Comparing Open Data TOU with ODC-by and ODC-PDDL
Locked
12
12
Comparing Open Data TOU with ODC-by and ODC-PDDL
 
|
All;
Below are links to some of Ottawa University - CIPPIC's latest work on the Canada Open Data City file with respect to licencing. The G4 + 1 Cities (Ottawa, Toronto, Vancouver and Edmonton + Montreal) will be discussing this file in the coming weeks. This report is the outcome of a request by the G4 + 1 who required a risk analysis to complement the first report submitted by CIPPIC who examined the City of Ottawa (aka the Vancouver) TOU.
CIPPIC's comparison of the City of Ottawa Terms of Use with ODC-By and ODC-PDDL - http://www.cippic.ca/uploads/open-licensing/Open_License_Comparison_Report-v2-10Feb2011.pdf
You will find other open-licencing reports here - http://www.cippic.ca/open-licensing/ These reports were prepared by David Fewer (http://www.commonlaw.uottawa.ca/en/david-fewer.html), oline Twiss, Articling Student and Kent Kent Mewhort.
Your comments are most welcome.
|
Re: Comparing Open Data TOU with ODC-by and ODC-PDDL
|
|
Reading through it I think it sums up the situation quite well. It was interesting to note that the City of Ottawa's license wasn't always the most protective of the rights of the City.
Personally I think that using a "standarised" CCby style license strikes the best balance between protecting the city's liability and the city's interests in getting its data used. People who want to use the data can understand it and since many applications basically use data from many cities it opens up these applications to the city's population. I note that the City releases data to Google under a Google license which is quite different. There it appears the usefulness of having Google provide a cycle map or Transit information such as the location of bus stops on Google maps is considered more important than the liability side. Cheerio John On 18 February 2011 10:09, Tracey P. Lauriault <[hidden email]> wrote: All; |
Re: Comparing Open Data TOU with ODC-by and ODC-PDDL
|
|
Thanks John;
Some are advocating for a ODC-PDDL there is a good discussion here - http://groups.google.com/group/opendatabc/browse_thread/thread/5a52335b3bc7fd3e
Cheers
t
On Fri, Feb 18, 2011 at 12:09 PM, john whelan <[hidden email]> wrote: Reading through it I think it sums up the situation quite well. It was interesting to note that the City of Ottawa's license wasn't always the most protective of the rights of the City. -- Tracey P. Lauriault 613-234-2805 |
Re: Comparing Open Data TOU with ODC-by and ODC-PDDL
|
|
I completely agree that a standardized licenses makes sense, though think that it shouldn't be of the CC variety. CC was created to deal with copyrighted material, not data. Yes, the license could (and sometimes does) get used for that purpose but I think governments have a responsibility not to use it. The main reason here is around education. Most citizens don't know that data cannot be copyrighted, Creative Commons has done a remarkable job setting up and educating the public on CC as an alternative to copyright. However, if we apply CC to data, it implies the data could have been copyrighted. This actually undermines the efforts of those of us (many who I believe are on this mailing list) who want to encourage a) a growth in the public's understanding of their legal rights around access to content. So having the government be complicit in this misrepresentation feels problematic. I think CC is a wonderful license and has done an enormous service to the effort to promote a free culture, but, from my perspective, it is the wrong tool for addressing data and facts. (In addition, requiring attribution is a problem as well - but check out my blog post and the databc mailing list for more on that). Cheers, dave On 11-02-18 9:22 AM, Tracey P. Lauriault wrote: Thanks John; |
Re: Comparing Open Data TOU with ODC-by and ODC-PDDL
|
|
OSM (OpenStreetMap) is going through some licensing issues of its own at the moment. OSM covers a wide range of people with different skill sets and interests. In Europe at the moment they don't have high quality data sources such as CANVEC but do have high densities of mappers on the ground and for many there is a social aspect to mapping. So one problem they have is finding things to map. Discouraging and even removing imports is high on the agenda for many.
The inner OSM committee has decided to move to CC-by-Odbl licensing rather than CC-by-SA which means that much of the current content is expected to disappear around June 2011. All imports will have to be verified to see if they meet the new rules. CANVEC does meet these new licensing requirements by the way. In Canada and Australia we have a different problem, more to map and fewer mappers. We also have more high quality data available to import than we can currently digest. End users typically want reliable data. The manual mapping method of mapping is fine but mappers equipment varies and often traces are made from satellite images. It's easy to do and you don't need to be on the ground to do them. Problem is the roads can be anything up to 100 meters from where they should be, aren't connected at junctions, street names aren't always correct, I came across at least 100 with the wrong name in Ottawa and the names aren't entered in a standard way. Routing and other software programs need streets joined up and named in standard ways. What has this got to do with licensing. Well Australia has a lot of government data that has been imported with CC-by-SA licensing. In Canada many would like to be able to send updates into CANVEC perhaps where we note a railroad track has been removed but we are unable to do so because of the way OSM is licensed. Places such as Canada and Australia where we have many fewer mappers and more ground to cover seem much more accepting of imports. As a result there is a real danger that OSM will split into OSM and CommonMap. CommonMap uses the same rich tools as OSM but more emphasis is placed on data imports such as house number information and CANVEC data and is licensed and will accept data licensed CC-by-SA. OpenStreetMap is not keen on making life easy for imports. So basically expect Canada, US, Australia and many other parts of the world to move to CommonMap, Europe will probably stay predominately OpenStreetMap. Quote from Frederik Ramm who is one of the European founders of OpenStreetMap and shows the difference in the European and my point of view. "We don't generally like data imports and we are keen on people discussing their imports and talking to the community about them before they commence. Too many people seem to think that any data is good data and just stuff it into OSM! So it is a desirable effect that things are not so easy with importing!" I note that the report does not seperate out CC-by-Odbl and CC-by-SA licensing. My person view is that CC-by is reasonable to both parties and widely accepted around the world and from a pragmatic point of view would work certainly for CommonMap. The current City license is too complex to understand easily but either of the other two standard approaches would work. Currently I can't import City data into OSM because of the license, however I note that I can tell that some city data has been imported. Not every one respects the licenses. I would very much like to be able to use the data for OSM or CommonMap in the future. Cheerio John On 18 February 2011 12:32, David Eaves <[hidden email]> wrote:
|
Re: Comparing Open Data TOU with ODC-by and ODC-PDDL
|
|
In reply to this post by David Eaves
I absolutely think cc-by (ideally sa) is the right license. I understand there are optics around "copyright," but cc-by adds enforceable links to data so people can easily "learn more," adds an inherent level of organization, and avoids abuse of data that doesn't require attribution. Past today's mash ups, there's a greater potential to enable a culture of constant learning (and reality checks), with flexible browser based semantic applications rather than one-use server-side mashups. With many integrated data sets, we will need something like CC REL. Open data as a movement needs more building on, less invented here, more signal and less noise (unorganized data), less acronyms that need specialist interpretation. CC introduces many people to a well developed alternative to copyright. We can (and should) count on Wikipedia and other efforts to be a significant part of integrated knowledge, so inline licenses are necessary regardless. David On 18 February 2011 12:32, David Eaves <[hidden email]> wrote:
|
Re: Comparing Open Data TOU with ODC-by and ODC-PDDL
|
|
I completely agree with this. We need to integrate data from multiple sources so standardized licensing helps here.
Cheerio John On 18 February 2011 17:20, David H. Mason <[hidden email]> wrote:
|
Re: Comparing Open Data TOU with ODC-by and ODC-PDDL
|
|
In reply to this post by David H. Mason
On Fri, Feb 18, 2011 at 5:20 PM, David H. Mason <[hidden email]> wrote:
I disagree with both the BY and SA recommendations. Everyone should really catch up on the discussion Tracey linked to http://groups.google.com/group/opendatabc/browse_thread/thread/5a52335b3bc7fd3e and on the CIPPIC papers at http://www.cippic.ca/open-licensing/ Rather than rehash what's been said, I'll selectively quote a few people.
In short: SA cuts out many commercial and non-commercial users who want to add new data to the licensed data but can't, either because the new data cannot be relicensed, or because the users want to retain certain rights over the data. BY introduces a technological hurdle in some use cases (e.g. integrating multiply BY data sources in a mobile app or in SMS or Twitter apps which have limited characters, merging multiple BY data sources and offering the data through an API, etc.) and it introduces uncertainty ("have I properly attributed the source?"). Here's quotes developing these ideas further:
SA is a bad idea with respect to open data. Under the ODbL, which is an SA license, Paul Ramsey writes: "if you use the data and add to it, you have to share your additions under the same license. So, once open, always open, no matter who takes it. It lowers the likelihood of people taking the data and building value-added on it, because the added value will also have to be freely released. [...] In practice [SA] freezes out a section of the community that can sometimes to very useful things with the data." David Eaves adds, "It basically kills the business communities (sic) interest in the open data. It also has a negative impact for some non-profits that may want to create mashups but have sensitive data they don't want to(or can't legally) share."
I don't think the semantic web argument (attribution links adding organization to the web) is strong enough. First, a bare link will not help organize the web. I can link to datasets from pages that are mere copies of the data, or from apps that use the data, or pages that modified the data, or I can link to datasets without using them at all. Even if bare links were of use, they would only help for those datasets that require attribution. The semantic web won't grow out of legal obligations to attribute. As far as I can tell, it will only grow out of voluntary inclusion of semantic markup, like RDFa, etc.
|
Re: Comparing Open Data TOU with ODC-by and ODC-PDDL
|
|
In reply to this post by David H. Mason
Not sure I understand. You can link to any license, so not sure how this is an unique advantage of CC. Also, any license with a -by is a non-starter for many people in the developer community I spoken with. See my post on open data and licenses. This is why the Open Government License in the UK is nice, as well as the PDDL, CC-0 could work but... ...your line "CC introduces many people to a well developed alternative to copyright." is exactly what concerns me. We don't need an alternative to copyright because this stuff was never copyrightable in the first place. Why would we educate people to believe that they don't have rights that they do? I'm also confident that the space can survive more than one license type. Certainly the open source community has thrived with several licenses, and indeed, the variety has made the community stronger, not more confused. I agree that simplicity is nice, but not at the expense of misinforming people about their rights. dave eaves On 11-02-18 2:20 PM, David H. Mason wrote:
|
Re: Comparing Open Data TOU with ODC-by and ODC-PDDL
|
|
If I may try to characterize... In your view, you envision the main users of the data now and in the future to be similar to the current open source community; a very small subset of the population, developers who produce apps for the majority of people. They want to build apps now in the simplest way possible. In my view, I'd like to see many users and producers of data, developers today but in a few years many people should be fluidly using data as they exchange professionally and in social networks. I'd like to see semantic applications that mix reusable public (city, etc) data with wikipedia data, subjective data, and so on, all with easy verifiability. I'd like to see a culture where if people are talking about something, their app will provide them with links to the original data sources and it's easy and normal for anyone to check sources (perhaps just the relative nerd of a group), meaningfully building on the trivia culture we have today with smartphones. It may scare official data producers now, but I don't think they should be afraid of developing relationships with data users (who will add value) and it seems kind of bizarre that professional organizations would work under fire-and-forget terms without these opportunities for feedback (which they're under no obligation to respond to). I think the NASCAR allusion is a failure of imagination, which will be overcome considering how valuable it would be for anyone to access a diverse set of data yet understand exactly where it all comes from, to go back to the sources. IMO what we are trying to get away from is overproduced consumer apps and instead enabling effective, easy to use apps that also invite casual individual inquiry and participation. Like search results that incorporate RDFa/microformats, if someone clicks on your pollution app, learns something, then clicks through to a deep, specific (semantic) link in the data source's web site, it's a win-win-win, with more opportunities for learning and engagement. So I think non-attribution public domain style licenses will stunt considerable potential, now and in the future. We're already working with richer data, and should be disappointed with efforts that don't produce it, so there's minimal technical overhead to including licenses if it's all inlined CC REL style. There do however have to be some kind of LGPL style conditions for mixing. Share-alike also has to be debated since many businesses won't want to contribute back. But regarding "viral" licenses, we can call many kinds of legal conditions viral, but perhaps it's appropriate for publicly funded data to be virally available? Regarding policing licenses, normally it won't be much of a burden, but it is useful to be able to police them when there is flagrant abuse. I don't see your point about copyright being important, certainly not enough to not use the well developed CC. It's all about arriving at the best terms of use going forward. To James McKinney's points, I hope people are not making data license concessions for the sake of Twitter's 140 character limit. And I absolutely think -by would contribute to development of the semantic web and the links will become more meaningful. DavidM On 18 February 2011 19:17, David Eaves <[hidden email]> wrote:
|
Re: Comparing Open Data TOU with ODC-by and ODC-PDDL
|
|
David - definitely not in my view that the main users of data will
the current open source community. Quite the opposite, that's why I
advise against a "by" license - very much see a much, much broader
community.
My sense is that the NASCAR problem is real (as I'm facing it right now) but perhaps I just lack imagination, but as we wrestle with this and the complete lack of clarity around what attribution means and where it needs to be, these are real problems being faced by real people, today. As for the viral stuff, I think we have very different visions - if you include that provision then you really will be limiting the data use to a tiny group of open source community users. I'll stop there, between my blog posts and comments here, feel like I've had my say and leave it at that. dave On 11-02-18 5:19 PM, David H. Mason wrote:
|
Re: Comparing Open Data TOU with ODC-by and ODC-PDDL
|
|
Attribution is something that is very difficult in distributed mapping contexts and I know that many scientists find it difficult to abide by as well. In both cases the geomaticians and the scientists are excellent when they attribute data in their papers, but it seems more problematic to do in in complex distributed or complex data model contexts. If the data are however wrapped with good metadata, this is less problematic as the attribution would be captured there.
I am paraphrasing loosely from what I have heard.
On Fri, Feb 18, 2011 at 8:30 PM, David Eaves <[hidden email]> wrote:
-- Tracey P. Lauriault 613-234-2805 |
Re: Comparing Open Data TOU with ODC-by and ODC-PDDL
|
|
I think the main issue here at the moment is can we find something that sounds reasonable to the City that they will accept and others can use. Saying we want full public domain maybe too far a stretch.
Cheerio John |
Re: Comparing Open Data TOU with ODC-by and ODC-PDDL
|
|
In reply to this post by Tracey P. Lauriault
David Mason, I've summarized many people's opinions for why both BY and SA licenses are not a good idea. From my reading, you defend BY by stating that the problem is a "failure of imagination", but you provide no solution, only confidence that it will be dealt with. Attribution is a problem many developers are currently facing, and have been facing for at least a few years now. If there were a solution, I don't think we'd still be talking about it. Maybe go off-list with Eaves to understand his encounter with that problem to enrich your understanding. As for SA, you offer no counter-argument, only the question, "perhaps it's appropriate for publicly funded data to be virally available?"
I think you're failing to consider issues that exist in the current, real world.
If you want to see people mix data from multiple sources, you should know that, today, it's hard to respect all the various BY licenses, especially when real estate is small (be that mobile screens, SMS messages, or Twitter updates -- all currently used far more often by far more people today than any semantic web app.) And if you want to mix data, you may not be able to unless all the SA licenses are compatible. If you want to see people mix data, a PDDL-like license is preferable.
There are many ways to incentivize attributing sources. I think putting it in the license is too hard an incentive. Attributing sources already increases people's trust in the data - I think that's incentive enough. There are also many ways in which data producers are connecting with data users; today, this is mostly by social means, including mailing lists, blog posts, feedback requests, Twitter conversations, etc. I don't see how a license or a linkback significantly facilitates this conversation.
> I hope people are not making data license concessions for the sake of Twitter's 140 character limit.
I hope people will not make data license concessions for the sake of a vision of a semantic web which doesn't yet exist. Solutions should solve real world problems. I still fail to see how a PDDL(-like) license would significantly hinder the development of a semantic web.
|
Re: Comparing Open Data TOU with ODC-by and ODC-PDDL
|
|
My two cents worth for SA-BY licenses, as the most feasible option right now...
Non-SA licenses (such as CC-BY): IMO, these licenses simply won't work for the vast majority of datasets that aren't covered by copyright in Canada. The reality is that this space has a lot less flexibility for licensing terms than in the copyright space. Data licenses need to be SA, or there's limited benefit in having license terms at all (no license terms -- ie. PDDL -- is, of course, also an option I support, but I'm not sure we're there yet). For copyrighted works, the rights "move with the work". The limitations within license terms are generally preserved even without a SA-clause. However, there's no such protection for data. Without SA, any middleman can simply clear the work of all of the license clauses. For example, if I download a dataset that's CC-BY, I could post up the exact same dataset on my website, as long as I attribute the data. However, anyone downloading it from my website has no such obligation. Same thing goes for any limitation of liability terms. As another quick thought, I wonder if it would be possible to limit the scope of the SA clause with some careful wording. With software licenses, SA generally only applies to (a) redistribution of (b) derivative works. For the liability issue with data, at some extent of permutation from an original dataset, I think the already small chance of liability would shift further towards the redistributor, thus making a similar scope restriction viable. It's not clear what the analog of this would be for data, but perhaps a middle ground could be found... Attribution generally: In my opinion, attribution is important for traceability. If I find a dataset on a third-party website, it's important to know the original source to verify, for example, that the dataset isn't a commercial database that someone has pulled off of bittorrent. The original source is equally important to know if there's ever a legal dispute about the rights. Kent On Sat, Feb 19, 2011 at 4:38 AM, James McKinney <[hidden email]> wrote:
|
Re: Comparing Open Data TOU with ODC-by and ODC-PDDL
|
|
In reply to this post by James McKinney
Responses below. On 18 February 2011 22:38, James McKinney <[hidden email]> wrote:
I mentioned some reasons BY and SA should be supported, and suggested public data should be published using CC versions of these licenses. Public domain type licenses could be a simple solution in the "current, real world." But in the (vast) majority of cases there would be no contribution back in terms of development or links and an opportunity would be lost to constantly better organize data. I'd suggest building on available capabilities that enable rather than deprecate a practical semantic web, where links are extremely important. Public domain terms are a huge compromise that creates many offline conversations of the nature I'm quite familiar with (not to diminish the importance of those discussions). "Failure of imagination" (no offence intended to anyone) was meant to refer to presentation and mechanics. We're not talking about cars or paper documents, and should determine the best approach considering data as hypermedia that anyone can access. Many would absolutely! like to see a digital network equivalent of stickered data, with embedded terms of use/source links easily available for human or computer user. RDFa/microformats are a significant representation of this today. Some data sets will need to be public domain. But without really limiting immediate applications, this is an opportunity for public data sources to support a structured re-use system and terms with long-term knowledge, participation and economic value, and not institutionalize a shortcut public domain approach. DavidM |
Re: Comparing Open Data TOU with ODC-by and ODC-PDDL
|
|
Just read this thread now and my apologies for resurrecting the thread, interesting conversation though. I stand mostly with Eaves and McKinney - CC and attribution will be a mess going forward. I do differ a bit with eaves and such in the sense that I think the main users of data will remain developers, journalists, businesses, community organizers and other professionals. The average joe will not spend his sunday afternoon mashing data up.
Public domain would be great by me. PDDL is as good as I have seen. The thing is, in my mind, you want to make it as easy as possible to use and re-use data (or parts of data). Attribution and respecting multiple licenses will put a serious stick in that wheel. The self-correcting nature of the web will ensure the quality of the data (see this open information bank called wikipedia). Last comment, fashion industry does not use attribution when creating new garments and has no real form of copyright on their creations, yet has a vibrant (and profitable) business model. I think that is our goal: http://blog.ted.com/2010/05/25/lessons_from_fa/ Gotta go! On 2011-02-19, at 11:49 AM, David H. Mason wrote:
|
Re: Comparing Open Data TOU with ODC-by and ODC-PDDL
|
|
http://eaves.ca/2011/03/01/lessons-from-fashions-free-culture-johanna-blakley-on-ted-com/
"This TEDx talk by Johanna Blakley is pure gold. A wonderful dissection of how patents and licenses are not only not necessary for innovation but can actually impede it - all while using the fashion industry as a case study." On Tue, Mar 1, 2011 at 10:25 AM, Jonathan Brun <[hidden email]> wrote: <snip> > Last comment, fashion industry does not use attribution when creating new > garments and has no real form of copyright on their creations, yet has a > vibrant (and profitable) business model. I think that is our > goal: http://blog.ted.com/2010/05/25/lessons_from_fa/ > Gotta go! > JB > MontrealOuvert.net > jonathanbrun.com <snip> |
Re: Comparing Open Data TOU with ODC-by and ODC-PDDL
|
|
Sorry, Michael, meant to credit you in that blog post. Back to add you
in now! On 11-03-01 8:43 AM, Michael Lenczner wrote: > http://eaves.ca/2011/03/01/lessons-from-fashions-free-culture-johanna-blakley-on-ted-com/ > > "This TEDx talk by Johanna Blakley is pure gold. A wonderful > dissection of how patents and licenses are not only not necessary for > innovation but can actually impede it - all while using the fashion > industry as a case study." > > > On Tue, Mar 1, 2011 at 10:25 AM, Jonathan Brun<[hidden email]> wrote: > <snip> >> Last comment, fashion industry does not use attribution when creating new >> garments and has no real form of copyright on their creations, yet has a >> vibrant (and profitable) business model. I think that is our >> goal: http://blog.ted.com/2010/05/25/lessons_from_fa/ >> Gotta go! >> JB >> MontrealOuvert.net >> jonathanbrun.com > <snip> > _______________________________________________ > CivicAccess-discuss mailing list > [hidden email] > http://lists.pwd.ca/mailman/listinfo/civicaccess-discuss |
Re: Comparing Open Data TOU with ODC-by and ODC-PDDL
|
|
No prob - but it was Jon Brun, not me. :)
On Tue, Mar 1, 2011 at 11:45 AM, David Eaves <[hidden email]> wrote: > Sorry, Michael, meant to credit you in that blog post. Back to add you in > now! > > On 11-03-01 8:43 AM, Michael Lenczner wrote: >> >> >> http://eaves.ca/2011/03/01/lessons-from-fashions-free-culture-johanna-blakley-on-ted-com/ >> >> "This TEDx talk by Johanna Blakley is pure gold. A wonderful >> dissection of how patents and licenses are not only not necessary for >> innovation but can actually impede it - all while using the fashion >> industry as a case study." >> >> >> On Tue, Mar 1, 2011 at 10:25 AM, Jonathan Brun<[hidden email]> >> wrote: >> <snip> >>> >>> Last comment, fashion industry does not use attribution when creating new >>> garments and has no real form of copyright on their creations, yet has a >>> vibrant (and profitable) business model. I think that is our >>> goal: http://blog.ted.com/2010/05/25/lessons_from_fa/ >>> Gotta go! >>> JB >>> MontrealOuvert.net >>> jonathanbrun.com >> >> <snip> >> _______________________________________________ >> CivicAccess-discuss mailing list >> [hidden email] >> http://lists.pwd.ca/mailman/listinfo/civicaccess-discuss > > _______________________________________________ > CivicAccess-discuss mailing list > [hidden email] > http://lists.pwd.ca/mailman/listinfo/civicaccess-discuss > |
«
Return to CivicAccess
|
1 view|%1 views
| Free forum by Nabble | Edit this page |