OUseful.Info, the blog…

Trying to find useful things to do with emerging technologies in open education

The Closed Route to Open Data

A couple of weeks ago, I gave a presentation to the WebScience students at the University of Southampton on the topic of open data, using it as an opportunity to rehearse a view of open data based on the premise that it starts out closed. In much the same way that Darwin’s Theory of Evolution by Natural Selection is based on a major presupposition, specifically a theory of inheritance and the existence of processes that support reproduction with minor variation, so too does much of our thinking about open data derive from the presupposed fact that many of the freedoms we associate with the use of open data in legal terms arise from license conditions that the “owner” of the data awards to us.

Viewing data in this light, we might start by considering what constitutes “closed” data and how it comes to be so, before identifying the means by which freedoms are granted and the data is opened up. (Sometimes it can also be easier to consider what you can’t do than what you can, especially when answers to questions such as “so what can you actually do with open data?” attract the (rather meaningless) response: “anything”. We can then contrast what you can do in terms of freedom complementary to what you can’t…)

So how can data be “closed”?

One lens I particularly like for considering constraints that are placed on actions and actors, particularly in the digital world (although we can apply the model elsewhere) I first saw described by Lawrence Lessig in Code and Other Laws of Cyberspace: What Things Regulate: A Dot’s Life.

Here’s the dot and the forces that constrain its behaviour:

4constraints

So we see, for example, the force of law, social norms, the market (that is, economic forces) and architecture, that is the “digital physical” way the world is implemented. (Architecture may of course be designed in order to enforce particular laws, but it is likely that other “natural laws” will arise as a result of any particular architecture or system implementation.)

Without too much thought, we might identify some constraints around data and its use under each of these separate lenses. For example:

  • Law: copyright and database right grant the creator of a dataset certain protective rights over that data; data protection laws (and other “privacy laws”) limit access to, or disclosure of, data that contains personal information, as well as restricting the use of that data for purposes disclosed at the time it was collected. The UK Data Protection Act also underwrites the right of individuals to claim additional limits on data use, for example the rights “to object to processing that is likely to cause or is causing damage or distress to prevent processing for direct marketing; to object to decisions being taken by automated means” (ICO Guide to the DPA, Principle 6 – The rights of individuals).
  • Norms: social mores, behaviour and taboos limit the ways in which we might use data, even if that use is not constrained by legal, economic or technical concerns. For example, applications that invite people to “burgle my house” based on analysing social network data to discover when they are likely to be away from home and what sorts of valuable product might be on the premises are generally not welcomed. Norms of behaviour and everyday workpractice also mean that much data is not published when theere are no real reasons why it couldn’t be.
  • Market: in the simplest case, charging for access to data places a constraint on who can gain access to the data even in advance of trying to make use of it. If we extend “market” to cover other financial constraints, there may be a cost associated with preparing data so that it can be openly released.
  • Architecture: technical constraints can restrict what you can do with data. Digital rights management (DRM) uses encryption to render data streams unusable to all but the intended client, but more prosaically, document formats such as PDF or the “release” of data charts are flat image files makes it difficult for the end user to manipulate as data any data resources contained in those documents.

Laws can also be used to grant freedoms where freedoms are otherwise restricted. For example:

  • the Freedom of Information Act (FOI) provides a mechanism for requesting copies of datasets from public bodies; in addition, the Environmental Information Regulations “provide public access to environmental information held by public authorities”.
  • the laws around copyright relax certain copyright constraints for the purposes of criticism and review, reporting, research, teaching (IPO – Permitted uses of copyright works);
  • in the UK, the Data Protection Act provides for “a right of access to a copy of the information comprised in their personal data” (ICO Guide to the DPA, Principle 6).
  • in the UK, the Data Protection Act regulates what can be done legitimately with “personal” data. However, other pieces of legislation relax confidentiality requirements when it comes to sharing data for research purposes. For example:
    • the NHS Act s. 251 Control of patient information; for example, the Secretary of State for Health may “make regulations to set aside the common law duty of confidentiality for medical purposes where it is not possible to use anonymised information and where seeking individual consent is not practicable” (discussion). Note that they are changes afoot regarding s. 251…
    • The Secretary of State for Education has specific powers to share pupil data from the National Pupil database (NPD) “with named bodies and third parties who require access to the data to undertake research into the educational achievements of pupils”. The NPD “tracks a pupil’s progress through schools and colleges in the state sector, using pupil census and exam information. Individual pupil level attainment data is also included (where available) for pupils in non-maintained and independent schools” (access arrangements).
  • the Enterprise and Regulatory Reform Bill currently making its way through Parliament legislates around the Supply of Customer Data (the “#midata” clauses) which is intended to open up access to customer transaction data from suppliers of energy, financial services and mobile phones “(a) to a customer, at the customer’s request; (b) to a person who is authorised by a customer to receive the data, at the customer’s request or, if the regulations so provide, at the authorised person’s request.” Although proclaimed as a way of opening up individual rights to access this data, the effect will more likely see third parties enticing individuals to authorise the release to the third party of the individual first party’s personal transaction data held by a second party (for example, #Midata Is Intended to Benefit Whom, Exactly?). (So you’ll presumably legally be able to grant Facebook access to your mobile phone records… Or Facebook will find a way of getting you to release that data to them without you realising you granted them that permission;-)

Contracts (which I guess fall somewhere between norms and laws from the dot’s perspective (I need to read that section of Lessig’s book again!) can also be used by rights holders to grant freedoms over the data they hold the rights for. For example, the Creative Commons licensing framework provides a copyright holder with a set of tools for relaxing some of the rights afforded to them by copyright when they license the work accordingly.

Note that “I am not a lawyer”, so my understanding of all this is pretty hazy;-) I also wonder how the various pieces of legislation interact, and whether there are cracks and possible inconsistencies between them? If there are pieces of legislation around the regulation and use of data that I’m missing, please post links in the comments below, and I’ll try and do a more thorough round up in a follow on post.

Written by Tony Hirst

February 19, 2013 at 10:03 am

Posted in Data, Policy

Tagged with

5 Responses

Subscribe to comments with RSS.

  1. You write: “Market: in the simplest case, charging for access to data places a constraint on who can gain access to the data even in advance of trying to make use of it. If we extend “market” to cover other financial constraints, there may be a cost associated with preparing data so that it can be openly released.”

    It’s one of the troubling things about Open Data that someone has to pay for the infrastructure to make it available. While I generally agree that many if not most attempts by public bodies to pay for that infrastructure by charging for access probably fail to meet their aims (as well as closing down use), opening up access may greatly increase use, adding un-recompensed costs onto the data provider.

    How do we persuade them to pay these costs?

    Chris Rusbridge

    February 19, 2013 at 12:11 pm

    • @Chris I think one of the ways of covering publishing costs is to make use of the published data. That is, if public bodies use opendata pathways as part of data-related workflows, some of the costs of publishing open data are replaced as costs of doing day-to-day business.

      Costs are also incurred by users of opendata, for example cleaning it, finding and reporting errors etc. Using a ‘many eyeballs’ argument, if public bodies use data that have an open window on to them: 1) the data is more likely to be useable/workable with (because people are using it/working with it); 2) if there are errors/problems with the data, there are more opportunities for them to be picked up; 3) given 1 is a lie, and folk use broken data and have to repair it themselves, if third parties help repair the data, the publishing organisation can get round some of the workarounds folk have to use to cope with working with broken data ;-)

      One view I take is that a primary user of opendata could be the publisher – if the data is not of interest/use to them, why would it be of interest to anyone else. (This feeds the transparency agenda in context of seeing what data folk are actually making decisions on; and innovation agenda by supporting *additional* uses and combinations of data.)

      This is of course all my own very idealised way of how data workflows do/don’t/could/should work. I need to get out into the real world more to see how things really do work… which is on my to do list…

      Tony Hirst

      February 19, 2013 at 12:38 pm

  2. The main other piece of legislation that deserves a mention is the European Directive on the re-use of public sector information (the “PSI Directive”), enacted into UK law as the Re-use of Public Sector Information Regulations 2005.

    FOI and EIR are crucial for providing access to public sector information, but until very recently did not mention licensing for re-use of that information. At the moment it’s unclear whether the recent additions to FOI will encourage more open data releases or more charging for data.

    The PSI Directive is the source of the presumption that public sector datasets should be available for re-use either for free or at marginal cost, so is quite important to the information rights argument underpinning the open data agenda.

    The European Commission is currently considering proposals to strengthen the PSI Directive, and UK MEPs are attempting to water down those proposals in order to protect existing commercial carve-outs (i.e. trading fund arrangements such as Ordnance Survey and Land Registry).

    • @owen Ah – yes – thanks for that. The point about the on-licensing of FOId data is also well made. Ta muchly:-)

      Tony Hirst

      February 19, 2013 at 12:39 pm

  3. [...] PS Thinks: when I download and use a software tool, accepting terms and license of use conditions as I do so, under what legal framework does the tool publisher claim ownership of the software application and the right to issue and enforce the license conditions or terms of use? Copyright applies to the source code, and I guess the compiled code? So to the extent that a musician may own copyeight over an expression (recording) of a song, maybe a software publisher owns copyright in the executable version of a piece of software, and running that application is like listening to a piece of recorded music. And just as you are limited in the extent to which you can play recorded music (not to a large audience, for example?) presumably similar conditions can be made to apply to the use to which you put a piece of software? (See for example, The Closed Route to Open Data.) [...]


Comments are closed.

Follow

Get every new post delivered to your Inbox.

Join 800 other followers

%d bloggers like this: