Report from the ChangeCamp Ottawa 2010 Session C2
July 17, 2010
Ottawa City Hall
Session Moderator: Tracey P. Lauriault (firstname.lastname@example.org, datalibre.ca, civicaccess.ca)
Note Taker: Mark Tovey, (email@example.com, WorldChanging Canada )
Input provided by:
Tracey P. Lauriault approached David Fewer, the director of Canadian Internet Policy and Public Interest Clinic (CIPPIC), about discussions from users and about the topic of access to public data in general. David Fewer agreed to have a summer law student conduct research on the issue of data licensing and ToU. Concurrently, the four open data cities have created a series of working groups between and among themselves to improve and advance their Open Data work. The City of Ottawa has agreed to be the point city on the issue of ToU.
Today’s session is about examining the existing City of Ottawa ToU and soliciting input from ChangeCamp 2010 attendees as part of the research process. The research will be ongoing during the summer and it was agreed that those who have provided input will be kept in the loop. In addition, there will be further consultation with the open data cities and potentially with the broader public who may have an opinion. The discussed output from this research may be in the form of advice to the cities, a modified ToU with annotations, or a CIPPIC FAQ. The Final output has not yet been decided.
- Share thoughts and comment on its content
- Mark Tovey took notes
- Tracey P. Lauriault wrote notes on the Flip Chart paper
Your open license to the datasets
- Operationalizing this is difficult particularly when there are many licenses in many jurisdictions
- License proliferation is considered problematic
- A lit of compatible licenses could be helpful
- Include a clause that would make this ToU compatible with inherent license classes, where ‘this or that shall prevail’ type of language
- Consider Open Data Commons or ODbl
- This makes derivative works difficult, e.g. what if a user creates a model using these data
- Attribution should not be a must
- This clause may be illegal
- Derivatives works license?
- Open data is impeded with this paragraph as a user cannot guarantee that the next user of the dataset or the derivative work can/will comply
- It was considered unreasonable, the original user can only be responsible for themselves and what if it was an underage user?
- One way to work with this is to make sure that everyone is using the same license. [laughter]. OR you can have a clause saying that this license is also compatible with these–and then you set specific licenses that you’re also permitting.
- OR Contact us if you have specific requirements around other licensing. OR absolutely you can’t use these licenses.
- (later someone said that it was helpful to specify particular licenses. You could always add new compatible licenses, but it would create problems if you kept things open-ended).
- What about mashups with many data sources and different licenses
- There’s the Open Data Commons, who have developed a couple of licenses – so really another side a is to encourage open data cities to use the same licenses.
- If you derive something from the data do you have to redistribute it under the same terms?Could be: if I’m doing something with the data it could have a different representation, then what
- With derivative works / mashup you would get a license on a license on ca license
- You could call it derivative work license in – e. g. model based on Ottawa data.
- Robin: Does attribution–does copyright law–imply that you must have attribution and the license can’t override? Are data copyrightable?
- Scientific communities are moving away from attribution, because of the mashing up multiple data sets such as in a climate change model. They may however site data sources.
- City can update the list of licenses that derivatives could be built under.
- The unrestricted user licenses has issues.
- Obama has done a lot of good stuff in terms of putting stuff into the public domain – so for example it can be Wikipedia is a great example. There’s no crown copyright – no attribution necessary.
- In the US government data are a public record and are in the public domain, there is a Westminster system in Canada which comes with Crown Copyright
- What this is saying then, that, in the end if I redistribute anything – I have to be sure that others agree to the terms?
- This is a kind of an impossibility: there’s really no point in saying something that’s illegal
Giving City credit
- To enable attribution, provide attribution information tags in XML
- Microattribution – metatags & XML attribution –
- Often this ideas does not work in practice because attribution is desired precisely as the dataset are.
- Data Citation may help
- When developing ToU, it is important to think in the very long term, for instance 50 years from now what could be the implications?
- Increase the limitations on the City’s ability to arbitrarily change the ToU
- Have a process in place to discuss ToU changes
- Notice of changes to the ToU on the City website is considered inadequate
- Devise a better way to advise on the changes
- Split paragraphs on changes to the datasets and to the ToU, have a paragraph on Changes to the Data and a separate one on Changes to the ToU
- A dataset versionning process can provide the means for users to know when changes have been made
- RSS feed of changes to users for dataset changes
- Feed of Feed on the datasets
- Provide an archive of previous data versions, recognizing that some data dynamically change and this may not be possible on a per change basis
- Ensure basic Metadata are provided
- Data Time Stamping
- Changes should be annotated
- Include a sunset clause whereby after a certain period of time the data go into the public domain.
- Create an archive
- What if – if it’s they’re responsibility – if the changes are effective immediately
- A consultation on the ToU could be useful
- You can’t force people who want to sign up to an RSS feed to do so.
- Imagine the lens on the distant future.
- What of sunset clauses – the information goes into the public domain after 50 years or something. This is a different thing than copyright.
- In some way it’s more related to archives.
- If you need some argument to go against the lawyers – this is extremely easy to game – all I have to do is add a 5 to the end of each data point – to make a derivative work a bot – it actually makes no sense for that term to exist.
- If you remove sole discretion – it does not give you a right to challenge.
- This is only justified is if someone has breached the licenses, but what are the types of breaches?
- Do not bring in morality clauses – harmful – even if it’s a w omen’s collective that sets up open bordello.com – it may still be considered harmful, who gets to decide what is and is not.
- This could be handled without cancellation, non-compliance.
- Address without cancellation
- Do not include morality, if it’s against the law, it’s against the law, and to avoid being sued the city would want the ability to go after you.
- An explanatory note, with a worked example. ie. so you can’t enable robmyhouse.com will help clarify this clause
- Allow recourse, for instance if a user corrects whatever the problem is, it will be restored automatically.
- The problem is whether or not a user notices the notice of changes, at the moment the method may not be sufficient and people may not be complying because they were not aware of the notice of change
- With this clause, at the moment, you could revoke something – so 15 years down the line all the derivative works have to be revoked too?
- Also need to consider when data are depreciated / archived
- Need a way to export the whole data – so you have a way to have a snapshot.
Compliance with law – Your responsibility
You assume sole responsibility for your use and reproduction of the datasets complying with all applicable laws and industry standards.
No warranty with datasets
You get NO WARRANTIES, none of any kind. By this, we mean, for example (but without limiting the total intended scope of the preceding sentence), (1) that while reasonable efforts have been made in preparing these datasets for use by you, the City cannot give any promises as to the completeness, currency, or accuracy of the datasets nor that access will be continuous, (2) the City cannot make any promise that the datasets are free and clear of any possible third party copyright, moral rights, or other claim, (3) the datasets have been modified from their original source, as data initially generated by the City for its internal uses, and (4) all data visualizations on maps are approximate and include only records that can be mapped.
You accept these datasets on an “as-is, where is” basis and agree to use them at your own risk.
Exclusion of liability
You agree that you will not and cannot sue the City for anything, which the City does or does not do (even if intentional or negligent) in connection with the datasets and your use or inability to use them. Without limiting the general scope of the preceding sentence, this means that the City and its agents are not liable on any legal theory or basis for any direct, incidental, indirect, special, punitive, exemplary, or consequential damages or losses, including without limitation, loss of revenue or anticipated profits, loss of goodwill, loss of business, loss of data, computer failure or malfunction, or any other damages or losses.
- This was not discussed, but some related information were discussed under different headings
Cancellation for non-compliance
- This was considered problematic and unreasonable
- Ensure there is a process and criteria in place
- Add explanatory notes
- What reasons could the city have to cancel or suspend access to a dataset?
- It was considered ‘easy to game’ this clause, so why bother having it?
- Remove the language around ‘sole discretion’
- Ensure there is process for recourse
- Morally should not creep into ToUs
- Errors in the data or a mistake in a breach of personal privacy are considered to be reasonable issues to ‘take back’ the data or not permit its use
- This could be addressed without cancellation, for instance, if there is a breach in the ToU then provide the user the time to rectify the problem
- Important to have an explanatory note and a clear example as to why the city would want to do this
- This is open to arbitrariness
- Remove credit – they can decide that they don’t want to be cited any more. What’s that all about?
- But what if you’ve got a great mashup of the really cool progressive sex shops – the city may decide that it does not mind you doing that but they may not want to be associated with it.
- I wonder if a possible solution could be that the city could make a request to remove credit for reason of public decency or public interest
- Public decency is a slippery slope – whenever possible remove morality and what remind ourselves that what is against the law is against the law.
- Agree, for instance, if you’re doing something scientific – let’s say a climate change model in 2100 – and let’s say science becomes a religious issue, would you want that model suppressed for morality reasons?
- Users want to be able to cite data to demonstrate they are using valid and reliable data from producer side.
- One of the goals is to make this license somehow exemplary. One thing that could be awesome would be to add explanatory notes to your licenses For instance “I want you to revoke your moral rights’ – but what did that mean? It meant that if we wanted to sell the code to a nuclear power plant you couldn’t revoke your moral rights. So then I understood.
- In the same way that a court will consider discussion in committee in figuring out how something will be interpreted. make discussions available.
You may not publicly represent or imply that the City is participating in, or has sponsored, approved, or endorsed the manner or purpose of, your use or reproduction of these datasets.
- Concern about the reasoning the City would include this.
- As an alternative, the City could ask that credit be removed but that the data user can reference the dataset’s source
- Validity – referencing provides the means to ensure that the user of the data are using reliable data from a credible source
You may not use any trade-mark, official mark, official emblem or logo, of the City, or any of its other references or means of promotion or publicity without the City’s prior written consent nor in any event to represent or imply an association or affiliation with the City.
Governing law and jurisdiction from which datasets are published
- Not discussed however there was mention an earlier question as to whether or not copyright applies to data
- A new ToU should include the formal language and be accompanied by explanatory notes with tangible examples.
- Disseminate the ToU in a way enable discussion e.g. comment-able in a wiki
- Consider the Creative Commons model which combines both plain language and legal language
- Consider either CCO (http://creativecommons.org/publicdomain/zero/1.0/) or ODbl (http://www.opendatacommons.org/licenses/odbl/)
- Develop and/or adopt a ToU with Open Government principles
- Provide the ability to cite these data, this is important when data are used in academic or other publications
- What about CC0 that’s public domain.
- Can you copyright a fact? – Teressa Scassa, Canada’s Research Chair in Information Law & Vice-Dean Research has written about this and id doing research with geomaticians on this topic . Sometimes you can copyright facts.