Re: [talk-au] How many NearMap users do you think have accepted the new CTs and ODbL?
On Mon, 2 May 2011 12:40:55 +0800 Andrew Gregory wrote: > Yes, it's all based on surveys where I've gone there in person. (How > else would I get the name?) I have been to towns already traced from Nearmap and simply got the road names. Not been up and down every street so they are tagged source=nearmap source:name=survey ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
Re: [talk-au] How many NearMap users do you think have accepted the new CTs and ODbL?
I'm curious about the legal position on tracing as a derived work (probably different in different countries). Perhaps Microsoft saying it does not consider tracing to be a derived work does not stem from their largess but their reading of the legal position. Amongst artist friends I've heard discussions about copies being considered autonomous artworks as long as they are not exact duplicates. I believe it stems from the legal protection offered to art works which parody or comment on other art works, collages, as well as art students creating their own copies of art works in the process of studying. Some artists have even created meticulous copies of famous paintings but at different scales and in different contexts as a way of comment. Did Andy Warhol breach copyright when he turned photos into screen print images? A hand traced copy of a feature produced from a photograph may well be sufficiently distinct to not be considered a derived work. Any copyright lawyers in the house? Gary On Mon, 2011-05-02 at 17:30 +1000, Ben Kelley wrote: > I'm not sure the source tag is enouh to identify a derived work. > > If person A adds a way by tracing (for example) Nearmap data which gets > rendered as a map, and person B adds a nearby street based on what they saw > on the map plus what they know of the area (source=local_knowledge), all are > derived from the original edit. (esp in the context of cc-by-sa) > > Removing everything with source=nearmap doesn't solve this. > > - Ben Kelley > > Sent from my HTC > > -Original Message- > From: David Murn > Sent: Monday, 2 May 2011 10:17 > To: Andrew Gregory > Cc: OSM Australian Talk List > Subject: Re: [talk-au] How many NearMap users do you think have accepted the > new CTs and ODbL? > > On Sun, 2011-05-01 at 23:18 +0800, Andrew Gregory wrote: > > > > In any case, I expect that when it comes time to actually apply the > > new license, any source=nearmap data will disappear leaving behind all > > my re-licensable data. > > That is what one would hope, but no-one has been able to give a straight > answer. The problem with this, is how many source= tags do they have to > check for and remove? The problem isnt specific to nearmap, it is a > general problem for all data derived from sources using differing > licences (for example, ABS, yahoo or data.gov.au, just in Australia). > It is easier to simply remove every edit from a user than for them to > automate the process of figuring out what was sourced from where. > > > ___ > Talk-au mailing list > Talk-au@openstreetmap.org > http://lists.openstreetmap.org/listinfo/talk-au ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
Re: [talk-au] How many NearMap users do you think have accepted the new CTs and ODbL?
I'm not sure the source tag is enouh to identify a derived work. If person A adds a way by tracing (for example) Nearmap data which gets rendered as a map, and person B adds a nearby street based on what they saw on the map plus what they know of the area (source=local_knowledge), all are derived from the original edit. (esp in the context of cc-by-sa) Removing everything with source=nearmap doesn't solve this. - Ben Kelley Sent from my HTC -Original Message- From: David Murn Sent: Monday, 2 May 2011 10:17 To: Andrew Gregory Cc: OSM Australian Talk List Subject: Re: [talk-au] How many NearMap users do you think have accepted the new CTs and ODbL? On Sun, 2011-05-01 at 23:18 +0800, Andrew Gregory wrote: > In any case, I expect that when it comes time to actually apply the > new license, any source=nearmap data will disappear leaving behind all > my re-licensable data. That is what one would hope, but no-one has been able to give a straight answer. The problem with this, is how many source= tags do they have to check for and remove? The problem isnt specific to nearmap, it is a general problem for all data derived from sources using differing licences (for example, ABS, yahoo or data.gov.au, just in Australia). It is easier to simply remove every edit from a user than for them to automate the process of figuring out what was sourced from where. ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
Re: [talk-au] How many NearMap users do you think have accepted the new CTs and ODbL?
On Mon, 2011-05-02 at 12:40 +0800, Andrew Gregory wrote: > > Unless youve realigned the ways based on GPS tracks after you traced > > from nearmap, tagging the ways as source=survey is incorrect. You > > can add a source:name=survey or similar, but if youve traced from a > > source, just because you verify it with another source if you havent > > made any changes Id suggest leaving source tag as is. > > > Yes, it's all based on surveys where I've gone there in person. (How > else would I get the name?) As I said, when you get the name, you should use source:name=survey and leave the source=nearmap tag in-place unless after you survey you not only enter the name but also realign all the nearmap-sourced nodes to your GPS trace. > I just can't see that happening. The damage to the map would be too > big! In any case, how do you select the people whose data is to be > deleted? The same list of unacceptable sources that is too hard to > determine in the first place? Whatever criteria that would be required > to identify users could just as easily be applied to ways and nodes, > in a much more targeted and far less damaging way. Youve basically summarised the whole problem here. The damage to the map is significant (figures range from between 50-80% loss of data in Australia). The method being used to select the data to delete is to ask users to allow OSM to relicence their contributions. Anyone who doesnt agree, has their data deleted. This also affects any revisions made to existing data by users who HAVE agreed. The 'unacceptable sources' isnt so much a pre-determined list, in general in Australia it is any data that is released under CC-BY or CC-BY-SA, which will soon be incompatible with the new licence that OSMF has drafted. One of the problems is that its not easy to determine which users are affected. Some data is obviously tagged as being sourced from somewhere, in these cases its easy to know if the data can remain or not under the new licence. But in Australia, a lot of users would have made edits (even minor edits) using CC-BY sources, such as the ABS data or simply using nearmap for a quick live edit on the OSM website, moving a toilet to the correct location without adding a source tag, for example. > Well source=nearmap is easily identified by an automated process. It > worked for you! As for ones I may have missed, well I will need to be > trusted that I haven't missed any, in exactly the same way the other > 3390-536=2854 users will have to be trusted that they've never used > nearmap. My simple test was simply to demonstrate that a minimum of 25% of those who agreed are unable to. That figure might be higher, but it can be guaranteed that its not lower. The figures of 3390, 536 and 2854 represent total number of Australian mappers, total that have accepted and total that havent. This means that 2854 users' data wont be included in the 'new' OSM as OSM cannot relicence the data from the contributor. David ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
Re: [talk-au] How many NearMap users do you think have accepted the new CTs and ODbL?
On 2 May 2011 08:17, David Murn wrote: > On Sun, 2011-05-01 at 23:18 +0800, Andrew Gregory wrote: > > > Thankfully, I've been careful to use source=nearmap. I've also been > > making a point to go around and survey streets I've traced, check > > their alignment, name them and set source=survey. > > Unless youve realigned the ways based on GPS tracks after you traced > from nearmap, tagging the ways as source=survey is incorrect. You can > add a source:name=survey or similar, but if youve traced from a source, > just because you verify it with another source if you havent made any > changes Id suggest leaving source tag as is. > Yes, it's all based on surveys where I've gone there in person. (How else would I get the name?) > > In any case, I expect that when it comes time to actually apply the > > new license, any source=nearmap data will disappear leaving behind all > > my re-licensable data. > > That is what one would hope, but no-one has been able to give a straight > answer. The problem with this, is how many source= tags do they have to > check for and remove? The problem isnt specific to nearmap, it is a > general problem for all data derived from sources using differing > licences (for example, ABS, yahoo or data.gov.au, just in Australia). > It is easier to simply remove every edit from a user than for them to > automate the process of figuring out what was sourced from where. > I just can't see that happening. The damage to the map would be too big! In any case, how do you select the people whose data is to be deleted? The same list of unacceptable sources that is too hard to determine in the first place? Whatever criteria that would be required to identify users could just as easily be applied to ways and nodes, in a much more targeted and far less damaging way. > > I short, I don't see any problems. All my current data conforms to the > > current license, and the data that doesn't conform to the new license > > is easily identifiable and removable. > > Is it easily identifiable by you or by an automated process also? Have > you tagged every single edit youve made, when sourcing nearmap, with > their source? I know personally Im sure theres been times when Ive made > a quick edit in potlatch and not thought about changing the source tag. > Well source=nearmap is easily identified by an automated process. It worked for you! As for ones I may have missed, well I will need to be trusted that I haven't missed any, in exactly the same way the other 3390-536=2854 users will have to be trusted that they've never used nearmap. If you don't trust me, then where do you stop? At a blank map? Perhaps I'm *more* trustworthy since I've made the effort to source my input? Maybe all un-sourced data needs to be deleted as well? How much of that is there? Seriously, where do you stop? IMHO, a list of unacceptable sources will *have* to be determined and applied to the map data. That will be the best that can be done. In fact, I don't see any other way. > Hopefully some general visualisation tools will be developed well > > before the license change takes place. > > Again, that is what one would hope, but as no-one is quite sure what > will be affected or how. Part of the problem also is that depending on > when you agreed to the new licence and CTs, they have quite possibly > changed since then, meaning that any visualisation of your data that is > impacted when you accepted it, would possibly look different now, if the > new wording became more compliant with sources you might have used. > > I think it's understood that any visualisation tools will have to keep up with policy changes. -- Andrew ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
Re: [talk-au] How many NearMap users do you think have accepted the new CTs and ODbL?
On Sun, 2011-05-01 at 23:18 +0800, Andrew Gregory wrote: > Thankfully, I've been careful to use source=nearmap. I've also been > making a point to go around and survey streets I've traced, check > their alignment, name them and set source=survey. Unless youve realigned the ways based on GPS tracks after you traced from nearmap, tagging the ways as source=survey is incorrect. You can add a source:name=survey or similar, but if youve traced from a source, just because you verify it with another source if you havent made any changes Id suggest leaving source tag as is. > In any case, I expect that when it comes time to actually apply the > new license, any source=nearmap data will disappear leaving behind all > my re-licensable data. That is what one would hope, but no-one has been able to give a straight answer. The problem with this, is how many source= tags do they have to check for and remove? The problem isnt specific to nearmap, it is a general problem for all data derived from sources using differing licences (for example, ABS, yahoo or data.gov.au, just in Australia). It is easier to simply remove every edit from a user than for them to automate the process of figuring out what was sourced from where. > I short, I don't see any problems. All my current data conforms to the > current license, and the data that doesn't conform to the new license > is easily identifiable and removable. Is it easily identifiable by you or by an automated process also? Have you tagged every single edit youve made, when sourcing nearmap, with their source? I know personally Im sure theres been times when Ive made a quick edit in potlatch and not thought about changing the source tag. > Hopefully some general visualisation tools will be developed well > before the license change takes place. Again, that is what one would hope, but as no-one is quite sure what will be affected or how. Part of the problem also is that depending on when you agreed to the new licence and CTs, they have quite possibly changed since then, meaning that any visualisation of your data that is impacted when you accepted it, would possibly look different now, if the new wording became more compliant with sources you might have used. David ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
Re: [talk-au] How many NearMap users do you think have accepted the new CTs and ODbL?
On 26 April 2011 21:24, David Murn wrote: > > How many OSM users have accepted the new terms, without fully > understanding that sources they have used in the past prohibit them from > doing so. > > Using my australian test extract from 21/03/2011, I found that 3390 > users have made edits in the area of interest (the Australian extract > available on osmaustralia.org). > > Of these 3390 users, 536 have used the tag source=nearmap at least once. > > Of these 536 users, 134 have agreed to the ODbL+CTs. > I'm almost certainly one of those 134. I must admit I didn't read the new license too closely, but nothing seemed out of order. In the back of my mind I was probably treating it like all the other various licenses I've agreed to in the past - as something that applies from the moment I agree to it. In hindsight, it's quite obvious that the new agreement has to be retrospective. Thankfully, I've been careful to use source=nearmap. I've also been making a point to go around and survey streets I've traced, check their alignment, name them and set source=survey. However, there have been times when I have traced buildings. That's much more difficult to survey. In any case, I expect that when it comes time to actually apply the new license, any source=nearmap data will disappear leaving behind all my re-licensable data. I short, I don't see any problems. All my current data conforms to the current license, and the data that doesn't conform to the new license is easily identifiable and removable. I would, of course, like to see what data is incompatible with the new license so that I can more focus my surveying efforts. Hopefully some general visualisation tools will be developed well before the license change takes place. -- Andrew ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
Re: [talk-au] How many NearMap users do you think have accepted the new CTs and ODbL?
On 27 April 2011 04:15, David Groom wrote: > But I thought that Nearmap has said that they did not think the CT's were > compatible with the use of their data. As I understood it this had nothing > to do with CC-BY-SA or ODbL. > > So the issue as I understand it is the CT's, and so anyone who has used > Nearmap as a source and has agreed to the CT's is in violation of both > Nearmap's licence, and the CT's. > > Of course my understanding of Nearmap's position may be wrong, and I suspect > they (Ben?) will be able to clarify matters. That's my understanding as well, even though some have suggested once you agree you can't unagree, even though there is clear breaches of contract with OSM-F. ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
Re: [talk-au] How many NearMap users do you think have accepted the new CTs and ODbL?
On Tue, 26 Apr 2011 23:24:09 +1000 David Murn wrote: > Using my australian test extract from 21/03/2011, I found that 3390 > users have made edits in the area of interest (the Australian extract > available on osmaustralia.org). > > Of these 3390 users, 536 have used the tag source=nearmap at least > once. > > Of these 536 users, 134 have agreed to the ODbL+CTs. In my recent foray into Victoria, I found spots which must have been mapped from Nearmap, judging from the quality of the mapping and the lack of street names or POIs. I haven't done any check to see if those mappers have attributed Nearmap on a changeset or otherwise. I believe 536 mappers is a minimum who have used Nearmap. And if I take 134 as the numerator, and 3390 as the denominator, then I get 4%. This represents a large community who have decided that they are staying CC-by-SA. Some of those mappers aren't local and don't count - like stae**er who traced parts of remote Australia from Google, admitted it and still hasn't had any attention to his edits from the DWG, although I pointed out that he had edited over the whole world from his armchair, and the source of those was likely to be Google as well. Rosscoe cleaned up Crystal Brook, I cleaned up Marree, and Halls Creek remains polluted. ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
Re: [talk-au] How many NearMap users do you think have accepted the new CTs and ODbL?
- Original Message - From: "Richard Weait" To: "OSM Australian Talk List" Sent: Tuesday, April 26, 2011 3:50 PM Subject: Re: [talk-au] How many NearMap users do you think have accepted the new CTs and ODbL? On Tue, Apr 26, 2011 at 9:24 AM, David Murn wrote: I was wondering this question tonight. [ ... ] This only shows where there is clear evidence of licence violations without having to look past the data's tags. [ ... ] Food for thought "Junk food" at best. ;-) OSM is currently published under CC-By-SA. There is no current "violation", even if what you suggest would be considered a "violation". We already know that some contributors have But I thought that Nearmap has said that they did not think the CT's were compatible with the use of their data. As I understood it this had nothing to do with CC-BY-SA or ODbL. So the issue as I understand it is the CT's, and so anyone who has used Nearmap as a source and has agreed to the CT's is in violation of both Nearmap's licence, and the CT's. Of course my understanding of Nearmap's position may be wrong, and I suspect they (Ben?) will be able to clarify matters. Regards David changesets that they wish to have removed before ODbL publication. That can't be done until Phase 4. ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
Re: [talk-au] How many NearMap users do you think have accepted the new CTs and ODbL?
On Tue, 2011-04-26 at 10:50 -0400, Richard Weait wrote: > On Tue, Apr 26, 2011 at 9:24 AM, David Murn wrote: > > > > I was wondering this question tonight. > > [ ... ] > > > This only shows where there is clear evidence of licence violations > > without having to look past the data's tags. > > [ ... ] > > > Food for thought > > "Junk food" at best. ;-) OSM is currently published under CC-By-SA. > There is no current "violation", even if what you suggest would be > considered a "violation". We already know that some contributors have > changesets that they wish to have removed before ODbL publication. > That can't be done until Phase 4. Good point, lets wait until the end of Phase 5 before we start thinking about users who have accepted the terms when they shouldnt have. You refer to 'some' contributors, I refer to over 100. You completely missed my point about the fact that some users (upto 25%) have accepted the terms which they cannot comply with. Whether that is a violation at this point in time, is irrelevant. Whether anyone bothers to look into the problem before relicencing data they dont own, IS relevant, also the fact that people will continue to agree to the terms, having used data which cannot be relicenced. Maybe a more serious note should be put there advising that if youve used a CC-BY-SA data source, you cannot accept the licence (without having to read through a dozen pages of legalese). As has been discussed here previously, many users of software have become accustomed to always accepting licence terms without reading them, as experience tells that declining an agreement means no further progress. This isnt a problem if someone breaks a licence by installing anti-virus on 2 computers, but it is a problem if someone breaks a licence by relicencing incompatible data into the OSM project, which is then used in-turn by millions. David ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
Re: [talk-au] How many NearMap users do you think have accepted the new CTs and ODbL?
On Tue, Apr 26, 2011 at 9:24 AM, David Murn wrote: > > I was wondering this question tonight. [ ... ] > This only shows where there is clear evidence of licence violations > without having to look past the data's tags. [ ... ] > Food for thought "Junk food" at best. ;-) OSM is currently published under CC-By-SA. There is no current "violation", even if what you suggest would be considered a "violation". We already know that some contributors have changesets that they wish to have removed before ODbL publication. That can't be done until Phase 4. ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
[talk-au] How many NearMap users do you think have accepted the new CTs and ODbL?
I was wondering this question tonight. How many OSM users have accepted the new terms, without fully understanding that sources they have used in the past prohibit them from doing so. So, I wrote a little script to find out and the numbers are surprising. Using my australian test extract from 21/03/2011, I found that 3390 users have made edits in the area of interest (the Australian extract available on osmaustralia.org). Of these 3390 users, 536 have used the tag source=nearmap at least once. Of these 536 users, 134 have agreed to the ODbL+CTs. So, approximately 25% of users who have attributed nearmap (and to be fair, a lot more data has probably used nearmap without proper attribution), have agreed to have their edits released under the new terms. Of 3390 total users in our region, 487 have agreed to the terms. Interesting numbers that show that a lot of users have been mislead or misunderstand the consequences of accepting the changes. If anyone is interested, I can provide the simple C code I used to generate these numbers and/or a list of usernames/uids that are involved. This only shows where there is clear evidence of licence violations without having to look past the data's tags. Im sure if similar figures were generated based on users in nearmap coverage areas, the numbers would be similar. These numbers were also generated from an extract, and not a full history dump, so there may well be cases where tags have changed and would not be counted here. Food for thought David ___ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
