When a government is looking at opening up data, how should it be? How does one define open-ness? In our submission to the Indian government on its Open Data License yesterday, we briefly explored this idea, and outlined a philosophy for openness, which I suspect we will use in subsequent submissions as well.
- The philosophy of open-ness: An open data license must be rooted in the philosophy of open-ness: that information must be made available in a manner that is transparent, free, non-discriminatory, neutral, unconditional, and without prejudice.To this end, it is essential that the data, once released under this license, imposes no restrictions on both the users ability acquire the data and its usage, based on:
- The identity of the entity using the information.
- The purpose of the usage of the information.
- The interpretation of the information
- How people react to the information once disclosed and/or reused, remixed or repurposed.
This openness allows citizens to remix, bundle and unbundle data and information, and reimagine the way the information may be understood or used by others. Openness thus needs to lead to creativity. To enable this, we need low friction, predictable systems, without having to negotiate agreements with the creators of the data or information.
We also reiterated a demand that government funded data, software and content must be open by default, a stand that we had taken when the Copyright Act was being amended. The rest of our submission:
- Openness not just limited to government: The idea of an open data license shouldn’t only be limited to the government releasing this information: the objective should be to foster the idea of anyone releasing their collected information, especially academics, resulting in analysis and interpretation, with the idea of spreading knowledge. To prescribe terms for an Open Data license merely from the perspective of the government would limit its usage to only government bodies and officials. Thus, we would recommend using “Open Data License”, instead of “Government Open Data Use License”.
- Commercial and non-commercial: No restriction should be placed on reuse of the data for commercial purposes, whether it is commercial sale of the exact same data set, or versions of this data. There is value add possible, in terms of even collating data sets relevant for a user base, not just in their presentation.
- Data from RTI, on all government and regulator websites, and government funded academic research: All data released under the Right to Information Act, or disclosed in Parliament, disclosed by all governments, State or Central, and by all regulators, in any medium, whether on their website, in any digital or physical format, as well as academic reserach funding by the government must automatically and compulsorily be under an open data license, since the data is funded by and owned by the taxpayers, i.e. the citizens of India.The idea of an open data license should not be used to exclude certain publicly available information from usage, by allocating the license to only limited sets of data and information.
Section specific comments
- #Section 3
Open data must be open access: that is, no identification must be necessary for the usage of open data. It must be available without a login, without the need for identification, or the need for an email address.
- #Section 4
Liability and attribution:
- Liability of data provider and user: the transfer of liability for the data set from the creator to the user of the data, will significantly limit the usage of data sets. The idea of open data is to allow remixing, repurposing and interpretation of data, with the explicit understanding that responsibility and liability for any issues with the data itself lie with the creator and/or the publisher of the data, upon whom the responsibility of verification lies.
- The need for attribution: Attribution is beneficial for identifying the source, but should not be a necessary condition for usage of open data. This appears to be a restrictive condition on the usage of data, and hence not in consonance with the philosophy of open-ness (point 1).The idea behind attribution is the transfer of liability to the source information, and a courtesy, but little else. If someone wishes to an interpretation of the data, without attribution, they should have the freedom to do so, given that, potentially, the liability for false data may also transfer to them if they don’t attribute the source. Even then, without explicit publicly available attribution, the liability of any discrepancies in data must lie with the source of the data, not with the end user. The liability of the user must only be limited to her value add to the data. In addition, if someone wishes to issue data without an attribution requirement, they should be allowed to do that.Thus, we oppose the idea of a compulsory attribution, and the format shared as a part of the Open Data license.
- #Section 5
The mandated template for attribution: Mandating a particular format for attribution is also a precondition for usage of data, which is also against the philosophy of open-ness (point 1.1). The format suggested is too long, and does not lend itself to visualizations:“Data has been published by Ministry of Statistics and Programme Implementation and sourced from Open Government Data (OGD) Platform of India: Overall Balance of Payments. (08/09/2015) . https://data.gov.in/catalog/overall-balance-payments. Published under Open Government Data License – India: [URL of Open Data License – India].”The format used for attribution, if chosen to be provided by a user, must be left to the user, and not mandated by the government.
- #Section 4d
Violation of laws: It is unreasonable to expect each citizen to be aware about whether a data set used violates the RTI Act, trademarks, official marks, patents etc. That’s an onerous condition. The liability for the release of the data set must lie only with the source of the data, not its user. As mentioned earlier, the liability for a user must only be limited to the value add, and only if applicable to the value add itself.
- #Section 4e
Hosting data in perpetuity: The requirement for hosting the source data in perpetuity is also a condition which may deter the release of the information. Data can often become quickly outdated or irrelevant. The requirement that the data needs to be hosted in perpetuity is a precondition that goes against the philosophy of open-ness. Given that we have recommended that any user should be allowed to release their data under this license, we need to be flexible to the ability of users to host data, and the cost of hosting the data in perpetuity: this can also be a condition that prevents the releasing of data. The license needs to be a bit more flexible, to encourage use: hosting data for perpetuity must be a recommendation but not a precondition.
- #Section 6The open data license should be in consonance with the Right to Information act, and the only exemptions applicable must be those that are a part of the Right to Information Act. If any the exemptions are not covered by the Right to Information Act, then should be excluded. It would be unusual if information is available under the Right to Information act, and yet cannot be released under the open data license.
Openness should not just be limited to data: Instead of limiting the license to data, we would recommend the creation of an open license, applicable to content and computer software technologies created by entities owned entirely by the Government of India. This content and technology is created with funding from the people of India, yet much of it is treated as if owned and meant for for commercial exploitation by the Government of India. As we had indicated in our submission to the government on the Copyright Amendment bill:
“the intellectual property created by the Government of India is the intellectual property of its people, and all citizens of India should have the right to access it, having already paid for it.”
In addition, we believe that this information should also be available for them to use and remix, as they deem fit, for commercial or not commercial purposes.
Just as with data, research work done by the government of India, related to digital technologies, would benefit society more if it were freely and openly available, since it can be used for further development. There is precedence in the Linux Operating System, which a collaborative community used as a base for creating free and open source computer operating systems, thus creating an alternative to expensive, proprietary software.
It is pertinent to note that works created by the judiciary i.e. judgments are not subject to copyright. The same option must be available for government data, technology and software, if not mandated.
Download MediaNama’s Open Data license Filing