creative commons licenses
Summary
- There are 11 posts — by 7 authors — in this topic.
- Latest post made by Brent Wood at 2010 Jul 17 06:49 NZST
While on the subject of making data available and licenses. I was wondering whether people felt that the existing CC licenses met the following need. We want you to have the data you can render it how you like you can use it to generate new data and renderings you need to attribute the source to us. you must preserve the integrity of the data. The last is the key one. creating a derivative work - such as a new visualization or using the data as input to another system is something to encourage. However modifying the data - while still attributing it to the source could be problematic. Cleaning up data is a dangerous activity - removing outliers, or using the wrong type of averaging might conceal genuine changes. Even trimming a time series to specific start and end dates can affect statistics - e.g. if periodic data is involved. So the license needs to say something like: 1. you must somehow point back to the original numbers 2. if you manipulate the numbers then that should be transparent - e.g. show working/source code. 3. you should not imply that the manipulated numbers are the same as the original source. the NoDeriv option seems too restrictive - to me it implies that you can't reprocess the data at all or generate visualizations. Yet Attribution alone doesn't seem to ensure fidelity with the data. Comments welcome.
Andrew
On 12 July 2010 13:33, Andrew Watkins <email obscured>> wrote: > While on the subject of making data available and licenses. I was > wondering whether people felt that the existing CC licenses met the > following need. > > We want you to have the data > you can render it how you like > you can use it to generate new data and renderings > you need to attribute the source to us. > you must preserve the integrity of the data. > [disclaimer: not a lawyer, don't rely on this, take your own advice, no point suing me anyway - i'm a student, etc etc] Licencing is a difficult issue. For data especially. An individual datum is probably not subject to copyright, because it probably wouldn't be a creative work. This means that the Creative Commons licences are not entirely appropriate for these situations. They would need to be handled by contract law more broadly, probably as part of an API usage agreement. One thing to make sure of when using contract law is to ensure that the parties are forming a contract. This requires legal consideration & and an intention to form legal relations. If an agency makes an API available without sign-in, there's an argument to say that there's no intention to form legal relations there, because it's so informal. Consideration is also a stretch. Many contracts have clauses that grant a single dollar in exchange for the use of the service. In the case of a zero-cost data API, it's a harder argument to establish that consideration was on both sides. I would like others' opinions on this - I don't want to spout out more lies if my understanding isn't correct.
Tim.
On 11/07/2010, at 7:49 PM, Tim McNamara wrote: > [disclaimer: not a lawyer, don't rely on this, take your own advice, > no > point suing me anyway - i'm a student, etc etc] Interesting questions, Tim, and I'm curious to know the answers. However, I'm also wary of conversations about the law between people who aren't lawyers. If we have a lawyer, or someone who has received a lawyer's opinion, I'd love to hear their take. But it's a side- issue from the point that Andrew raised. Andrew's question was about CC licenses failing to distinguish between derivative works and distorted works. My understanding of CC is that "derivative" is as granular as you get, that there's no legal mechanism for identifying or prohibiting distortion of your data. I suspect this is because "distortion", "manipulation", and so on are subjective. Folks think they've done the right thing until shown otherwise, and in some cases not even then.
Cheers; Nat
In the public sector context, NZGOAL will deal with most of these issues: - Creative Commons Attribution (BY) is the default licence for copyright works. - The presumption for that licence can be displaced if a restriction applies but agencies are urged to think seriously before utilising a no-derivatives form of licence, given the restrictions that such a licence could place on creative and economically valuable uses of datasets. - While individual items of data are unlikely to attract copyright protection, datasets and databases are often capable of qualifying as copyright works (they are a subset of "literary works"). NZGOAL explains all this. (The draft version is already available online on the e.govt.nz website.) In some instances there may be a need for separate API terms, e.g., to regulate the frequency of pinging of an API. NZGOAL does not deal with this issue as it's conceptually distinct. This may, however, be a good subject for a separate guidance note (NZGOAL contemplates discrete guidance notes on given topics). - To some extent the integrity issue can be dealt with by requiring a specific form of attribution, i.e., if the dataset is modified in any way then the attribution statement needs to say so. The finalised version of NZGOAL will provide such sample attribution statements for agencies to include in their copyright and licensing statements if they wish. Bear in mind also that, if attribution is required, there should usually be a link back to the source data, enabling people to consult or verify against it if required. I hope this helps. Feel free to raise more questions if you like. Richard Best
> Date: Mon, 12 Jul 2010 13:33:49 +1200 > Subject: [ninjas] creative commons licenses > From: <email obscured> > To: <email obscured> > > While on the subject of making data available and licenses. I was > wondering whether people felt that the existing CC licenses met the > following need. > > We want you to have the data > you can render it how you like > you can use it to generate new data and renderings > you need to attribute the source to us. > you must preserve the integrity of the data. > > The last is the key one. creating a derivative work - such as a new > visualization or using the data as input to another system is > something to encourage. However modifying the data - while still > attributing it to the source could be problematic. > > Cleaning up data is a dangerous activity - removing outliers, or using > the wrong type of averaging might conceal genuine changes. Even > trimming a time series to specific start and end dates can affect > statistics - e.g. if periodic data is involved. > > So the license needs to say something like: > 1. you must somehow point back to the original numbers > 2. if you manipulate the numbers then that should be transparent - > e.g. show working/source code. > 3. you should not imply that the manipulated numbers are the same as > the original source. > > the NoDeriv option seems too restrictive - to me it implies that you > can't reprocess the data at all or generate visualizations. Yet > Attribution alone doesn't seem to ensure fidelity with the data. > > Comments welcome. > > Andrew > > ----------------------------------------- > Full text of this topic in The Open Government Ninjas: > http://groups.open.org.nz/r/topic/lmvuRvd8lsdk9KVZq8KMX > > To leave The Open Government Ninjas, email > <email obscured>?Subject=unsubscribe > > Start your own free groups and site with > OnlineGroups.Net http://onlinegroups.net > > Host your own online groups site with > GroupServer http://groupserver.org _________________________________________________________________ Your E-mail and More On-the-Go. Get Windows Live Hotmail Free. https://signup.live.com/signup.aspx?id=60969
Generally I think Creative Commons do not recommend their content licenses for databases. More on open data licensing at: http://www.opendefinition.org/guide/data/ Data licenses like the Open Data Commons licenses and CC0 are meant for data: http://www.opendatacommons.org/ http://wiki.creativecommons.org/CC0_FAQ Regarding pedigree of data and ensuring that it is not 'falsified' -- I'm not convinced that licenses are necessarily the right way to do this. E.g. a fair bit of data that we've got from government is of questionable quality too! There are also technical and social measures to track changes. E.g. wiki-like versioning for data so one can track / roll back changes. Also, plagiarism is a good example of dealing with malpractice with social norms rather than purely legal measures. All the best, Jonathan On Mon, Jul 12, 2010 at 7:40 AM, R B <email obscured>> wrote: > > In the public sector context, NZGOAL will deal with most of these issues: > > - Creative Commons Attribution (BY) is the default licence for copyright works. > - The presumption for that licence can be displaced if a restriction applies but agencies are urged to think seriously before utilising a no-derivatives form of licence, given the restrictions that such a licence could place on creative and economically valuable uses of datasets. > - While individual items of data are unlikely to attract copyright protection, datasets and databases are often capable of qualifying as copyright works (they are a subset of "literary works"). NZGOAL explains all this. (The draft version is already available online on the e.govt.nz website.) In some instances there may be a need for separate API terms, e.g., to regulate the frequency of pinging of an API. NZGOAL does not deal with this issue as it's conceptually distinct. This may, however, be a good subject for a separate guidance note (NZGOAL contemplates discrete guidance notes on given topics). > - To some extent the integrity issue can be dealt with by requiring a specific form of attribution, i.e., if the dataset is modified in any way then the attribution statement needs to say so. The finalised version of NZGOAL will provide such sample attribution statements for agencies to include in their copyright and licensing statements if they wish. Bear in mind also that, if attribution is required, there should usually be a link back to the source data, enabling people to consult or verify against it if required.
> > I hope this helps. Feel free to raise more questions if you like. > > Richard Best > >> Date: Mon, 12 Jul 2010 13:33:49 +1200 >> Subject: [ninjas] creative commons licenses >> From: <email obscured> >> To: <email obscured> >> >> While on the subject of making data available and licenses. I was >> wondering whether people felt that the existing CC licenses met the >> following need. >> >> We want you to have the data >> you can render it how you like >> you can use it to generate new data and renderings >> you need to attribute the source to us. >> you must preserve the integrity of the data. >> >> The last is the key one. creating a derivative work - such as a new >> visualization or using the data as input to another system is >> something to encourage. However modifying the data - while still >> attributing it to the source could be problematic. >> >> Cleaning up data is a dangerous activity - removing outliers, or using >> the wrong type of averaging might conceal genuine changes. Even >> trimming a time series to specific start and end dates can affect >> statistics - e.g. if periodic data is involved. >> >> So the license needs to say something like: >> 1. you must somehow point back to the original numbers >> 2. if you manipulate the numbers then that should be transparent - >> e.g. show working/source code. >> 3. you should not imply that the manipulated numbers are the same as >> the original source. >> >> the NoDeriv option seems too restrictive - to me it implies that you >> can't reprocess the data at all or generate visualizations. Yet >> Attribution alone doesn't seem to ensure fidelity with the data. >> >> Comments welcome. >> >> Andrew >> >> ----------------------------------------- >> Full text of this topic in The Open Government Ninjas: >> http://groups.open.org.nz/r/topic/lmvuRvd8lsdk9KVZq8KMX >> >> To leave The Open Government Ninjas, email >> <email obscured>?Subject=unsubscribe >> >> Start your own free groups and site with >> OnlineGroups.Net http://onlinegroups.net >> >> Host your own online groups site with >> GroupServer http://groupserver.org > > _________________________________________________________________ > Your E-mail and More On-the-Go. Get Windows Live Hotmail Free. > https://signup.live.com/signup.aspx?id=60969 > ----------------------------------------- > Full text of this topic in The Open Government Ninjas: > http://groups.open.org.nz/r/topic/6XtFU07X4yrS8Y6Ua6vKda > > To leave The Open Government Ninjas, email > <email obscured>?Subject=unsubscribe > > Start your own free groups and site with > OnlineGroups.Net http://onlinegroups.net > > Host your own online groups site with > GroupServer http://groupserver.org > -- Jonathan Gray Community Coordinator The Open Knowledge Foundation http://blog.okfn.org http://twitter.com/jwyg http://identi.ca/jwyg
On 15/07/2010, at 8:29 AM, Jonathan Gray wrote: > Generally I think Creative Commons do not recommend their content > licenses for databases. When Richard was working on the NZGOAL recommendations, I specifically queried this. I wanted CC0 as the default for govt data. I connected Richard with Puneet Kishor, a fellow at Science Commons. Puneet initially said CC0 too, but Richard pointed to some work that suggests the government *can't* waive its copyright. Puneet did some research and said this: > The bottom-line advice -- the approach you are taking is appropriate > and correct. If you can't waive copyright, go ahead and apply an > attribution license, and then waive or dilute the attribution > requirement. A CC-BY 3.0 with a waiver of attribution is effectively > as good as CC0, and in practice, it can work out best in most > situations. If you look at CC-BY 3.0 (see http://creativecommons.org/licenses/by/3.0/us/) > , it offers the capability that "Any of the (attribution) conditions > can be waived..." > > So, to summarize, CC-BY 3.0 with a weak/diluted/or completely waived > attribution requirement, with a link back to the CC-BY 3.0 license > page, and preferably, with the license info embedded in the work (if > the work is digital and online), is a very sound option for your > needs. Hope this helps,
Nat
On Thu, Jul 15, 2010 at 5:50 PM, Nathan Torkington <email obscured>> wrote: > On 15/07/2010, at 8:29 AM, Jonathan Gray wrote: >> Generally I think Creative Commons do not recommend their content >> licenses for databases. > > When Richard was working on the NZGOAL recommendations, I specifically > queried this. I wanted CC0 as the default for govt data. I connected > Richard with Puneet Kishor, a fellow at Science Commons. Puneet > initially said CC0 too, but Richard pointed to some work that suggests > the government *can't* waive its copyright. We definitely understand that public domain is not always the best or only way to go for data. This is exactly why we've developed attribution and attribution-sharealike licenses specifically for data, which are currently being reviewed for use by Open Street Map and the UK government: http://www.opendatacommons.org/licenses/ Discussion about licenses at: http://lists.okfn.org/mailman/listinfo/odc-discuss >> The bottom-line advice -- the approach you are taking is appropriate >> and correct. If you can't waive copyright, go ahead and apply an >> attribution license, and then waive or dilute the attribution >> requirement. A CC-BY 3.0 with a waiver of attribution is effectively >> as good as CC0, and in practice, it can work out best in most >> situations. If you look at CC-BY 3.0 (see http://creativecommons.org/licenses/by/3.0/us/) >> , it offers the capability that "Any of the (attribution) conditions >> can be waived..." >> >> So, to summarize, CC-BY 3.0 with a weak/diluted/or completely waived >> attribution requirement, with a link back to the CC-BY 3.0 license >> page, and preferably, with the license info embedded in the work (if >> the work is digital and online), is a very sound option for your >> needs. Various folks (including folks at CC! [1]) are still not convinced about the suitability of CC *content* licenses (which mainly cover copyright) for rights in *data* (such as EC Database Directive). Some jurisdictions have 'hacked' licenses to include rights in data, but not yet clear whether this is a robust, long term legal solution (as opposed to a shorter term hack...). [1] http://bibwild.wordpress.com/2008/11/24/creative-commons-is-not-appropriate-for-data/
-- Jonathan Gray Community Coordinator The Open Knowledge Foundation http://blog.okfn.org http://twitter.com/jwyg http://identi.ca/jwyg
On Fri, Jul 16, 2010 at 3:50 AM, Nathan Torkington <email obscured>> wrote: > On 15/07/2010, at 8:29 AM, Jonathan Gray wrote: > > Generally I think Creative Commons do not recommend their content > > licenses for databases. > > When Richard was working on the NZGOAL recommendations, I specifically > queried this. I asked this too. Richard said that under NZ law, CC licenses could apply to data much more clearly than in other jurisdictions where it all becomes a grey area (ala ODbL and other solutions that are being worked on). Rob :)
Hi All I thought it might help to follow up on a few of the points in this thread. (1) EXCHANGE WITH PUNEET KISHOR When Nat very helpfully put us in contact with Puneet Kishor (a Science Commons Fellow), he said he thought it might help to run the NZGOAL framework past him to make sure that it's going to work for data and not just documents. We gladly took up this opportunity. In my email to Nat and Puneet, I set out an excerpt from another document which contained our current thinking on utilising Creative Commons licences (and not CC Zero) for datasets. That excerpt was as follows: ******* "As far as I'm aware, there are three main arguments against using CC licences for copyright datasets. Those arguments can be summarised as follows: potential attribution stacking problems where multiple datasets are mashed up; complexities arising from those who select a CC licence variant that imposes a share-like requirement (the share-alike requirement potentially having a stifling effect on downstream exploitation); and so-called "category errors", i.e., where non-copyright data has a copyright licence placed on it inappropriately or where a copyright dataset does properly carry a copyright licence but an end user takes less than a "substantial part" yet thinks it is bound by the licence terms when (because it's taking less than a substantial part) he/she/it is not. Those concerned by such apparent problems tend to suggest one of two different solutions: The first solution, and the one which Science Commons advocates, is for the dataset owner to waive all copyright and related works in the dataset, by using CC Zero or the older (and US-centric) CC Public Domain Dedication and Certification tool. The use of these tools largely assumes that it is legally possible to completely waive statutory copyright in a copyright work when - at least in certain Commonwealth countries - that may not in fact be the case (as argued by the UK's Phillip Johnson in "'Dedicating' Copyright to the Public Domain" (2008) 71(4) Modern Law Review 587-610). In the governmental context, it would also raise significant policy issues around the propriety of waiving Crown copyright. The second solution is to advocate the use of an alternative to CC licences, i.e., a dataset-specific licence that contains less restrictions. Our current view is that there is merit in advocating the use of CC licences for all kinds copyright works that can be released for re-use, including copyright datasets and databases. To advocate the use of separate licences for datasets would run the risk of further licence proliferation which is one of the current problems NZGOAL is trying to resolve. There is also the fact that the CC model is very widely known and respected and benefits from the CC infrastructure, including the three different forms for each licence, i.e., human-readable, lawyer-readable and machine-readable. As you'll gather from my comments above, we do not support the CCZero / full waiver of copyright approach, as it raises complex legal and policy issues, the resolution of which is - in our view - unnecessary at this point. The current approach in the draft NZGOAL is, if accepted by Cabinet, expected to see the freeing-up of copyright works and datasets more quickly than an approach which could be legally questionable and more complex from a policy angle. We have, nevertheless, been mindful of the potential for attribution-stacking problems and we have received further feedback suggesting we ought to take this expressly in account within NZGOAL. The approach we have adopted at the moment is to add a new policy principle into NZGOAL, one which would encourage agencies releasing datasets to think long and hard about potential attribution stacking problems downstream and, where there is any likelihood of that, to impose minimal attribution requirements on end users. It is also perfectly open to agencies to completely waive all attribution requirements whilst still retaining the other benefits of the CC licences. As such, we consider that the CC framework already has sufficient flexibility to accommodate potential attribution stacking problems. So far as the share-alike issue is concerned, NZGOAL cautions agencies against its use. As you'll know, it recommends CC-BY as default and warns agencies of the potentially stifling effect of using either share-alike or no derivative variants. We may need to consider adding a paragraph or two on this in the specific context of datasets. Something we'll ponder. The potential for category errors would remain with the CC model, as it would for other copyright licensing models, in contrast to full waivers (if they were in fact legally permissible). While it's a problem that may be more common in the data arena, it can exist in other areas as well. It's really a facet of the complexity of copyright law and not something, in our view, that can or should be 'fixed' in the governmental context by adoption of CCZero or other form of waiver. There may, nevertheless, be merit in the Crown releasing a specific guidance on this topic in due course. Again, something we'll ponder." ******* Nat has reproduced Puneet's response above. Puneet also prefaced his remarks by noting that he had (1) digested the Philip Johnson paper, (2) had conversations with faculty at the University of Wisconsin Law School, and (3) discussed matters Science Commons counsel, Thinh Nguyen. We did, of course, find Puneet's comments both helpful and reassuring. Again, many thanks Nat for putting us in touch. (2) CREATIVE COMMONS AND DATASETS I'm not sure it's entirely right to say Creative Commons do not recommend their licences for datasets/databases. Originally, it was quite the contrary. Their FAQs positively recommended CC licences for copyright databases/datasets. It was Science Commons that moved to an alternative model due to the three concerns set out above in the context of primarily scientific/research data. However, as discussed above, NZGOAL seeks to address two of the three concerns, whilst the third is an issue that is not unique to datasets/databases, and Puneet agreed with the approach NZGOAL advocates. (3) OPEN DATA LICENCES The fact that data licenses like the Open Data Commons licenses and CC0 may be designed with data in mind doesn't mean that Creative Commons licences are not suitable for copyright databases and datasets. I would also add that CC0 was not designed only for data; it was designed for any copyright work. Its purpose is to achieve a waiver of all copyright in a work and, if that fails, it has a fall back position of a very broad licence to re-use (with no attribution or other requirements on licensees). As noted above, it is not a part of the NZGOAL framework. (4) "DATABASE RIGHT" AND ALTERNATIVE DATABASE LICENCES There will probably always be competing views as to whether CC licences or alternative database licences are more appropriate for data and it's an issue we will follow with interest. These things evolve and nothing is necessarily set in stone. I would urge caution, however, when considering both discussions and alternative licences that are premised on concerns arising from "the database right". The database right is a European creation that has no equivalent under New Zealand law. (5) OTHER COMMENTS It is fair to say that copyright and licensing issues around datasets and chunks of data are not straight-forward. It is also correct to say that, while datasets and databases can qualify as copyright works, individual data items seldom do. To infringe copyright in a copyright dataset or database, ordinarily one needs to establish that the alleged infringer copied all or a "substantial part" of the dataset/database. This raises interesting issues. If, for example, a mashup were to take only a small/insubstantial amount of data from multiple publicly available datasets, it may not require any licence at all. As I say, there are some complex issues here (and nothing I've said should be taken as legal advice; sorry - regulatory requirement). It is also an area that is in a state of flux, internationally, in the courts. There has, for example, been recent litigation in Australia which, on a first read, restricts the copyright protection of databases, even where significant effort has gone into their creation. Last time I checked, the Australian case I have in mind was on appeal. So to some extent it's a question of 'watch this space'. It may well be that the various issues in this area should be the subject of a discrete guidance note in due course. Kind regards
Richard > Date: Fri, 16 Jul 2010 11:08:18 +1200 > Subject: Re: [ninjas] creative commons licenses > From: <email obscured> > To: <email obscured> > > On Fri, Jul 16, 2010 at 3:50 AM, Nathan Torkington <email obscured>> wrote: > > > On 15/07/2010, at 8:29 AM, Jonathan Gray wrote: > > > Generally I think Creative Commons do not recommend their content > > > licenses for databases. > > > > When Richard was working on the NZGOAL recommendations, I specifically > > queried this. > > > > I asked this too. Richard said that under NZ law, CC licenses could apply to > data much more clearly than in other jurisdictions where it all becomes a > grey area (ala ODbL and other solutions that are being worked on). > > Rob :) > > ----------------------------------------- > Full text of this topic in The Open Government Ninjas: > http://groups.open.org.nz/r/topic/6mLVpvvdMfENK8wfrECjXZ > > To leave The Open Government Ninjas, email > <email obscured>?Subject=unsubscribe > > Start your own free groups and site with > OnlineGroups.Net http://onlinegroups.net > > Host your own online groups site with > GroupServer http://groupserver.org _________________________________________________________________ Hotmail: Powerful Free email with security by Microsoft. https://signup.live.com/signup.aspx?id=60969
Many thanks for this detailed and interesting email Richard! For info I'm copying in Rufus Pollock and Jordan Hatcher -- who know more about this than I do! ;-) All the best,
Jonathan On Fri, Jul 16, 2010 at 8:12 AM, R B <email obscured>> wrote: > > > > Hi All > > > > I thought it might help to follow up on a few of the points > in this thread. > > > > (1) EXCHANGE WITH PUNEET KISHOR > > > > When Nat very helpfully put us in contact with Puneet Kishor > (a Science Commons Fellow), he said he thought it might help to run the NZGOAL framework > past him to make sure that it's going to work for data and not just documents. We gladly took up > this opportunity. In my email to Nat and Puneet, I set out an excerpt from another > document which contained our current thinking on utilising Creative Commons > licences (and not CC Zero) for datasets. That excerpt was as follows: > > > > ******* > > > > "As far as I'm aware, there are three main arguments > against using CC licences for copyright datasets. Those arguments can be > summarised as follows: > > > > > potential attribution stacking problems where > multiple datasets are mashed up; > > > > > complexities arising from those who select a CC > licence variant that imposes a share-like requirement (the share-alike > requirement potentially having a stifling effect on downstream exploitation); > and > > > > > so-called "category errors", i.e., > where non-copyright data has a copyright licence placed on it inappropriately > or where a copyright dataset does properly carry a copyright licence but an end > user takes less than a "substantial part" yet thinks it is bound by > the licence terms when (because it's taking less than a substantial part) > he/she/it is not. > > > > Those concerned by > such apparent problems tend to suggest one of two different solutions: > > > > > The first solution, and the one which Science > Commons advocates, is for the dataset owner to waive all copyright and related > works in the dataset, by using CC Zero or the older (and US-centric) CC Public > Domain Dedication and Certification tool. The use of these tools largely > assumes that it is legally possible to completely waive statutory copyright in > a copyright work when - at > least in certain Commonwealth countries - that may not in fact be the case (as argued > by the UK's Phillip Johnson in "'Dedicating' Copyright to the Public > Domain" (2008) 71(4) Modern Law Review 587-610). In the governmental > context, it would also raise significant policy issues around the propriety of > waiving Crown copyright. > > > The second solution is to advocate the use of an > alternative to CC licences, i.e., a dataset-specific licence that contains less > restrictions. > > > > Our current view is > that there is merit in advocating the use of CC licences for all kinds > copyright works that can be released for re-use, including copyright datasets > and databases. To advocate the use of separate licences for datasets would run > the risk of further licence proliferation which is one of the current problems > NZGOAL is trying to resolve. There is also the fact that the CC model is very > widely known and respected and benefits from the CC infrastructure, including > the three different forms for each licence, i.e., human-readable, > lawyer-readable and machine-readable. > > > > As you'll gather > from my comments above, we do not support the CCZero / full waiver of copyright > approach, as it raises complex legal and policy issues, the resolution of which > is - in our view - unnecessary at this point. The current approach in the draft > NZGOAL is, if accepted by Cabinet, expected to see the freeing-up of copyright > works and datasets more quickly than an approach which could be legally > questionable and more complex from a policy angle. > > > > We have, > nevertheless, been mindful of the potential for attribution-stacking problems > and we have received further feedback suggesting we ought to take this > expressly in account within NZGOAL. The approach we have adopted at the moment > is to add a new policy principle into NZGOAL, one which would encourage > agencies releasing datasets to think long and hard about potential attribution > stacking problems downstream and, where there is any likelihood of that, to > impose minimal attribution requirements on end users. It is also perfectly open > to agencies to completely waive all attribution requirements whilst still > retaining the other benefits of the CC licences. As such, we consider that the > CC framework already has sufficient flexibility to accommodate potential attribution > stacking problems. > > > > So far as the > share-alike issue is concerned, NZGOAL cautions agencies against its use. As > you'll know, it recommends CC-BY as default and warns agencies of the > potentially stifling effect of using either share-alike or no derivative > variants. We may need to consider adding a paragraph or two on this in the > specific context of datasets. Something we'll ponder. > > > > The potential for > category errors would remain with the CC model, as it would for other copyright > licensing models, in contrast to full waivers (if they were in fact legally > permissible). While it's a problem that may be more common in the data arena, > it can exist in other areas as well. It's really a facet of the complexity of > copyright law and not something, in our view, that can or should be 'fixed' in > the governmental context by adoption of CCZero or other form of waiver. There > may, nevertheless, be merit in the Crown releasing a specific guidance on this > topic in due course. Again, something we'll ponder." > > > > ******* > > > > Nat has reproduced Puneet's response above. Puneet also > prefaced his remarks by noting that he had (1) digested the Philip Johnson > paper, (2) had conversations with faculty at the University of Wisconsin Law > School, and (3) discussed matters Science Commons counsel, Thinh Nguyen. We > did, of course, find Puneet's comments both helpful and reassuring. Again, many > thanks Nat for putting us in touch. > > > > (2) CREATIVE COMMONS AND DATASETS > > > > I'm not sure it's entirely right to say Creative Commons do not > recommend their licences for datasets/databases. Originally, it was quite the > contrary. Their FAQs positively recommended CC licences for copyright databases/datasets. > It was Science Commons that moved to an alternative model due to the three > concerns set out above in the context of primarily scientific/research data. > However, as discussed above, NZGOAL seeks to address two of the three concerns, > whilst the third is an issue that is not unique to datasets/databases, and > Puneet agreed with the approach NZGOAL advocates. > > > > (3) OPEN DATA LICENCES > > > > The fact that data licenses like the Open Data Commons > licenses and CC0 may be designed with data in mind doesn't mean that Creative > Commons licences are not suitable for copyright databases and datasets. I would > also add that CC0 was not designed only for data; it was designed for any > copyright work. Its purpose is to achieve a waiver of all copyright in a work > and, if that fails, it has a fall back position of a very broad licence to > re-use (with no attribution or other requirements on licensees). As noted > above, it is not a part of the NZGOAL framework. > > > > (4) "DATABASE RIGHT" AND ALTERNATIVE DATABASE > LICENCES > > > > There will probably always be competing views as to whether > CC licences or alternative database licences are more appropriate for data and > it's an issue we will follow with interest. These things evolve and nothing is > necessarily set in stone. I would urge caution, however, when considering both discussions > and alternative licences that are premised on concerns arising from "the > database right". The database right is a European creation that has no > equivalent under New Zealand law. > > > > (5) OTHER COMMENTS > > > > It is fair to say that copyright and licensing issues around > datasets and chunks of data are not straight-forward. It is also correct to say > that, while datasets and databases can qualify as copyright works, individual > data items seldom do. > > > > To infringe copyright in a copyright dataset or database, > ordinarily one needs to establish that the alleged infringer copied all or a > "substantial part" of the dataset/database. This raises interesting > issues. If, for example, a mashup were to take only a small/insubstantial > amount of data from multiple publicly available datasets, it may not require > any licence at all. > > > > As I say, there are some complex issues here (and nothing > I've said should be taken as legal advice; sorry - regulatory requirement). It is also an area that is in a state of flux, internationally, in the courts. There has, for example, been recent litigation in Australia which, on a first read, restricts the copyright protection of databases, even where significant effort has gone into their creation. Last time I checked, the Australian case I have in mind was on appeal. So to some extent it's a question of 'watch this space'. > > It > may well be that the various issues in this area should be the subject of a > discrete guidance note in due course. > > > > Kind regards > > Richard > > > >> Date: Fri, 16 Jul 2010 11:08:18 +1200 >> Subject: Re: [ninjas] creative commons licenses >> From: <email obscured> >> To: <email obscured> >> >> On Fri, Jul 16, 2010 at 3:50 AM, Nathan Torkington <email obscured>> wrote: >> >> > On 15/07/2010, at 8:29 AM, Jonathan Gray wrote: >> > > Generally I think Creative Commons do not recommend their content >> > > licenses for databases. >> > >> > When Richard was working on the NZGOAL recommendations, I specifically >> > queried this. >> >> >> >> I asked this too. Richard said that under NZ law, CC licenses could apply to >> data much more clearly than in other jurisdictions where it all becomes a >> grey area (ala ODbL and other solutions that are being worked on). >> >> Rob :) >> >> ----------------------------------------- >> Full text of this topic in The Open Government Ninjas: >> http://groups.open.org.nz/r/topic/6mLVpvvdMfENK8wfrECjXZ >> >> To leave The Open Government Ninjas, email >> <email obscured>?Subject=unsubscribe >> >> Start your own free groups and site with >> OnlineGroups.Net http://onlinegroups.net >> >> Host your own online groups site with >> GroupServer http://groupserver.org > > _________________________________________________________________ > Hotmail: Powerful Free email with security by Microsoft. > https://signup.live.com/signup.aspx?id=60969 > ----------------------------------------- > Full text of this topic in The Open Government Ninjas: > http://groups.open.org.nz/r/topic/7KQ4cSHMKPywcmM0W1w3gD > > To leave The Open Government Ninjas, email > <email obscured>?Subject=unsubscribe > > Start your own free groups and site with > OnlineGroups.Net http://onlinegroups.net > > Host your own online groups site with > GroupServer http://groupserver.org > -- Jonathan Gray Community Coordinator The Open Knowledge Foundation http://blog.okfn.org http://twitter.com/jwyg http://identi.ca/jwyg
The attribution stacking concern is one I have expressed before, and I'm pleased to see it so well described here. The solution I have suggested is for NZGOAL to provide & recommend a standard, facilitative attribution clause, with a ",,,, [Your organisation here] ..." approach, so both the licence & attribution are provided. The other issue, not mentioned in this thread, is that metadata, particularly "fitness for use" information is needed for a dataset release to be useful, but no such requirement exists for creative works. There is also the discoverability issue, ie: enabling online datasets to be found is a useful precursor to using them. Sufficient metadata, including abstracts & keywords, needs to be provided to give search engines traction in locating relevant datasets (and filtering less relevant ones). Any project, such as NZGOAL, looking at what is required to faciliate useful public release of data, should, IMHO, cover licence, attribution and metadata recommendations, as a one-stop-shop for how to go about releasing data to the public. Cheers, Brent Wood
--- On Fri, 7/16/10, Jonathan Gray <email obscured>> wrote: > From: Jonathan Gray <email obscured>> > Subject: Re: [ninjas] creative commons licenses > To: <email obscured>, "R B" <email obscured>> > Cc: "Jordan S Hatcher" <email obscured>>, "Rufus Pollock" <email obscured>> > Date: Friday, July 16, 2010, 9:06 PM > Many thanks for this detailed and > interesting email Richard! > > For info I'm copying in Rufus Pollock and Jordan Hatcher -- > who know > more about this than I do! ;-) > > All the best, > > Jonathan > > On Fri, Jul 16, 2010 at 8:12 AM, R B <email obscured>> > wrote: > > > > > > > > Hi All > > > > > > > > I thought it might help to follow up on a few of the > points > > in this thread. > > > > > > > > (1) EXCHANGE WITH PUNEET KISHOR > > > > > > > > When Nat very helpfully put us in contact with Puneet > Kishor > > (a Science Commons Fellow), he said he thought it > might help to run the NZGOAL framework > > past him to make sure that it's going to work for data > and not just documents. We gladly took up > > this opportunity. In my email to Nat and Puneet, I set > out an excerpt from another > > document which contained our current thinking on > utilising Creative Commons > > licences (and not CC Zero) for datasets. That excerpt > was as follows: > > > > > > > > ******* > > > > > > > > "As far as I'm aware, there are three main arguments > > against using CC licences for copyright datasets. > Those arguments can be > > summarised as follows: > > > > > > > > > > potential attribution stacking problems where > > multiple datasets are mashed up; > > > > > > > > > > complexities arising from those who select a CC > > licence variant that imposes a share-like requirement > (the share-alike > > requirement potentially having a stifling effect on > downstream exploitation); > > and > > > > > > > > > > so-called "category errors", i.e., > > where non-copyright data has a copyright licence > placed on it inappropriately > > or where a copyright dataset does properly carry a > copyright licence but an end > > user takes less than a "substantial part" yet thinks > it is bound by > > the licence terms when (because it's taking less than > a substantial part) > > he/she/it is not. > > > > > > > > Those concerned by > > such apparent problems tend to suggest one of two > different solutions: > > > > > > > > > > The first solution, and the one which Science > > Commons advocates, is for the dataset owner to waive > all copyright and related > > works in the dataset, by using CC Zero or the older > (and US-centric) CC Public > > Domain Dedication and Certification tool. The use of > these tools largely > > assumes that it is legally possible to completely > waive statutory copyright in > > a copyright work when - at > > least in certain Commonwealth countries - that may > not in fact be the case (as argued > > by the UK's Phillip Johnson in "'Dedicating' Copyright > to the Public > > Domain" (2008) 71(4) Modern Law Review 587-610). In > the governmental > > context, it would also raise significant policy issues > around the propriety of > > waiving Crown copyright. > > > > > > The second solution is to advocate the use of an > > alternative to CC licences, i.e., a dataset-specific > licence that contains less > > restrictions. > > > > > > > > Our current view is > > that there is merit in advocating the use of CC > licences for all kinds > > copyright works that can be released for re-use, > including copyright datasets > > and databases. To advocate the use of separate > licences for datasets would run > > the risk of further licence proliferation which is one > of the current problems > > NZGOAL is trying to resolve. There is also the fact > that the CC model is very > > widely known and respected and benefits from the CC > infrastructure, including > > the three different forms for each licence, i.e., > human-readable, > > lawyer-readable and machine-readable. > > > > > > > > As you'll gather > > from my comments above, we do not support the CCZero / > full waiver of copyright > > approach, as it raises complex legal and policy > issues, the resolution of which > > is - in our view - unnecessary at this point. The > current approach in the draft > > NZGOAL is, if accepted by Cabinet, expected to see the > freeing-up of copyright > > works and datasets more quickly than an approach which > could be legally > > questionable and more complex from a policy angle. > > > > > > > > We have, > > nevertheless, been mindful of the potential for > attribution-stacking problems > > and we have received further feedback suggesting we > ought to take this > > expressly in account within NZGOAL. The approach we > have adopted at the moment > > is to add a new policy principle into NZGOAL, one > which would encourage > > agencies releasing datasets to think long and hard > about potential attribution > > stacking problems downstream and, where there is any > likelihood of that, to > > impose minimal attribution requirements on end users. > It is also perfectly open > > to agencies to completely waive all attribution > requirements whilst still > > retaining the other benefits of the CC licences. As > such, we consider that the > > CC framework already has sufficient flexibility to > accommodate potential attribution > > stacking problems. > > > > > > > > So far as the > > share-alike issue is concerned, NZGOAL cautions > agencies against its use. As > > you'll know, it recommends CC-BY as default and warns > agencies of the > > potentially stifling effect of using either > share-alike or no derivative > > variants. We may need to consider adding a paragraph > or two on this in the > > specific context of datasets. Something we'll ponder. > > > > > > > > The potential for > > category errors would remain with the CC model, as it > would for other copyright > > licensing models, in contrast to full waivers (if they > were in fact legally > > permissible). While it's a problem that may be more > common in the data arena, > > it can exist in other areas as well. It's really a > facet of the complexity of > > copyright law and not something, in our view, that can > or should be 'fixed' in > > the governmental context by adoption of CCZero or > other form of waiver. There > > may, nevertheless, be merit in the Crown releasing a > specific guidance on this > > topic in due course. Again, something we'll ponder." > > > > > > > > ******* > > > > > > > > Nat has reproduced Puneet's response above. Puneet > also > > prefaced his remarks by noting that he had (1) > digested the Philip Johnson > > paper, (2) had conversations with faculty at the > University of Wisconsin Law > > School, and (3) discussed matters Science Commons > counsel, Thinh Nguyen. We > > did, of course, find Puneet's comments both helpful > and reassuring. Again, many > > thanks Nat for putting us in touch. > > > > > > > > (2) CREATIVE COMMONS AND DATASETS > > > > > > > > I'm not sure it's entirely right to say Creative > Commons do not > > recommend their licences for datasets/databases. > Originally, it was quite the > > contrary. Their FAQs positively recommended CC > licences for copyright databases/datasets. > > It was Science Commons that moved to an alternative > model due to the three > > concerns set out above in the context of primarily > scientific/research data. > > However, as discussed above, NZGOAL seeks to address > two of the three concerns, > > whilst the third is an issue that is not unique to > datasets/databases, and > > Puneet agreed with the approach NZGOAL advocates. > > > > > > > > (3) OPEN DATA LICENCES > > > > > > > > The fact that data licenses like the Open Data > Commons > > licenses and CC0 may be designed with data in mind > doesn't mean that Creative > > Commons licences are not suitable for copyright > databases and datasets. I would > > also add that CC0 was not designed only for data; it > was designed for any > > copyright work. Its purpose is to achieve a waiver of > all copyright in a work > > and, if that fails, it has a fall back position of a > very broad licence to > > re-use (with no attribution or other requirements on > licensees). As noted > > above, it is not a part of the NZGOAL framework. > > > > > > > > (4) "DATABASE RIGHT" AND ALTERNATIVE DATABASE > > LICENCES > > > > > > > > There will probably always be competing views as to > whether > > CC licences or alternative database licences are more > appropriate for data and > > it's an issue we will follow with interest. These > things evolve and nothing is > > necessarily set in stone. I would urge caution, > however, when considering both discussions > > and alternative licences that are premised on concerns > arising from "the > > database right". The database right is a European > creation that has no > > equivalent under New Zealand law. > > > > > > > > (5) OTHER COMMENTS > > > > > > > > It is fair to say that copyright and licensing issues > around > > datasets and chunks of data are not straight-forward. > It is also correct to say > > that, while datasets and databases can qualify as > copyright works, individual > > data items seldom do. > > > > > > > > To infringe copyright in a copyright dataset or > database, > > ordinarily one needs to establish that the alleged > infringer copied all or a > > "substantial part" of the dataset/database. This > raises interesting > > issues. If, for example, a mashup were to take only a > small/insubstantial > > amount of data from multiple publicly available > datasets, it may not require > > any licence at all. > > > > > > > > As I say, there are some complex issues here (and > nothing > > I've said should be taken as legal advice; sorry - > regulatory requirement). It is also an area that is in a > state of flux, internationally, in the courts. There has, > for example, been recent litigation in Australia which, on a > first read, restricts the copyright protection of databases, > even where significant effort has gone into their creation. > Last time I checked, the Australian case I have in mind was > on appeal. So to some extent it's a question of 'watch this > space'. > > > > It > > may well be that the various issues in this area > should be the subject of a > > discrete guidance note in due course. > > > > > > > > Kind regards > > > > Richard > > > > > > > >> Date: Fri, 16 Jul 2010 11:08:18 +1200 > >> Subject: Re: [ninjas] creative commons licenses > >> From: <email obscured> > >> To: <email obscured> > >> > >> On Fri, Jul 16, 2010 at 3:50 AM, Nathan Torkington > <email obscured>> > wrote: > >> > >> > On 15/07/2010, at 8:29 AM, Jonathan Gray > wrote: > >> > > Generally I think Creative Commons do > not recommend their content > >> > > licenses for databases. > >> > > >> > When Richard was working on the NZGOAL > recommendations, I specifically > >> > queried this. > >> > >> > >> > >> I asked this too. Richard said that under NZ law, > CC licenses could apply to > >> data much more clearly than in other jurisdictions > where it all becomes a > >> grey area (ala ODbL and other solutions that are > being worked on). > >> > >> Rob :) > >> > >> ----------------------------------------- > >> Full text of this topic in The Open Government > Ninjas: > >> http://groups.open.org.nz/r/topic/6mLVpvvdMfENK8wfrECjXZ > >> > >> To leave The Open Government Ninjas, email > >> <email obscured>?Subject=unsubscribe > >> > >> Start your own free groups and site with > >> OnlineGroups.Net http://onlinegroups.net > >> > >> Host your own online groups site with > >> GroupServer http://groupserver.org > > > > > _________________________________________________________________ > > Hotmail: Powerful Free email with security by > Microsoft. > > https://signup.live.com/signup.aspx?id=60969 > > ----------------------------------------- > > Full text of this topic in The Open Government > Ninjas: > > http://groups.open.org.nz/r/topic/7KQ4cSHMKPywcmM0W1w3gD > > > > To leave The Open Government Ninjas, email > > <email obscured>?Subject=unsubscribe > > > > Start your own free groups and site with > > OnlineGroups.Net http://onlinegroups.net > > > > Host your own online groups site with > > GroupServer http://groupserver.org > > > > > > -- > Jonathan Gray > > Community Coordinator > The Open Knowledge Foundation > http://blog.okfn.org > > http://twitter.com/jwyg > http://identi.ca/jwyg > > ----------------------------------------- > Full text of this topic in The Open Government Ninjas: > http://groups.open.org.nz/r/topic/4J37AUTGFbdVqki7vqWvpv > > To leave The Open Government Ninjas, email > <email obscured>?Subject=unsubscribe > > Start your own free groups and site with > OnlineGroups.Net http://onlinegroups.net > > Host your own online groups site with > GroupServer http://groupserver.org >
Loading…
Privacy | Acceptable Use | Terms of Service | About OnlineGroups.Net | Contact OnlineGroups.Net
Start an OnlineGroups.Net site for easier email collaboration in your organization.
Powered by GroupServer, the open source web-based mailing list manager.