Talk:Open Database License/Implementation Plan

From OpenStreetMap Wiki
Jump to: navigation, search

Translation

Da wird schon von "Implementierung" gesprochen - wo doch noch nicht mal klar ist, wer was genau will und warum, und welche Vor- und welche Nachteile damit verbunden sind.

Vielleicht kann das mal jemand übersetzen?

Danke, --Markus 11:46, 18 July 2010 (UTC)


Talk: Implementation Plan

  • Someone has commented on the list that briefing of major contributors such as AND should be included in the plan MikeCollinson 13:29, 28 February 2009 (UTC)
* This has now been added, date is outline and suggested waiting until OSMF members have voted to proceed. -- Firefishy 00:22, 6 March 2009 (UTC)
  • I think the +1 week agreement ought to be provisional. Then I think there needs to be a tool which allows you to see the effect on the map according to people's provisional agreement. That way we can see what the effect will be if we agree to move forward - in effect we take a straw poll. Then people's decision becomes binding just before the +1 month meeting. So if the effects look locally disastrous, one can change one's mind about agreeing. The tool to see the potential effects is vital even without a provisional agreement so at least there is scope to identify and repair damage afterwards. David.earl 14:01, 28 February 2009 (UTC)
* Partial: +1 week has been expended to 2 weeks. The "+" items are there to detail that they are provisional and not hard and fast dates. An email to reminder to non-responders at end of first week could likely carry acceptance stats. -- Firefishy 00:22, 6 March 2009 (UTC)
  • Discussion on legal-talk also indicates that people want feedback on the amount of data that will be lost to help make their final decision. Having an initial straw poll, later making it binding, seems like a good idea. I think that all should be reminded that they can change their mind and be able to see current acceptance stats before the time is up.—Sward 08:20, 16 July 2010 (UTC)
  • I think it is a bad idea to plan for the release of 1.0 at the OKCon event because it is not unlikely that the feedback in the project mandates further work on the license and a delay. If 1.0 is on a fixed track to be launched at the OKCon then this would mean that OSM would have to wait for 1.1. --Frederik Ramm 21:27, 28 February 2009 (UTC)
* Jordan is going to post additional work-in-progress revisions on http://www.co-ment.net/text/844/ . OSMF/Working-group was unfortunately not involved with the choice of release date of of ODbL 1.0. ODbL 1.0 will be reviewed by OSMF board and members on it's merits for adoption, later adoption or new direction after it is released. -- Firefishy 00:22, 6 March 2009 (UTC)
  • The timeframe is much too short. There should be more time for community discussion of the draft license. I cannot see how the community members can make an informed decision about the license with almost no discussion of its merits. Longbow4u 22:20, 28 February 2009 (UTC)
* Timeframe has now been expanded. Entire plan is up for discussion. Please make suggestion. -- Firefishy 00:22, 6 March 2009 (UTC)
  • I do not believe the licence should be presented as a fait accompli to which you can vote only 'yes' or 'no'. At a minimum a range of options should be presented. -- Ed Avis <eda@waniasset.com>
* Emails, messages and wording still need to be defined. 'yes'/'no' make a lot more sense when you know what the question is. -- Firefishy 00:22, 6 March 2009 (UTC)

I don't like this at all

  • First:

it is 10 pages of ugly legaleese.

I could argue that it is not ugly :-) . Is Creative Common's legal text also ugly or not? -- Firefishy 00:58, 6 March 2009 (UTC)
Creative commons is actually much better. Plus, it is a copyright license, not legal agreement, so you can ignore it if you are not making derivative works. Pavel 16:33, 7 March 2009 (UTC)
  • Second:

ODbL explicitely does not cover art (etc) you merge it with. Under CC-by-SA, if you create beatiful painted map partially based on openstreetmap, everyone automatically has right to copy that map, along with the paintings... and that's good. ODbL only covers database license...

If ODbL is adopted, people are alowed to create maps based on openstreetmap, and we will not be able to copy/use those maps. That just does not sound fair.

There are ongoing discussion around these issues. Slightly different argument but shared ideas, If I use GIMP to create images do they need to be Free (libre/gratis)? No, because it would not benefit GIMP. If I use Photoshop to create images do they need to be closed? No, because it would make no difference to Photoshop. If I use OSM to create map images (arguably) they are of no (or very little) benefit back to open geodata. Are we trying to free geo-data or paper maps? I would argue freeing the geo-data makes more sense because then I and others can make the paper maps. Mapnik, Tiles@home, etc are there and will continue to be great apps for me to do this. Open Source! -- Firefishy 00:58, 6 March 2009 (UTC)
The gimp example is misapplied. gimp is a tool, not a dataset. And yes, I do want free maps, both paper and online. If you render openstreetmap and sell it, I expect to be able to copy that output. If you create online service based on openstreetmap data, I expect to be able to be able to print the renderings and use them. Pavel 16:33, 7 March 2009 (UTC)
Actually, the case "underlying data are free, rendered maps are not" is similar to gimp being BSD licensed, and people putting restrictive copyright notices on generated binaries. There are some reasons I prefer GPL... Pavel 16:44, 7 March 2009 (UTC)

Third:

+ 1 week
   * Website to allow users to voluntarily agree to new...

Yeah, that sounds voluntary. "Agree to the crap we invented, or we delete your data".

+1 - Just finished deleting all original contributions I made. If anyone undeletes them I will write a vandal-bot --DiverCTH 22:42, 11 May 2010 (UTC)
That sounds like a very immature and stupid decision to me. If you once accepted the CC-BY-SA terms, why not keep your data available under those terms? Your contributions will now be lost in the final CC-BY-SA planet map. And in case of a fork continuing to use CC-BY-SA, that fork will also lack your contribitions. It would have been a lot smarter just not doing anything. Then your contributions would have been removed AFTER the last CC-BY-SA map was made. --Henriko 15:11, 26 March 2011 (UTC)
We are not going to delete any data, but we might have to remove it from the central database if the greater community decides that we want to release the data under ODbL. The removed data will continue to be available in old planet.osm files. The circle of groups of people are growing and eventually everyone will need to be asked. -- Firefishy 00:58, 6 March 2009 (UTC)
You are forcing people to fork. Pavel 16:33, 7 March 2009 (UTC)

Fourth:

+ 2 months
   * Final cut-off.

Hmm. openstreetmap is in big part covered by imports, and companies provided data for imports under CC-by-SA. Having to negotiate imports once was bad, but having to do it second time is stupid... and 2 months do not seem like enough time.

An item has been added to the proposed plan. 2 months is only provisional and will likely need extra time to conclude with all the parties involved. But like with all deadlines most things only happen close to the deadline anyway. If we said 1 year from now, I suspect people would only start to act near the end of that deadline -- Firefishy 00:58, 6 March 2009 (UTC)

So prepare to delete Czech Republic.

You the only editor of Czech Republic then? :-) We are trying to be open and invite people to take part in the process. -- Firefishy 00:58, 6 March 2009 (UTC)
Not the only one, but I did some quite big imports. Pavel 16:42, 7 March 2009 (UTC)

Fifth:

What are the benefits? I don't see any

This is covered by http://www.opengeodata.org/?p=262 . Discussion at State of the Map 2007 and 2008 have been positive in this direction. CC-BY-SA is too ambiguous a license for geodata and our needs. -- Firefishy 00:58, 6 March 2009 (UTC)
The linked article looks like a lot of FUD. (And it should be in the wiki in the first place, so that it can be commented on.) Plus, proposed new contract is not solving any of it; if someone wants the data, he can get his cat accept the EULA and then use the data in any way. So... the official excuse for switching to contract (evil corp not giving changes back) is really an excuse, as evil corp + cat can still not give the data back. Pavel 16:33, 7 March 2009 (UTC)



-- Pavel 00:50, 4 March 2009 (UTC)

Revising a vote?

Is the answer on the vote at +2 weeks final? Or is it possible to swing a yes to no or the other way around for as long as no actual decision is made on moving to the next license? Given how things evolve during the process opinions can change, so in no way should the first vote be final. --Eimai 13:51, 15 March 2009 (UTC)

"What do we do with the people who have said no or not responded? "

Well, no response means obviously that the conditions & terms apply that we all have agreed when joining the project. So basically "no answer" is "Yes, I want the orignal license", or "No" as is the modern version of the answer ;-) And what to do? Except from disabling log in, see if some have enough spare time to keep an OSM alternative project running... --Amai 00:04, 23 December 2009 (UTC)

Notice on website

Are you going to tell users that agreeing to ODbL may remove their content from the future ODbL database? I think many innocenct people may erroneously agree on the fork, but later on feel pissed off by seeing their contributions not showing up cause they were enhancing previous data. --amai 18:38, 26 February 2010 (UTC)

What about users who've made use of sources under CC licenses?

If users have used data sources that were explicitly under a CC-By or CC-SA style license, they can't legally licence all their contributions under ODbL. I'm thinking in particular about any contributors that have made use of Ordnance Survey Opendata (including tracing or taking street names from the StreetView layer in Potlatch). This creates two issues: (1) how to allow users to licence only some of their previous contributions under ODbL, and (2) The need to negotiate with Ordnance Survey and any other known CC-By sources that aren't already recorded as Imports to allow use under ODbL. -- Rjw62 17:11, 12 May 2010 (UTC)

This would also include users who have contributed data based on NearMap's aerial photography. For Australia, that's a significant proportion of the data. -- benlast 05:29, 29 Jul 2010 (UTC)

"Week x" timings?

How are those actually counted? For example, if it says "Week 8 (approximate)" - does it mean abount 8 weeks starting from some specific important date (like 12 May 2010 or 8 June 2010)? Or does it mean 8 weeks starting from *today* (in which case it should be updated at least weekly)? Or something different? It would be nice for it to be clarified (or even replaced with day.month.year approximate dates in order to be more clear; even if it means taking a few minutes more when updating wiki) --mnalis 17:04, 20 June 2010 (UTC)

My guess (not being a member of the licensing group) is that they are fairly arbitrary. The whole process has taken so much longer than expected and thus so much behind schedule, that probably any date (in the future) on the Implementation Plan should probably be treated with a mountain of salt... ;-) But originally, I believe they were intended as relative to a "starting data". However, what you probably can assume is that "Week 8" is likely to be after "Week 3". Amm 19:32, 20 June 2010 (UTC)
There is an outstanding LWG task on me to convert the 'Weeks' into the bold titles. --Firefishy 08:37, 21 June 2010 (UTC)
Has phase 2 now begun officially? Should the text be updated to reflect this and the minor changes to the process that appear to have been made? Or is it not yet ready to be announced more widely? Amm 01:42, 12 August 2010 (BST)
Yes phase 2 has started. Publicity should start later today. -- Firefishy 06:01, 12 August 2010 (BST)

Anonymous edits?

What would happen to (so long ago made, but still present I assume) anonymous edits, and data based off them? --No signature

Related question: will the "anonymous" status of his edits be kept after the license change, if an anonymous user (who not confirmed "make all my edits public") agrees to ODbL? --Plasmon 23:34, 15 November 2010 (UTC)
For example, 22,73% of the nodes in Södermanlands län (the county immediately south of Stockholm, Sweden) have been edited by anonymous users [1]. It would be absurd to simply delete them. When editing anonymously, one can probably assume that these people won't be claiming any copyright for them, and relicence the edits. These users are also impossible to get in touch with. It's a legitimate concern, and deserves an honest answer! /Grillo 05:49, 23 November 2010 (UTC)
No, they are not "impossible to get in touch with". The OSM Foundation has the e-mail address with which they registered, just as it does for every other user. Also, please don't assume bad faith by implying that there won't be an "honest answer". --Richard 14:30, 23 November 2010 (UTC)
Anonymous user does only mean, that their edits are not public. Their whole history of edits is still stored on the OSM servers, but you are not able to have a look on it. --Plasmon 20:03, 23 November 2010 (UTC)
OK, I was not aware of this. Have these users been contacted? /Grillo 01:21, 3 December 2010 (UTC)
Not individually, as no mapper has been contacted individually (e.g. per e-mail), see Phase 2. They can find the information in the wiki or on OSM homepage, of course. If they want, they are able to accept the license. --Plasmon 17:46, 6 December 2010 (UTC)
Individual OpenStreetMap contributors were contacted via their registration email addresses in May 2011. - User:Rw 23:05, 25 June 2011
All contributors were contacted? Or just those who hadn't answered the question yet? I didn't receive any such contact. Eric2 11:14, 26 June 2011 (BST)

Fork?

I've just spent quite some time entering data into JOSM, and when I try to commit the changes to OSM, I'm told I'm not allowed to any more.

Does anyone have any news about possible fork projects to which I can submit my collected data? Eric2 20:29, 25 June 2011 (BST)

See Forking and [List of forks in planning or production] --mnalis 14:57, 9 July 2011 (BST)
Thanks. I was looking at fosm.org but it still seems at quite an early stage, and I can't even see any maps on there. Plus none of the links work and I can't see any forum / wiki / community either. It doesn't even bring up the home page on my regular browser. Similarly the Ccbysa fork still seems to be just ideas. :( Eric2 22:34, 14 July 2011 (BST)

How to find CC-BY data and replace them

Hi, I'm looking for something (a map?) to help me to find all data still under the old licence, and that are going do be deleted. I would like to focus on them to replace with brand new ones. --Marc 09:59, 11 July 2011 (BST)

Agreed - there doesn't seem to be anything I can find in graphical form of what will disappear when old data is removed. If there was, I could ensure that in my local patch I replace this data with new surveys.
Also, how is the data to be removed? Is it going to be based on last editor, or creator of the way? What about where nodes are tagged on non odbl people's ways, e.g. with a maxspeed tag, by someone who has accepted the terms? What happens with way splits for these people's ways? C2r 18:38, 25 July 2011 (BST)
Those are very good questions. Have a look at the so-called Backup plan (and its talk page) for more discussions on this. The way I understand it, if you take two hypothetical contributors, Peter (who has not accepted the switch) and Paul (who has accepted the switch), then the following happens:
  • Peter creates a way. Paul modifies the way, adding nodes, moving nodes, adding maxspeed and other tags. => The way gets deleted and both Peter and Paul's contributions are removed.
  • Paul creates a way. Peter modifies the way, adding various things. Paul then further modifies his own way with lots more changes. => The way gets reverted to before Peter changed it, so Peter's changes and Paul's subsequent changes (to the way he himself created) are removed.
Note that Peter may have deliberately declined the license switch, or he may simply have never heard about it having left the project. Either way, Paul's efforts (even though he accepted the license switch) are lost.
In my personal opinion, that's a terrible way to go about things, but it's the only way to handle this license switch. It's important to note that if you agreed to the license switch, you agreed to this deletion. If it doesn't sound good to you, or you're surprised by the implications, then things weren't adequately explained to you and in my opinion you should have declined.
As for split ways, that's still a mess and I don't think anyone has a good idea how to reasonably handle this. See the backup plan link for more on this. Eric2 08:57, 9 August 2011 (BST)

Proposal for implementation

I haven't heard how the changeover is going to happen. I have heard some suggestions that all ways that have been partly edited by non-ODbL-acceptors will be deleted. This of course will lose data added by ODbL-acceptors, including moved nodes and other tags. I recently suggested the following way of reducing the data loss on talk-au - I thought it might be useful to add it here. It will need a bot to do this (not my field of expertise!):

1. NODES The bot will need to go through the version list for every single node. It should start with version one of the node (or if this is by a decliner, then the first version by an ODBL-acceptor), then for each subsequent version it should compare the new data with the previous version and overwrite it (this includes the position of the node as well as tags) - unless the owner of the version has not agreed to the new licence, in which case the information should be unchanged. At the end of this process we will have a collection of nodes with ODBL-only data. Some of these nodes will have only been edited by ODBL-decliners - these will have no information (no tags, no positions) and will be deleted. Other nodes who have been edited by decliners and acceptors will have as much information as possible preserved. (e.g. decliner adds nodes on a way, I subsequently move some nodes when adding new ways - the nodes I have moved will have their positions preserved, although they would lose any tags added by decliners.)

2. WAYS In the same way, the bot goes through the version list for each way, for each version making a note of which nodes are included and in what order, and any changes in tags by ODBL-acceptors. At the end of this process, any ways by ODBL-decliners will have no tags (and no nodes as these will have been deleted already in step 1). Ways partly edited by acceptors will have at least some nodes left intact, and some information. (It might need someone to go back later to re-add nodes to the ways, and re-add some tags, but at least some tags will be left behind.)

3. RELATIONS In the same way, the bot goes through the version list for each relation, for each version making a note of which ways are included, and any changes in tags by ODBL-acceptors. At the end of this process, any relations by ODBL-decliners will have no tags (and no ways as these will have been deleted already in step 1). Relations partly edited by acceptors will have at least some ways left intact, and some information. (It might need someone to go back later to re-add ways to the relations, and re-add some tags, but at least some tags will be left behind.)

An examples:

Acceptor adds a way. Acceptors and decliners both add / move nodes. Decliner-bot adds maxspeed tag + source 'default maxspeed'. I correct maxspeed (if needed) + 'source:maxspeed=voice (or sign or survey)' + possibly add tags. The bot I have just described removes CC-BY-SA-only data. As many nodes as possible are kept, and as many tags as possible including mine are kept. (Much better than deleting the entire way.)

I hope this all makes sense!

Mrpulley 10:51, 22 December 2011 (UTC)

Perspective on post-implementation data

I've read updates including acceptance rate of contributors in the UK, List of users asked about ODbL, and updates on the status of the redaction bot. But what's the best way to get a clearer sense of the type, volume, and location of data OSM is at risk of losing? That is, the data whose contributors have not yet accepted, or declined, the new contributor terms (and is being redacted)? Ckrahe 18:53, 11 June 2012 (BST)

For a "clearer sense of the type, volume, and location of data OSM is at risk of losing" see Remapping#Tools to help you Both OSM inspector and the Cleanmap tool will help with that. A zoomed out view on OSM inspector can make things look worse than they are. e.g. in London we're losing a bunch of building outlines. The data isn't really that valuable. The redaction bot is more sophisticated. It's deciding upon actions to the put the map elements into a clean state, which sometimes won't require an element to be deleted. We don't have a map display to show exactly what the results would be, but the remapping tools give a good idea. -- Harry Wood 11:09, 12 June 2012 (BST)