Re: [OSM-legal-talk] feedback requested
I think it's relevant that node changes as suggested should involve stand alone nodes only (such as POI). Once they are part of a structure of say a building or a road, water or any area, the nodes should be considered a composition rather then 4 nodes. While the underlying structure is a geographic fact, the choice of place nodes and the number to represent the structure is a creative work. I think this seems an obvious conclusion, but should be made clear. In addition the LWG should also pay some attention to relations. Regards, Gert -Oorspronkelijk bericht- Van: Richard Weait [mailto:rich...@weait.com] Verzonden: Wednesday, December 21, 2011 2:44 AM Aan: Licensing and other legal discussions. Onderwerp: Re: [OSM-legal-talk] feedback requested On Tue, Dec 20, 2011 at 4:39 PM, Jean-Marc Liotier j...@liotier.org wrote: On 12/20/2011 10:11 PM, Apollinaris Schoell wrote: Of particular interest are: - can node positions be cleaned by moving to a new position? While you are at it, I would love to hear about a specific subset of the cases encompassed by this question : the cases where the edit is correlated with a change of source. I asked this question a week ago in the Are objects still tainted when they are edited from a better source ? thread here and it has not been answered yet. And we're listening. Tell us and be specific. Some of us have been remapping for a while. Give examples. ___ legal-talk mailing list legal-talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/legal-talk ___ legal-talk mailing list legal-talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/legal-talk
Re: [OSM-legal-talk] feedback requested
On 20 December 2011 21:27, Richard Weait rich...@weait.com wrote: Dear All, LWG would like feedback on a couple of items relating to cleaning tainted data as we all prepare for the data base transition. Draft minutes are here. https://docs.google.com/document/pub?id=1ZIQSl0xXpUFbqTeknz61BYgfCINDTzlAWomOiGxhgG8 Of particular interest are: - can node positions be cleaned by moving to a new position? These question should really be asked to a lawyer who also knows how OSM works. I understand the LWG may want to ask the community if this is worth consulting. But ultimately the cleanness criteria are not up for voting or discussion in a circle of people who obviously want to be done with this process without losing their contributions (or anyone else who's not expert in IPR really) (on the other hand there is a lot of things that could, and maybe should, be decided by all of community but instead are decided in a small group, like the date for the switch to a next phase of the license change process) - is a mapper declaration of odbl=clean interesting and helpful in reconciling the data base? Definitely, and I think odbl=no would also be useful to mark objects that are known to come from ODbL-incompatible sources but whose contributors accepted Contributor Terms 1.2.4, of which there is a significant number. Cheers ___ legal-talk mailing list legal-talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/legal-talk
Re: [OSM-legal-talk] feedback requested
I think the test must be the same as for any other data which OSMF does not have permission to use. If a mapper added a node by copying from Google Maps, but then another mapper moved it to a different position using a permitted data source, is it okay to keep that node in the database? -- Ed Avis e...@waniasset.com ___ legal-talk mailing list legal-talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/legal-talk
Re: [OSM-legal-talk] feedback requested
2011/12/21 ce-test, qualified testing bv - Gert Gremmen g.grem...@cetest.nl: I think it's relevant that node changes as suggested should involve stand alone nodes only (such as POI). Once they are part of a structure of say a building or a road, water or any area, the nodes should be considered a composition rather then 4 nodes IMHO rights on this composition can also faint, e.g. years ago a (non-ct) mapper was drawing a rough street with nodes every 300 meters. Now those initial way has five times more nodes then it had in its initial version (most probably the initial way would also be split into different pieces now, due to details like speed limits, turn-restrictions, bus routes, lane-count, ...). I think there must also be a point where nothing from the initial way is actually contained in the current data (often these initial ways don't have much attributes, it is common in here to find ways which only have/had a highway-tag (the value is now often changed, so not even one tag is the same). If you assume that other tags (like the name) would also have been inserted by the following mappers you could extend this to ways which had a name (or some other frequent tag, for which a following mapper guarantees that he would have added it if it were missing). While the underlying structure is a geographic fact, the choice of place nodes and the number to represent the structure is a creative work. +1, but where is the point that this structure is significantly changed? How many nodes do you have to move and insert/delete to be something different? What if someone takes a river, moves it aside and lets it become a track (deletes the river tags and sets highway-tag, changes name). Now he copies this way as a new way (new nodes and way) to the old position of the river and sets tags. Is the track-way now tainted because it consists of old nodes, while the river is OK because it was newly created? Admittedly a rare corner case, but IMHO one that shows that there is a point where there is no more original information in the following versions of a way. cheers, Martin ___ legal-talk mailing list legal-talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/legal-talk
Re: [OSM-legal-talk] feedback requested
A common way to adjust a node position is to move it halfway between the old one and the new one. For example, if there is already a way on the map traced from GPS but you have a new GPS trace for it which is a bit different, it would be unwise to adjust it to exactly fit your new trace. But you may expect to improve accuracy a bit if you adjust it to about halfway between the old and new positions. Similarly a node such as a bus stop may have its position tweaked to somewhere in between where it was and the new observed position. So I don't think you can assume that when a node is moved its position information is 'cleaned' somehow. The new position as often as not is derived from the old position. -- Ed Avis e...@waniasset.com ___ legal-talk mailing list legal-talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/legal-talk
Re: [OSM-legal-talk] feedback requested
Richard Weait richard@... writes: We consider that the creation of an object and its id to be a system action rather than individual creative contribution. However, 'the creation of an object and its id' never occurs by itself. At a minimum, you create an object with id and lat/lon, and that location data is part of the OSM map. The next version of the node, even if its position has been adjusted, is likely to be derived from this data. -- Ed Avis e...@waniasset.com ___ legal-talk mailing list legal-talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/legal-talk
Re: [OSM-legal-talk] feedback requested
Please don't confuse the matter by treating tagged and untagged notes the same. If somebody is improving the geometry of a way because he is interpolating from the available information (may that be GPS traces of other ways) then he is doing exactly that, just because he is reusing an existing object (pre-numbered sheet remember) to mark a new interpolated position doesn't mean it is a derived work (because it is simply no different than taking a new sheet of paper with a different number). Now if you wish to state that interpolation itself creates a derived work, please argue that. Simon Am 21.12.2011 13:10, schrieb Ed Avis: A common way to adjust a node position is to move it halfway between the old one and the new one. For example, if there is already a way on the map traced from GPS but you have a new GPS trace for it which is a bit different, it would be unwise to adjust it to exactly fit your new trace. But you may expect to improve accuracy a bit if you adjust it to about halfway between the old and new positions. Similarly a node such as a bus stop may have its position tweaked to somewhere in between where it was and the new observed position. So I don't think you can assume that when a node is moved its position information is 'cleaned' somehow. The new position as often as not is derived from the old position. -- Ed Avise...@waniasset.com ___ legal-talk mailing list legal-talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/legal-talk ___ legal-talk mailing list legal-talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/legal-talk
Re: [OSM-legal-talk] feedback requested
Simon Poole simon@... writes: If somebody is improving the geometry of a way because he is interpolating from the available information (may that be GPS traces of other ways) then he is doing exactly that, That is exactly it: improving the geometry of a way. Not replacing it. If you take an existing street and adjust its position it is hard to argue that you have taken a completely clean-room approach to doing so, not using the existing geometry at all. The existing geometry is there on your screen while you are editing! Yes, there are some cases where you might totally ignore the existing geometry (perhaps because it has been messed up by a newcomer hitting the wrong buttons in Potlatch) and recreate it wholesale. But those are a small minority. just because he is reusing an existing object (pre-numbered sheet remember) to mark a new interpolated position doesn't mean it is a derived work Agreed - that in itself is not enough. If a mapper grabbed some existing node from the database and removed its location data entirely (perhaps taking it from a global stock of 'spare nodes' kept in the Pacific ocean) then it would clearly not be derived. But why do that when you can just click to create a new node? If the mapper starts with a node that's already in roughly the right place and just adjusts it a little bit, then the new position is derived from the old one. Now if you wish to state that interpolation itself creates a derived work, please argue that. By interpolation I was referring to the practice of taking two paths (be they two GPS traces, one GPS trace and one existing way on the map, a way on the map and a path visible in an aerial photograph, etc) and combining them to make a new path which is roughly halfway between the two. For example if mapping from GPS plus an existing out-of-copyright map you may trace a way which is about halfway between your GPS trace and what you see on the old map - since neither of them by itself is entirely accurate. Doing this makes the new path derived from both the old one and the new one. -- Ed Avis e...@waniasset.com ___ legal-talk mailing list legal-talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/legal-talk
Re: [OSM-legal-talk] feedback requested
On 21 December 2011 12:43, Ed Avis e...@waniasset.com wrote: andrzej zaborowski balrogg@... writes: - is a mapper declaration of odbl=clean interesting and helpful in reconciling the data base? Definitely, and I think odbl=no would also be useful to mark objects that are known to come from ODbL-incompatible sources but whose contributors accepted Contributor Terms 1.2.4, of which there is a significant number. Hold on - is OSMF going to delete contributions even from some people who *did* accept the new contributor terms? (I'm not saying it should or it should not, but this needs to be made clear.) This has been made clear many times: whether to delete an object or not needs to be decided looking at some part of its edits history, so even if your contribution is clean, the object may be tainted. The case I'm thinking about though is where a mapper accepted CT but had previously (or later) contributed data incompatible with ODbL. Which is something that seems to be allowed by CT, as long as 1. you grant OSMF the rights that you have in the data, 2. what you uploaded was compatible with current licensing. Cheers ___ legal-talk mailing list legal-talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/legal-talk
Re: [OSM-legal-talk] feedback requested
Am 21.12.2011 13:34, schrieb Ed Avis: Simon Poolesimon@... writes: If somebody is improving the geometry of a way because he is interpolating from the available information (may that be GPS traces of other ways) then he is doing exactly that, That is exactly it: improving the geometry of a way. Not replacing it. If you take an existing street and adjust its position it is hard to argue that you have taken a completely clean-room approach to doing so, not using the existing geometry at all. The existing geometry is there on your screen while you are editing! If you take an existing tainted way and move it they way is still going to go, so what is your point again? Now if you wish to state that interpolation itself creates a derived work, please argue that. By interpolation I was referring to the practice of taking two paths (be they two GPS traces, one GPS trace and one existing way on the map, a way on the map and a path visible in an aerial photograph, etc) and combining them to make a new path which is roughly halfway between the two. For example if mapping from GPS plus an existing out-of-copyright map you may trace a way which is about halfway between your GPS trace and what you see on the old map - since neither of them by itself is entirely accurate. Doing this makes the new path derived from both the old one and the new one. You are using derived in a common language sense, please argue why this is a derived work in the IP/legal sense (choose any jurisdiction you would like). Simon ___ legal-talk mailing list legal-talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/legal-talk
Re: [OSM-legal-talk] feedback requested
Simon Poole simon@... writes: If you take an existing tainted way and move it they way is still going to go, so what is your point again? Are we not talking about the following situation: - mapper A (who has agreed to the CTs) creates a way - mapper B (who has not agreed) adjusts the way's geometry, creating some new nodes - mapper C (who has agreed) adjusts the position of those nodes In this case the third edit would have to be reverted because the new position of the nodes is still based on work contributed by mapper B, even though they have been moved since he created them. You are using derived in a common language sense, please argue why this is a derived work in the IP/legal sense (choose any jurisdiction you would like). That is a question for lawyers. I do not know whether it is a derived work under copyright law or sui generis database rights. Normally the approach of the project is to not import data from sources that do not have permission, and if it gets into the database, to delete it (reverting the changeset) as soon as possible. We don't get into the business of judging whether we might get away with including it anyway, because we are not lawyers. So we have to use the common-sense judgement of whether one piece of work builds on another. -- Ed Avis e...@waniasset.com ___ legal-talk mailing list legal-talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/legal-talk
Re: [OSM-legal-talk] feedback requested
Am 21.12.2011 14:15, schrieb Ed Avis: Simon Poolesimon@... writes: If you take an existing tainted way and move it they way is still going to go, so what is your point again? Are we not talking about the following situation: - mapper A (who has agreed to the CTs) creates a way - mapper B (who has not agreed) adjusts the way's geometry, creating some new nodes - mapper C (who has agreed) adjusts the position of those nodes In this case the third edit would have to be reverted because the new position of the nodes is still based on work contributed by mapper B, even though they have been moved since he created them. IMHO no, if we assume that C is editing in good faith and actually improving the geometry (we might want to have a minimum distance requirement for a move to be considered ok). You are using derived in a common language sense, please argue why this is a derived work in the IP/legal sense (choose any jurisdiction you would like). That is a question for lawyers. I do not know whether it is a derived work under copyright law or sui generis database rights. Normally the approach of the project is to not import data from sources that do not have permission, and if it gets into the database, to delete it (reverting the changeset) as soon as possible. We don't get into the business of judging whether we might get away with including it anyway, because we are not lawyers. So we have to use the common-sense judgement of whether one piece of work builds on another. In general we have assumed that for example tracing from aerial imagery and similar sources does not create a derived work in which the creator of the imagery has rights (not that I necessarily agree with that). The requirement has always been that we have had permission to trace at the point in time that the tracing happened (forgetting about special cases like NearMap) . The argument of the proponents that IP exists at all in ways and similar objects has been that the tracing (regardless of source) was an expression of creativity and that that expresses itself in, among other properties, the placement of nodes where it is found aesthetically pleasing. So why is one a derived work and the other not? Simon ___ legal-talk mailing list legal-talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/legal-talk
Re: [OSM-legal-talk] feedback requested
Am 21.12.2011 14:50, schrieb Ed Avis: Simon Poolesimon@... writes: In general we have assumed that for example tracing from aerial imagery and similar sources does not create a derived work in which the creator of the imagery has rights (not that I necessarily agree with that). The requirement has always been that we have had permission to trace at the point in time that the tracing happened Right - we require permission. So for example tracing from Google Maps is not allowed, even if the legal theory about not creating a derived work turns out to be correct. I contend that mappers' contributions would need to be treated no different to any other external data source. If we have permission, we can use them, if not, we can't. If one mapper illegitimately adjusted the position of a way by using Google Earth as a backdrop, but then a second mapper moved the position of the nodes some more, normal OSM practice would still be to delete the tainted data. So you contend that there was no permission to use positional information entered in the DB by other mappers to interpolate prior to the current CTs (obviously this is not covered by CC-by-SA 2.0)? Simon ___ legal-talk mailing list legal-talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/legal-talk
Re: [OSM-legal-talk] feedback requested
Quoting Ed Avis e...@waniasset.com: Simon Poole simon@... writes: - mapper A (who has agreed to the CTs) creates a way - mapper B (who has not agreed) adjusts the way's geometry, creating some new nodes - mapper C (who has agreed) adjusts the position of those nodes In this case the third edit would have to be reverted IMHO no, if we assume that C is editing in good faith and actually improving the geometry I would say that C is mapping in good faith and improving the existing geometry by building on the previous work done by B as well as his or her own surveying. I've often walked around with a GPS, then come back and adjusted road positions. I would not claim that the resulting position was all my own work, particuarly as I did not wholesale replace the layout with my GPS track but rather tried to find a middle path that lay between the existing geometry and the GPS. Think about it another way - if mapper B had never existed, would the final result be the same? If mapper C is wholly replacing the geometry then the answer is yes, and the new layout is entirely C's work. If the answer is no, then C is contributing useful information but relying on B's earlier work. I contend that the second case is more common. Certainly if you started mapping and decided to totally ignore the existing layout of a way and replace it with your GPS trace or aerial tracing, you would be criticized by most of your fellow mappers. It is more normal to make small adjustments, combining both the new survey information and what went before. (If you did decide to throw away the earlier work and start from scratch, you might well just delete the existing object and make a new one.) In general we have assumed that for example tracing from aerial imagery and similar sources does not create a derived work in which the creator of the imagery has rights (not that I necessarily agree with that). The requirement has always been that we have had permission to trace at the point in time that the tracing happened Right - we require permission. So for example tracing from Google Maps is not allowed, even if the legal theory about not creating a derived work turns out to be correct. I contend that mappers' contributions would need to be treated no different to any other external data source. If we have permission, we can use them, if not, we can't. If one mapper illegitimately adjusted the position of a way by using Google Earth as a backdrop, but then a second mapper moved the position of the nodes some more, normal OSM practice would still be to delete the tainted data. If there is to be an exception (which does seem to me like one rule for ordinary mappers, another rule for the OSMF) then I think the onus is on those proposing it to do the legal research making sure it is safe. Of course it is far more tempting to wave hands and pick whatever policy helps to get the whole business over with, but that is not a sound way to make legal decisions. With respect to the legal fraternity, I think we are the experts here as this thread is showing from all parties ... collectively making a set of highly precise technical decisions based on general legally-enshrined principles. What we should be considering in making our final decisions is 1) risk - If we do/don't do something, what is the future risk to the project? and 2) good faith - are we making a reasonable effort to remove the IP of folks who have not given us permission to continue? I certainly agree with Ed that we should treat ex-contributors no differently to any IP owner ... but feel we are already doing that in this and other conversations. I would also point out that Safe Harbour provisions also apply, if an ex-contributor later points out remaining instances of IP, then the OSMF will take steps to remove them. We naturally want that to be negligible or zero but it is an extra safe-guard. Mike LWG ___ legal-talk mailing list legal-talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/legal-talk
Re: [OSM-legal-talk] feedback requested
Sorry, I appreciate your taking the time to go through the arguments on this but I think I have said all I have to say about node positions. I'll let others decide whether what I wrote makes sense. -- Ed Avis e...@waniasset.com ___ legal-talk mailing list legal-talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/legal-talk
Re: [OSM-legal-talk] feedback requested
mike@... writes:
2) good faith - are we making a reasonable effort to remove the IP of
folks who have not given us permission to continue? I certainly agree
with Ed that we should treat ex-contributors no differently to any IP
owner ... but feel we are already doing that in this and other
conversations.
Mike, in that case I would ask you to apply the 'Google Maps test'. If some
map data were entered by copying from a third party who did not give permission,
but then edited in good faith, how much of the data needs to be unpicked?
In the past OSMF has taken a very cautious approach to this, which I believe
is the right one.
If after careful consideration you do formulate a policy ('the LWG declares that
creating a node is not a creative operation, so it can be kept as long as it has
been moved by somebody else afterwards', or whatever you decide), then it should
also be applied to such third-party-copyright situations going forward.
Originally it was promised that no big deletion would go ahead if it would cause
too much damage to the OSM data. Is that still the case and if so who is tasked
with deciding whether to pull the switch?
--
Ed Avis e...@waniasset.com
___
legal-talk mailing list
legal-talk@openstreetmap.org
http://lists.openstreetmap.org/listinfo/legal-talk
Re: [OSM-legal-talk] feedback requested
Quoting Ed Avis e...@waniasset.com:
mike@... writes:
2) good faith - are we making a reasonable effort to remove the IP of
folks who have not given us permission to continue? I certainly agree
with Ed that we should treat ex-contributors no differently to any IP
owner ... but feel we are already doing that in this and other
conversations.
Mike, in that case I would ask you to apply the 'Google Maps test'. If some
map data were entered by copying from a third party who did not give
permission,
but then edited in good faith, how much of the data needs to be unpicked?
In the past OSMF has taken a very cautious approach to this, which I believe
is the right one.
Ed,
Yes, that certainly seems reasonable to me though I would bow to the
more technically clued up, such as the Data Working Group. I would
speculate that it is an issue that has not come up, where there has
not been any significant subsequent edit activity the simplest and
most effective use of everyone's time is a simpler revert.
If after careful consideration you do formulate a policy ('the LWG
declares that
creating a node is not a creative operation, so it can be kept as
long as it has
been moved by somebody else afterwards', or whatever you decide),
then it should
also be applied to such third-party-copyright situations going forward.
Again, very reasonable to me going forward.
Originally it was promised that no big deletion would go ahead if it
would cause
too much damage to the OSM data. Is that still the case and if so
who is tasked
with deciding whether to pull the switch?
Community assent ... sounds vague to some may be, but has worked well
so far. Someone gave a good assessment in this or the Editing of
Content thread, sorry I cannot access it at the moment. By a number
of measures, we are in the range of 95% of data good to go, so I'd
personally say we are already at the no big deletion stage. We can
still increase that though, so we should. We've also said that we also
want to take local hotspots into consideration ... the UK, Germany and
Spain have a lot of red spots for example. The LWG's main task at the
moment is to get more undecided and non-responders on board and to
facilitate a small number of contributors who can say yes to some but
not all their contributions. Any chance of you changing your decline
now, that is the easiest way of decreasing deletions?
Mike
___
legal-talk mailing list
legal-talk@openstreetmap.org
http://lists.openstreetmap.org/listinfo/legal-talk
Re: [OSM-legal-talk] feedback requested
Quoting Ed Avis e...@waniasset.com:
mike@... writes:
2) good faith - are we making a reasonable effort to remove the IP of
folks who have not given us permission to continue? I certainly agree
with Ed that we should treat ex-contributors no differently to any IP
owner ... but feel we are already doing that in this and other
conversations.
Mike, in that case I would ask you to apply the 'Google Maps test'. If some
map data were entered by copying from a third party who did not give
permission,
but then edited in good faith, how much of the data needs to be unpicked?
In the past OSMF has taken a very cautious approach to this, which I believe
is the right one.
Ed,
Yes, that certainly seems reasonable to me though I would bow to the
more technically clued up, such as the Data Working Group. I would
speculate that it is an issue that has not come up, where there has
not been any significant subsequent edit activity the simplest and
most effective use of everyone's time is a simpler revert.
If after careful consideration you do formulate a policy ('the LWG
declares that
creating a node is not a creative operation, so it can be kept as
long as it has
been moved by somebody else afterwards', or whatever you decide),
then it should
also be applied to such third-party-copyright situations going forward.
Again, very reasonable to me going forward.
Originally it was promised that no big deletion would go ahead if it
would cause
too much damage to the OSM data. Is that still the case and if so
who is tasked
with deciding whether to pull the switch?
Community assent ... sounds vague to some may be, but has worked well
so far. Someone gave a good assessment in this or the Editing of
Content thread, sorry I cannot access it at the moment. By a number
of measures, we are in the range of 95% of data good to go, so I'd
personally say we are already at the no big deletion stage. We can
still increase that though, so we should. We've also said that we also
want to take local hotspots into consideration ... the UK, Germany and
Spain have a lot of red spots for example. The LWG's main task at the
moment is to get more undecided and non-responders on board and to
facilitate a small number of contributors who can say yes to some but
not all their contributions. Any chance of you changing your decline
now, that is the easiest way of decreasing deletions?
Mike
___
legal-talk mailing list
legal-talk@openstreetmap.org
http://lists.openstreetmap.org/listinfo/legal-talk
[OSM-legal-talk] Apologies for misleading munin graphs
Hi, as you probably know I'm running statistics on the raw count of objects processed by the OSMI view and making Munin graphs of them here: http://tools.geofabrik.de/osmi/munin.html I'm afraid that there has been an error in some of the graphs (example graph with problem shown here http://www.remote.org/frederik/tmp/odbl_ways-month.png) where the Y axis did not start at 0, giving the impression that the number of problematic was smaller than it in fact is. Especially the way graph looked as if, if the trend continues, all problematic ways would be eliminated by January which was a bit over-optimistic! I've fixed the configuration and the graphs are less euphemistic now. They are meant to inform, not to manipulate. Bye Frederik -- Frederik Ramm ## eMail frede...@remote.org ## N49°00'09 E008°23'33 ___ legal-talk mailing list legal-talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/legal-talk
