered to be unsafe.

  • This is confirmed to be false, at least partial history from before 2007 is available. (Balrog 19:53, 16 December 2011 (UTC))


If a way is split, one half of the way will be new and does not inherit the history of the other half.

Ah, but the nodes of the original way will still remain. So the new way that is split from the original way is "tainted" with the history of the nodes it references.
Nodes only affect geometry, not attributes of ways. Split, merges or copy-pasted objects (e.g. moving tags from a node to building) can be detected by comparing identical tags deleted from one object and added to another inside the same changeset. In the case of way splits it's even easier because the script can notice that nodes that previously were part of one way now are part of the new way. This is very basic and I'm amazed that even dumber rules are being proposed on this page.


If two ways are merged the history for one of the ways will have been discarded.

Suggestions by FK270673

  • Moving a node creates a new work (if we consider nodes a single node as artistic work).OK!
  • Tags like street names are considered facts, at least in these countries where legislation says so.
  • Ways are considered joint work, any way last edited by agreeing mapper is considered his work and thus clean.
    • Mapper A creates a new way with 10 nodes
    • Mapper B splits this ways into two ways with 5 nodes each.
      • -> Mapper B is called creator of the first part though he did not add any node
      • -> Mapper A is called creator of the second part though he added all nodes of the first part as well
    • Mapper C moves all 10 nodes to a better position
      • -> Mapper C has contributed 100% of all nodes. He has "created" the current geometry
    • Mapper D adds another five nodes to both ways
      • -> Mapper D has contributed 100% of actual geometry and 33% of nodes. Mapper C has contributed 67% of nodes. Mapper A and B have no influence in actual geometry, but their nodes may be used as a reserve.
  • Relations are considered facts not protected by copyright.
  • Non-responding users with less than 1,000 nodes who did not respond to multiple mails should be treated as acceptors. They have already implicitly transferred their copyright to OSMF (at least according to Continental copyright law).
  • Some mappers have lost their password or email account (e.g. while graduating or being laid-off) and cannot reach the Accept button any more. Some have died, have become ill or have lost their relatives. Solicitating widows with license agreements is absolutely indecent!
  • Don't forget: Currently I know no mapper intending to sue OSMF, but many, many mappers who have stopped contributing or even disagreed because of the excessive deletion policy. At least two or three mappers in my neighborhood responded they would agree if OSMF had a less restrictive deletion policy. With their non-agreement, they want to demonstrate the devastating and irreparable damages of a rude deletion policy. One declining mapper has even said he would like to wait with his acceptance until a new LWG with less restictive policy comes into office.

--FK270673 12:20, 15 December 2011 (UTC)

Suggetion for FK270673

  • Replace all your points with one "Lets name stealing as finding" and then treat imported data in OSM as found. Owner of copyright for imported data in OSM

Your copyright (or, to be more precise, your database right) covers those nodes and ways which were imported from UMP. But does your copyright also covers those single nodes which have been MOVED meanwhile by other editors? I don't think so, because Creative Commons explicitly allows the creation of derivative works. You may reuse these moved nodes, but you may not prohibit the relicencing of moved nodes. There might be a legal necessity to remove some UMP imports. However, my arguments applied to those individual mappers who have contributed about 50 nodes five years ago in a city where agreeing mappers have contributed about 98% of data. If you have explicitly declined AND contributed substantial data (not just five nodes and two ways who have been edited by others), then you have a legal right to remove your data. It would be helpful if you would tell us the number of objects that is covered by your claim. In your case, it would be helpful to present a list of generic UMP nodes in order to assess which unchanged nodes need to be removed. --FK270673 10:10, 23 December 2011 (UTC)

My suggestions rather applied to countries like Hungary, where only 19,278 of 2,672,897 nodes, only 2,489 of 216,274 ways and only 22 of 9,890 relations have been edited by decliners or non-responders: So it would make sense to consider Hungary as a "clean" country where no deletions are legally necessary. In Poland, 3,266,296 of 10,106,083 unchanged nodes (26.96 %) are in danger of being deleted: That's a really substantial part and not an assumption of marginality like in Hungary. So it would make sense to assess each country differently. I have already suggested to offer separate licences for each country (e.g. to keep CC-by-sa valid for Poland, but not for Hungary) as Creative Commons allows to assemble objects (countries) with different licenses within an ODbL database. However, they told me that OSMF would rather delete one third of Poland than allowing a special license for a single country. It would make sense to apply the ODbL licence for the UMP database as well, as it is very doubtful whether a moved node is still covered by copyright of the original creator. If you continue to refuse your consent, it's only your personal responsability if Poland disappears from the map. --FK270673 11:16, 23 December 2011 (UTC)

Objects that lost all information from V1

What about objects (primarily nodes) created by a non-agreeing mapper that lost all information from that version, so no tags, no way or relation memberships that have been created by the non agreeing mapper(s) live on and the coordinates changed so much that nothing of the original creation remains? Such nodes (and ways and relations) might exist and shoud probably not be deleted. --Errt 11:47, 18 December 2011 (UTC)

User accepted but source of their work was CC

Data isn't already clean if only users accepted the new termsn. If they use any cc-licensed-material (especially cc-by-sa) their contributions aren't clean and the data have to be removed. So all changeset-descriptions have to be filtered out, if there is any cc-hint and this data have to be deleted. --Quedel 17:51, 20 January 2012 (UTC)

Public domain data

Data in the public domain does not have an owner. With no owner, there is nobody to grant a license. Thus, such data is not being licensed under the current license, and won't be licensed under any change in license, and there is no reason to delete public domain data, EVEN IF it was contributed by a mapper who doesn't accept the license change.

I hope that all current "is it clean?" software looks at the "public domain" checkbox on user settings. RussNelson 14:46, 28 January 2012 (UTC)