Re: [OSM-legal-talk] share-alike on generalized data?
On 06/02/16 11:32 AM, Tobias Wendorff wrote: > > Sure, I understand that. But I thought the main concept behind share-alike > is to make data better by foreign "investitions". In general the idea of share-alike is to make sure that downstream users of data have the same ability to work with the data as upstream users. So it's about the users continuing to be able to use the data rather than improving it, although that may be a side effect in some cases. - Rob. ___ legal-talk mailing list legal-talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/legal-talk
Re: [OSM-legal-talk] share-alike on generalized data?
First of all: sorry, my mail client had problems while parsing the mail. That's what the message-id got lost. I hope that won't happen again :( > The relevant question here is if during this process you generate a > derivative database containing both OSM and proprietary data. If you > do you'd need to share this database. I'll contact the programmer of the generalization tool, but since input files are seperate shapefiles, there isn't a mixture of data from me as a data provider. I can't tell you, if the tool will create a mixed database ... the resulting files will be stored in ONE file, like a PDF or AI file or anything else. So that's bad ... or even worse. > From my perspective this is not relevant - first because usefulness is a > subjective assessment and second because the ODbL share-alike > provisions are not about usefulness, they are about giving something > back in case the ODbL data is useful for you in connection with other > data. Sure, I understand that. But I thought the main concept behind share-alike is to make data better by foreign "investitions". > But in your specific use case i see no real problem. If you move the > OSM street data to match your other information you can easily make > available the modified street data. If you do it the other way round > modifying other data using OSM data things are more tricky. But you > can try to avoid that. I think, it's vice-versa. Since some elements will also move (push, pull, snap), f.e. landuse polygons, they will interact with OSM data. So, according to ODbL, everything which is interacting with OSM data needs to be released nevertheless it's in the same database or now. That's something, the employer didn't take into account and it might break the use he had in idea. We are about using OpenStreetMap data on an area of about 5000 square kilometer. Best regards, Tobias ___ legal-talk mailing list legal-talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/legal-talk
Re: [OSM-legal-talk] share-alike on generalized data?
On Saturday 06 February 2016, Tobias Wendorff wrote: > > Since this process includes an (automatic processed) interaction > between foreign and OpenStreetMap data, share-alike might step in. The relevant question here is if during this process you generate a derivative database containing both OSM and proprietary data. If you do you'd need to share this database. > Since the process of generalization doesn't make the quality of > OpenStreetMap data any better (f.e. higher position error), the > results won't be useful at all or could even have a bad effect, > if they get imported back into OpenStreetMap. From my perspective this is not relevant - first because usefulness is a subjective assessment and second because the ODbL share-alike provisions are not about usefulness, they are about giving something back in case the ODbL data is useful for you in connection with other data. But in your specific use case i see no real problem. If you move the OSM street data to match your other information you can easily make available the modified street data. If you do it the other way round modifying other data using OSM data things are more tricky. But you can try to avoid that. -- Christoph Hormann http://www.imagico.de/ ___ legal-talk mailing list legal-talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/legal-talk
Re: [OSM-legal-talk] Use of data from the EU GMES/Copernicus programme
Since there do not seem to be any more objections or comments on the matter i am going to add this to the contributors page so people can use the data where it seems useful. I have no specific use cases at the moment but over time it will probably be of interest to add some Sentinel-2 imagery to the osmim (http://maps.imagico.de/#osmim). -- Christoph Hormann http://www.imagico.de/ ___ legal-talk mailing list legal-talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/legal-talk
Re: [OSM-legal-talk] share-alike on generalized data?
Am So, 7.02.2016, 00:53 schrieb Rob Myers: > In general the idea of share-alike is to make sure that downstream users > of data have the same ability to work with the data as upstream users. > > So it's about the users continuing to be able to use the data rather > than improving it, although that may be a side effect in some cases. The problem of the project is that data from the local mapping agencies can't be released under share-alike. Of cause, the project would release all the changed made to OSM street data, but it wouldn't be possible to release the laduse data. It was planned to release tools and information about differences between cadastral land register and OpenStreetMap in terms of completeness and other dimensions. I'm sad that this new project can't be realised because of the strict share-alike. I think, it's really impossible to convince the LMA to release the generalized data under ODbL and I don't think it's enough to describe the way and tools to do the generalization, is it? I mean, this won't be enough, will it? - get OSM data extract from 2016-02-07 - filter streets - get LMA data extract from 2015-12-31 - open in generalization tool XY with parameters XY Since the generalization tool isn't free, nearly no user could take advantage of this. ___ legal-talk mailing list legal-talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/legal-talk
Re: [OSM-legal-talk] share-alike on generalized data?
On 2/6/2016 9:41 PM, Tobias Wendorff wrote: I mean, this won't be enough, will it? - get OSM data extract from 2016-02-07 - filter streets - get LMA data extract from 2015-12-31 - open in generalization tool XY with parameters XY When publicly using a derivative database (or produced work from one) you need to share either the derivative database, alterations to the original database that form the derivative database, or contents plus method to perform the alterations. The tool isn't really an issue here, as you've indicated the external data isn't available under an open license. It sounds like you have two databases, one the landuse data, the other the roads data, together which form a collective database. This might be in one file, or more than one file. Because both databases are derived from OSM data, by design you have to release them (or another option under s 4.6). Alternately, you might have one database with both landuse and roads data, which might be in one file or more than one file, but the results are the same. What you don't have to do is release the original proprietary database from the LMA, but it would be able to be partially reconstructed from what you do release. Again, this is by design. ___ legal-talk mailing list legal-talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/legal-talk
[OSM-legal-talk] share-alike on generalized data?
Desr members of Legal-Talk! I'm working on a project, which mixes feature classes of different sources: buildings, landuse and other data from a local mapping agency and streets streets from OpenStreetMap Since the produced work will be published, share-alike needs to be discussed. At lower scales, the resulting map should be generalized. While the licenses of the mapping agency don't have a share-alike licens, ODbL does of course. At some places, OpenStreetMap objects might get moved or pushed to prevent cartographic collisions with the other objects. Also, landuse areas could be snapped to OSM streets to fill the wholes. Since this process includes an (automatic processed) interaction between foreign and OpenStreetMap data, share-alike might step in. Since the process of generalization doesn't make the quality of OpenStreetMap data any better (f.e. higher position error), the results won't be useful at all or could even have a bad effect, if they get imported back into OpenStreetMap. According to ODbL, I'd need to publish the changed streets, a changeset (or equal) or a description, how to change the data - but only upon a request. Is this useful for generalized data at all? Best regards, Tobias ___ legal-talk mailing list legal-talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/legal-talk