Post: 9 December 2013
The problem of licence proliferation in open data has been highlighted in several recent posts, notably one by Leigh Dodds and a couple by Mike Linksvayer.
In this post I want to look at some examples of UK data publishers who ostensibly apply the Open Government Licence (OGL) to re-use of their datasets, but then add their own terms to effectively create a separate licence.
Leigh’s post has already covered the open data licences used by Ordnance Survey and Natural England. Those licences are less than ideal in that they are slightly more restrictive than the OGL. However they do at least conform to the Open Definition.
The six examples below are a bit different. In each case the data publisher has started with the Open Government Licence, but modified it in some way that puts the terms of re-use outside the Open Definition.
What looks like open data at first glance may not be on closer scrutiny …
UK Hydrographic Office
UKHO’s INSPIRE Portal And MEDIN Bathymetry Data Archive Centre makes a range of datasets available via a geospatial viewer. Although the viewer is mainly designed to visualise the datasets, users can also download the bulk data in a range of formats. (You might need to faff around with the interface a bit.)
According to UKHO’s licence terms the datasets are available under the Open Government Licence. However, additionally:
1. The data sets must not be used for navigation or in the creation of navigational products.
2. If you reuse any of the data, you must acknowledge the source by including the following attribution statement: “Contains public sector information, licensed under the Open Government Licence v2.0, from [the Owner/Originator of the data as indicated in the metadata]”.
The second point is fine; the OGL allows and encourages the data publisher to specify an attribution statement, so that isn’t a material change to the terms of the OGL.
However the first point is an additional term that restricts re-use of the datasets, beyond the scope of restrictions already included in the OGL. It also puts the datasets outside the Open Definition, because the Open Definition says a licence “must not restrict anyone from making use of the work in a specific field of endeavour.”
I’m sure this additional term is well meant. If the UKHO doesn’t think these versions of their datasets are suitable for navigation purposes then I would take that advice seriously.
However the same point could be expressed as an information warning and included in documentation with the datasets. It doesn’t need to be a licence term.
Coal Authority
The Coal Authority (a non-departmental public body of the DECC) makes some spatial datasets available via web mapping services. According to the terms and conditions of use:
The data is supplied under the terms of the Open Government Licence subject to the following overriding re-use condition:
Re-use is not permitted for any activity that is part of the Coal Authority’s public task responsibilities.
As in the previous example, this additional condition creates a new licence because it modifies the OGL, and also falls outside the Open Definition because it restricts re-use of the data in a specific field of endeavour.
Unlike UKHO, the Coal Authority does not have the excuse of good intentions. The Coal Authority has a sophisticated commercial data licensing operation and the “overriding re-use condition” clearly operates to discourage competition.
Most of the datasets available via the Coal Authority’s web mapping services are, in any case, not open data because they cannot be downloaded. It’s a bit difficult to understand why the Coal Authority has bothered to invoke the OGL at all.
Higher Education Statistics Agency
HESA publishes the Unistats dataset, which includes the Key Information Set (KIS) and other data on university and college courses across the UK. The landing page for this data says:
These data are made available to third parties through the Open Government Licence. Anyone wishing to use the data for their own purposes must follow the guidelines for appropriate use as explained in this licence.
However when you click through to download the data there’s a whole ‘nother set of terms and conditions.
While these terms include analogues of the terms in version 1 of the OGL, HESA has made extensive additions. Probably of greatest significance is the following:
Your access and use of the Unistats Dataset is subject to your acceptance of, and compliance with, the provisions of the terms and conditions below, HESA’s Website Privacy Policy set out at http://www.hesa.ac.u/privacy and any other legal notices and/or instructions which may appear on http://www.hesa.ac.uk/unistatsdata (“this Website”) from time to time …
That clause seems to give HESA the right to vary the conditions for ongoing re-use of the data (and not just to create new conditions for re-use of data downloaded subsequently), which conflicts with the “perpetual” grant of licence in the OGL.
Spogo
Spogo is “a not-for-profit site funded by a lottery grant awarded to ukactive by Sport England”. The site includes a developer area with two groups of dataset available for download: ActivePlaces data under the Active Places Open Data Licence and spogo data under the spogo Open Data Terms and Conditions.
Neither of these are actually open data licences.
The Active Places licence “incorporates” the Open Government Licence “as amended by” a shopping list of additional terms “specific to Sport England”. The clearest breach of the Open Definition is the “responsible use” clause, which requires licensees to refrain from using the data in any way that “might be considered offensive” or “might otherwise adversely affect the reputation of Sport England”.
The spogo terms and conditions do not incorporate the OGL but fail as an open data licence on numerous points. One clause requires licensees to warrant that they will not use the data “in any way that causes detriment to spogo or brings spogo into disrepute”. Another requires licensees to warrant that the data will be reproduced “only in the form provided” and not “amended or dismantled in any way”.
(The spogo terms and conditions also, believe it or not, require the licensee to provide an uncapped indemnity. Frankly I wouldn’t touch this thing with a barge pole if it was a commercial licence, let alone an open data licence.)
The Charity Commission
The Charity Commission maintains the statutory register of charities in England and Wales. The Charities Act requires the Commission to provide copies of any documents it holds open for inspection to any person that requests them.
The Commission’s website has a page that sets out its policy on providing an electronic copy of the register for re-use.
I don’t quite know where to start with this policy. I can’t decide whether it’s more Lewis Carroll or Kafka.
The position of the Commission seems to be as follows:
In addition there’s some stuff about names of trustees etc. in the register being personal data, which is highly debatable.
I can’t begin to parse all of that so it makes sense; I think just reading it has made me stupider. But constructively it’s not any kind of open data licensing process.
The Cabinet Office has said that by March 2014 the Charity Commission will make some or all of the register data available as free open data, so perhaps the above nonsense will go away then.
Land Registry
Last but not least: Land Registry’s INSPIRE Index Polygons. This set of potentially very useful spatial data was released in September as a series of downloads on the Land Registry website.
The conditions of use start off well by saying the polygons and attributes “are available for use and re-use under the Open Government Licence”. But then the wheels come off:
However, as the licence says, OGL does not cover the use of third party rights which we are not authorised to license. Land Registry uses Ordnance Survey data in the preparation of the polygons and therefore you should contact Ordnance Survey for the relevant licence conditions if you require to:
(a) use the polygons (including the associated geometry, namely x,y co-ordinates) for a purpose other than personal, non-commercial use or commercial or non-commercial use within your organisation; or
(b) sub-license, distribute, sell or make available the polygons (including the associated geometry, namely x,y co-ordinates) to third parties.
Ordnance Survey has confirmed in correspondence that no, we can’t use the polygons as if they are open data. As a practical example, an attempt by Kent County Council to put the INSPIRE Index Polygons online was quickly shut down at the request of OS.
In theory the attributes in the dataset can be used as open data. But there’s not much to those attributes; just the INSPIRE ID and a couple of dates. And what use are attributes without the substance of the dataset, the polygons themselves?
The INSPIRE Index Polygons are not an open dataset and it was disingenuous of Land Registry to pretend otherwise.
Conclusions
It’s apparent from the above examples that data publishers will munge up the OGL for all sorts of reasons: perhaps their legal advisors have leaned on them, or they want to jump on the open data bandwagon without actually opening their data, or they’re just unclear on the concept.
However I think it’s important to call out data publishers (and public authorities in particular) on examples of bad practice like these that create an unnecessary proliferation of licences.
The Open Government Licence is highly regarded for its simplicity. The terms are relatively easy to understand for people who don’t usually read licences and contracts. The utility of the OGL is undermined if users have to interpret it in the context of additional, often contradictory, terms and conditions.
This confusion is not in the interests of data publishers either, since as a principle of contract law any ambiguity in meaning will usually be interpreted against the interests of the party who provided the wording.
Photo credit: Open_Data_stickers.jpg by Jonathan Gray (CC0 1.0)