Post: 12 February 2016
The Country Register is a dataset curated by the Foreign & Commonwealth Office that contains British English-language names and descriptive terms for countries. It was launched yesterday and is re-usable as open data under the Open Government Licence.
This is the first government register to be released under a GDS-led project to produce “authoritative lists you can trust” or “canonical sources of truth” within a cohesive cross-government digital infrastructure.
I’m looking at this dataset from the perspective of re-use as open data, but am mindful that GDS’s approach to registers is very much based on utility, the internal data needs of government, building APIs, etc.
At the moment it’s not quite clear to me how GDS’s work on registers will fit in with previous initiatives on open data, the “national information infrastructure”, or older regulatory concepts of public registers – so I found the Country Register instructive as a practical example to go along with the theory.
For me there are two problems with this first release of the Country Register dataset: lack of documentation, and the contents of the bulk downloads.
Lack of documentation
The Country Register is a small dataset and builds on FCO’s existing geographical names index, with the addition of some start and end dates. But I know that only from reading a blog post by FCO’s Tony Worron. The landing page for the dataset provides little or no explanation of the content.
It’s important for re-users to understand that this dataset is “canonical” only as an expression of British foreign policy, and not a neutral presentation of global political geography. These are the countries that Britain recognises; other countries, as well as the UN, have their own authoritative lists. That’s the sort of key point that would be covered in good documentation, but seems to fall outside the scope of GDS’s minimalist, transaction-orientated approach.
It should also be made much clearer that this version of the Country Register is an alpha release; particularly as the landing page suggests it is beta. (It’s the registers service itself that is in beta, not this specific register.)
Bulk download
It is good to see a prominent link to a bulk download of the Country Register dataset, and that the download is available in formats (csv, tsv) that are accessible to non-coders.
However while the link says “download all entries” the bulk downloads actually contain “all records”. There are 201 and 199 records, and the distinction between a “record” and an “entry” is left unexplained.
At first I thought “records” were the current contents of the Country Register, and “entries” included those plus some expired records. But it’s not that simple. East Germany remains in the bulk download as a separate record, but West Germany does not because that code has been re-assigned to the current German state. This is potentially confusing and it would help to know the rules that define the current contents of the register.
It would also be useful if the register datasets were packaged for download with a metadata file containing definitions of each of the fields, and a clear statement of the re-use licence. (ONS’s approach to packaging of administrative data is a good template.)
Image credit: The World Showing British Empire in Red (1922) by Eric Fischer (CC BY 2.0)