Re: [Talk-us] press from SOTM US
I do hope to come to an agreement within OSM along the lines you just hashed out, Frederik (while not quite advocating for it): The data extracted by geocoding should just not lead to a substantial extract of the database, hence not producing a derivative database in the sense of the ODbL. I feel this would be within the spirit of why the ODbL was adopted (to encourage contribution) while clarifying an important use of OSM data that would create a huge incentive to improve data. Right now we largely don't have functioning municipal boundaries in OSM. Obviously, any data that is mixed into OSM data for _powering_ the geocoder would fall under share alike stipulations. You bring up the important problem of properly bounding the geocoding case. I'm thinking if all that can be extracted from OSM's database are names and addresses for lat/lon pairs or lat/lon pairs for names or addresses, it would be arguably impossible or at least impractically hard to recreate a functioning street network from it and the extracted data would be a narrow subset of OSM no matter how many locations are being geocoded. Thoughts? On Oct 20, 2012, at 1:38 PM, Frederik Ramm frede...@remote.org wrote: Hi, On 10/20/2012 09:59 AM, Toby Murray wrote: The discussion was about the fact that some companies are very afraid of share-alike licenses and it is preventing them from using our data to its fullest potential. There are several sides to this. Of course the share-alike license prevents companies from using our data to the same potential as a hypothetical PD counterpart (or a licensed-for-money competitor); excluding some kinds of use-without-sharing-back is the reason for a share-alike license and was desired by a large majority of the stakeholders. On the other hand, the license does not have to be feared, and some users might actually let their fear of share-alike shy away from some totally legal uses of OSM. There is some uncertainty about when exactly the share-alike clause is activated. One specific example that was mentioned: If you use OSM data to geocode a user's address, does the user database then have to be shared? No, but the database of locations, which might let others guess who your users are. That's apparently how the lawyers tend to read it but in my mind this would be silly. We have no use for a company's user database even if it were possible to release it without breaking every privacy law on the books. I agree that we have little use for that database of locations but I think that it is crystal clear this is a derived database. The only way to not require share-alike for that would be - as Richard has recently mentioned on legal-talk, where this discussion should be held -, to define any amount of geocoded locations to be insubstantial. However that would raise the question - could you not, by mass-geocoding every single address on every single street - re-create our whole street network? That could hardly be insubstantial then. Bye Frederik -- Frederik Ramm ## eMail frede...@remote.org ## N49°00'09 E008°23'33 ___ Talk-us mailing list Talk-us@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-us Alex Barth http://twitter.com/lxbarth tel (+1) 202 250 3633 ___ Talk-us mailing list Talk-us@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-us
Re: [Talk-us] press from SOTM US
On Oct 22, 2012, at 1:12 PM, Alex Barth wrote: The data extracted by geocoding should just not lead to a substantial extract of the database, hence not producing a derivative database in the sense of the ODbL. I feel this would be within the spirit of why the ODbL was adopted (to encourage contribution) while clarifying an important use of OSM data that would create a huge incentive to improve data. Right now we largely don't have functioning municipal boundaries in OSM. Obviously, any data that is mixed into OSM data for _powering_ the geocoder would fall under share alike stipulations. MySociety is working on derived municipal boundaries from OSM data: http://global.mapit.mysociety.org/ E.g.: http://global.mapit.mysociety.org/area/168844.html There's data in there, and code out there that you could build on. The MapIt service itself is non-commercial, but the code that drives it is freely-available. http://code.mapit.mysociety.org/ You bring up the important problem of properly bounding the geocoding case. I'm thinking if all that can be extracted from OSM's database are names and addresses for lat/lon pairs or lat/lon pairs for names or addresses, it would be arguably impossible or at least impractically hard to recreate a functioning street network from it and the extracted data would be a narrow subset of OSM no matter how many locations are being geocoded. Thoughts? This seems to match the spirit of the license as far as I understand it. -mike. michal migurski- m...@stamen.com 415.558.1610 ___ Talk-us mailing list Talk-us@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-us
Re: [Talk-us] press from SOTM US
Hi, On 22.10.2012 22:12, Alex Barth wrote: I do hope to come to an agreement within OSM along the lines you just hashed out, Frederik (while not quite advocating for it): This really ought to be discussed on legal-talk where there are many people with a year-long involvement into the finer details of the license - Cc+Followup there. Right now we largely don't have functioning municipal boundaries in OSM. Obviously, any data that is mixed into OSM data for _powering_ the geocoder would fall under share alike stipulations. I'm not sure about this obviously. I can imagine situations where someone collects geocoding queries and OSM's answers and perhaps even records which of the results the user clicked on afterwards, giving them a distinct advantage over other OSM users who don't have all that extra data. IIRC, geocoder.ca has proven that they can build a valuable geocoding database with such techniques. If we were to make a blanket declaration that geocoding doesn't trigger share-alike, we'd give that away, we'd allow people to build their own improved upon OSM geocoding databases and sell them on. If we allow it, then it *will* happen, because there's a commercial gain to be had. We would even open the door to services where someone geocodes with OSM and then says wrong result? just move the marker to the right position on this map, and keeps the corrections to himself, in a separate corrections database. I haven't thought this through enough to actually say which of the unwanted use cases are indeed possible even with the current substantial guidelines (http://wiki.openstreetmap.org/wiki/Open_Data_License/Substantial_-_Guideline) and which additional unwanted use cases would be possible with a weakened form of those. We should perhaps not only make a list of what people would like to do with geocoding, but a second list of what we don't want people to do (things like I sketched above - build improved database on top of OSM and market that), then we can maybe check any guidelines we draft against these points. You bring up the important problem of properly bounding the geocoding case. I'm thinking if all that can be extracted from OSM's database are names and addresses for lat/lon pairs or lat/lon pairs for names or addresses, it would be arguably impossible or at least impractically hard to recreate a functioning street network from it and the extracted data would be a narrow subset of OSM no matter how many locations are being geocoded. Thoughts? I'm not sure that a functioning street network is the bit that share-alike intends to protect and the rest is not: This whole discussion arose from the fact that there is heightened commercial interest in OSM-based geocoding - that there even seem to be people who are not interested in a functioning road network at all but who would be prepared to invest quite a bit of money to switch2osm their geocoding. So it seems that maybe address data is as valuable as the street network and should have the same level of protection? Bye Frederik -- Frederik Ramm ## eMail frede...@remote.org ## N49°00'09 E008°23'33 ___ Talk-us mailing list Talk-us@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-us
Re: [Talk-us] press from SOTM US
Thanks for kicking over to legal list. Responses inline. On Oct 22, 2012, at 6:34 PM, Frederik Ramm frede...@remote.org wrote: Hi, On 22.10.2012 22:12, Alex Barth wrote: I do hope to come to an agreement within OSM along the lines you just hashed out, Frederik (while not quite advocating for it): This really ought to be discussed on legal-talk where there are many people with a year-long involvement into the finer details of the license - Cc+Followup there. Right now we largely don't have functioning municipal boundaries in OSM. Obviously, any data that is mixed into OSM data for _powering_ the geocoder would fall under share alike stipulations. I'm not sure about this obviously. I can imagine situations where someone collects geocoding queries and OSM's answers and perhaps even records which of the results the user clicked on afterwards, giving them a distinct advantage over other OSM users who don't have all that extra data. IIRC, geocoder.ca has proven that they can build a valuable geocoding database with such techniques. If we were to make a blanket declaration that geocoding doesn't trigger share-alike, we'd give that away, we'd allow people to build their own improved upon OSM geocoding databases and sell them on. If we allow it, then it *will* happen, because there's a commercial gain to be had. I think a blanket declaration on geocoding isn't quite necessary. It's about clarifying what happens to the dataset that is being geocoded (a user database, a picture database, etc.). Say we clarified that geocoding a dataset with an OSM powered geocoder (e. g. Nominatim) does not extend the ODbL license to such a dataset. This clarification would not apply to the dataset that actually powered the geo coder. So if I went and gathered improvement suggestions of my users (move the marker to the right position on the map) and I added them into that OSM dataset that powers the geocoder, this OSM dataset would still constitute a derivative DB. More below... We would even open the door to services where someone geocodes with OSM and then says wrong result? just move the marker to the right position on this map, and keeps the corrections to himself, in a separate corrections database. I haven't thought this through enough to actually say which of the unwanted use cases are indeed possible even with the current substantial guidelines (http://wiki.openstreetmap.org/wiki/Open_Data_License/Substantial_-_Guideline) and which additional unwanted use cases would be possible with a weakened form of those. We should perhaps not only make a list of what people would like to do with geocoding, but a second list of what we don't want people to do (things like I sketched above - build improved database on top of OSM and market that), then we can maybe check any guidelines we draft against these points. You bring up the important problem of properly bounding the geocoding case. I'm thinking if all that can be extracted from OSM's database are names and addresses for lat/lon pairs or lat/lon pairs for names or addresses, it would be arguably impossible or at least impractically hard to recreate a functioning street network from it and the extracted data would be a narrow subset of OSM no matter how many locations are being geocoded. Thoughts? I'm not sure that a functioning street network is the bit that share-alike intends to protect and the rest is not: This whole discussion arose from the fact that there is heightened commercial interest in OSM-based geocoding - that there even seem to be people who are not interested in a functioning road network at all but who would be prepared to invest quite a bit of money to switch2osm their geocoding. So it seems that maybe address data is as valuable as the street network and should have the same level of protection? Fair point. Still - I would ask what is the purpose of this protection and how does it benefit OSM on this particular level? OSM clearly benefits of being used. The usage of OSM data in maps has been clarified, I believe the ability to unencumberedly leverage OSM data to create produced works is a huge benefit for OpenStreetMap as a whole as it creates more versatile map styles, (and yes, abilities to monetize them) and in turn have more map users and thousands of micro incentives of improving our common map. Important similar incentives are routing or geo coding. The latter is where I think the shoe starts to hurt. In my mind there's much to be gained by giving better incentives to contribute to OSM by clarifying the geocoding situation and little to be lost by allowing narrow extracts of OSM. I believe we can do this within the letter of the ODbL and within the spirit of why the ODbL was adopted. BTW, I don't want to know how many people out there have used Nominatim for geocoding without having any idea... Bye Frederik -- Frederik Ramm
Re: [Talk-us] press from SOTM US
On 10/19/2012 11:48 PM, Richard Welty wrote: we got some. Carl Frantzen of Talking Points Memo asked me about coming to SOTM US and i urged him to do so. he did and here we are: That was fantastic coverage; looking forward to the next 2 parts. I wonder how far the series will reach into the general population (the part generally unaware of OSM). ___ Talk-us mailing list Talk-us@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-us
Re: [Talk-us] press from SOTM US
From: Toby Murray [mailto:toby.mur...@gmail.com] Sent: Saturday, October 20, 2012 12:59 AM To: Talk Openstreetmap Subject: Re: [Talk-us] press from SOTM US On Fri, Oct 19, 2012 at 10:48 PM, Richard Welty rwe...@averillpark.net wrote: we got some. Carl Frantzen of Talking Points Memo asked me about coming to SOTM US and i urged him to do so. he did and here we are: http://idealab.talkingpointsmemo.com/2012/10/openstreetmap-part-1-new- cartographers.php Not quite sure what the little paragraph about move away from its open source roots. is all about. He kind of dropped that in there like a live hand grenade and then didn't say any more about it. Apparently this will be in parts 2 and 3. Having been at the session where I think this statement came from, I can assure everyone that there was absolutely no talk about moving away from an open license! :) Well, there's a difference between open source and open data. My understanding is that OSM is an open data project, but the OSM services run by OSMF are open source. There are non-open source tools for working with OSM data like ESRI's ArcGis OSM editor or Maperitive. The downside to closed-source software in a project like OSM is that you're a lot less likely to get help from the people in the community who have already faced similar challenges. The discussion was about the fact that some companies are very afraid of share-alike licenses and it is preventing them from using our data to its fullest potential. There is some uncertainty about when exactly the share-alike clause is activated. One specific example that was mentioned: If you use OSM data to geocode a user's address, does the user database then have to be shared? That's apparently how the lawyers tend to read it but in my mind this would be silly. We have no use for a company's user database even if it were possible to release it without breaking every privacy law on the books. Not even the OSMF shares its user table, so in my mind also this is a silly reading of the ODbL. I believe part of the confusion is because the case law around what is and isn't a derivative work is not well defined. There might be a scenario on which after distributing a user's exact location you also have to release the address corresponding to that location if requested... but if the data is accurate and the geocoder is any good, you can just do the reverse geocoding. ___ Talk-us mailing list Talk-us@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-us
Re: [Talk-us] press from SOTM US
Hi, On 10/20/2012 09:59 AM, Toby Murray wrote: The discussion was about the fact that some companies are very afraid of share-alike licenses and it is preventing them from using our data to its fullest potential. There are several sides to this. Of course the share-alike license prevents companies from using our data to the same potential as a hypothetical PD counterpart (or a licensed-for-money competitor); excluding some kinds of use-without-sharing-back is the reason for a share-alike license and was desired by a large majority of the stakeholders. On the other hand, the license does not have to be feared, and some users might actually let their fear of share-alike shy away from some totally legal uses of OSM. There is some uncertainty about when exactly the share-alike clause is activated. One specific example that was mentioned: If you use OSM data to geocode a user's address, does the user database then have to be shared? No, but the database of locations, which might let others guess who your users are. That's apparently how the lawyers tend to read it but in my mind this would be silly. We have no use for a company's user database even if it were possible to release it without breaking every privacy law on the books. I agree that we have little use for that database of locations but I think that it is crystal clear this is a derived database. The only way to not require share-alike for that would be - as Richard has recently mentioned on legal-talk, where this discussion should be held -, to define any amount of geocoded locations to be insubstantial. However that would raise the question - could you not, by mass-geocoding every single address on every single street - re-create our whole street network? That could hardly be insubstantial then. Bye Frederik -- Frederik Ramm ## eMail frede...@remote.org ## N49°00'09 E008°23'33 ___ Talk-us mailing list Talk-us@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-us
[Talk-us] press from SOTM US
we got some. Carl Frantzen of Talking Points Memo asked me about coming to SOTM US and i urged him to do so. he did and here we are: http://idealab.talkingpointsmemo.com/2012/10/openstreetmap-part-1-new-cartographers.php ___ Talk-us mailing list Talk-us@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-us
