Opening Public Data in South Africa: legal complications

Written by Andrew Rens on April 15th, 2013

Adi Eyal, data  experiences opening public data in South Africa asks government should claim copyright over  knowledge funded by taxpayers.

Describing his experiences dealing with the Municipal Demarcation Board, the National Statistics Agency and the Electoral Commission,  provincial and municipal administrations he reports that many of these government offices claim that the databases they administer are subject to copyright and even deny permission to re-use the data.

I have bad news for Adi, the practises that he is tracking are layered onto an even more complicated and contradictory  legal situation.  To appreciate the complex legal rules bearing on open data in South Africa one needs to consider the interaction of four five distinct sets of legal rules.

The Bill of Rights grants everyone rights to information, especially information held by the State. The Promotion of Access to Information Act creates mechanisms for the enforcement of the right information held by the State. The Copyright Act dates to 1978, long before the Constitution or the rise of Big Data. Which aspects of of data are subject to copyright? In what circumstances does copyright over data vest in the State? Is the 1979 delegation of control of state copyright to the Government Printer permissible under the 1996 Constitution especially the provisions governing the powers of provinces and municipalities? How does the legal duty of co-operative government, including information sharing, affect the duty of government offices to share data? Are databases subject to the State Information Technology Agency Act 1998? Should the Ministerial Interoperability Standards specify data formats?

The interaction of these different legal and policy instruments that creates uncertaintly for government offices is not the result of design but of a contingent process in which the interacting factors developed in different contexts but now must be understood together. South Africa could benefit from open government data, and is in some exceptional cases already benefiting. But if the successul experiments are to scale the legal complications need to be addressed.


2 Comments so far ↓

  1. Adi Eyal says:

    Thanks for writing this article. It’s useful that you’re putting the issue in context.

    My understanding of the copyright act is that it does not directly cover data. Data can possibly be interpreted under the literary work definition. In any case, to qualify as a literary work, the data needs to include some component of creativity or innovation.

    Facts are not applicable under copyright. Databases of facts may be, depending on whether the process of the selection of the facts to be included underwent a creative or innovative process. For example, I don’t think that the telephone book qualifies for copyright protection. On the other hand a list of names of the 10 most powerful leaders in South Africa may, as the process of creating that list is innovative.

    Interestingly enough, the act explicitly excludes official texts from copyright but doesn’t specifically define what an official text is. Arguably, most government data can be considered to be an official text.

    PAIA defines the process through which data can be requested but does not deal with ownership of the intellectual property as far as I understand. It is not clear what my rights are with regards to data received through the PAIA process.

    The State Information Technology Act includes a bizarre clause that vests all intellectual property in SITA.

    Legal clarification is important but shouldn’t stop users of data from continuing their work. Legal action from government agencies is highly unlikely and positive examples of how data is being used to produce information products and improve citizen engagement with government will only improve the case for making more data open.

  2. Andrew Rens says:

    Thank you for the comment Adi

    First let me agree with you that the need for legal clarification should not stop users from continuing their work, nor government from discovering the potential of open data.

    Of course I can’t comment on whether legal action from government is likely.

    But that doesn’t mean that people can’t engage with government, get access to data as they are able, and demonstrate the desirability of open data.

    As you know the Copyright Act came in to force in 1978, BIG DATA hadn’t been imagined yet. The Act includes in the definition of “literary work” the following: “(g) tables and compilations, including tables and compilations of data stored or embodied
    in a computer or a medium used in conjunction with a computer,”.

    Whether a compilation of data falls under copyright depends on the same three factors that affect all South African works; it must be original, reduced to a material form and made by a qualified person. According to the Act copyright may adhere in the state as a qualified person. Storage as ‘digital data’ is enough for a work to be regarded as in material form. For a work to be regarded as original it must be the product of independent production i.e. not a copy of another work.

    What does that mean for databases? South African law is not at all clear on this. South African law has not adopted the standard of requiring creativity found in the Feist decision in the US, nor even the originality by dint of selection and arrangement of the European Union Database Directive.

    Most of the cases dealing with the relevant section don’t deal with computerised databases but the one leading case on computerised databases Payen v Bovis dealt with a work created under the 1965 Act. That case states that a computer generated database is subject to copyright where it is the result of independent skill and effort.

    The vesting of not all but ‘technological’ intellectual property in SITA is matched by the appointment of the Office of the Government Printer to manage copyright that vests in government.

    But there are many statutes that appoint Registrars of particular ‘registers’ (by now all computerised databases) which it could be argued grant those Registrars the responsibility to manage the copyright in the registers (assuming that there is any copyright in the register).