New Data.gov.uk, same as the old Data.gov.uk

Update: 20 April 2018

GDS has now added ‘Additional licence information’ links to DGU records affected by the main technical issue highlighted in this post.

Access to ancillary material on the old DGU platform, including the Contracts Finder archive and the Glossary, has been restored.

Post: 13 April 2018

Yesterday the Government Digital Service relaunched the user-facing part of Data.gov.uk as ‘Find open data’.

The new site provides a minimum viable metadata catalogue for government data, on a more stable platform and wrapped in the now familiar GDS design ethos.

Most of the extra features that remained on the old platform are gone: no more user comments, data requests, openness ratings, ODI certificates, site analytics, document library, etc.

The Glossary of Public Sector Information and Open Data Terminology is gone.

The Contracts Finder archive is also apparently gone, though it’s still listed in the catalogue.

I understand nightly ’dumps’ of the DGU catalogue will remain available, even if they are no easier to discover than any other dataset.

I don’t want to belabour this too much – as a regular user of DGU I had low expectations for the new platform and GDS has duly met them. It’s the same uncurated bucket of metadata we had before, with less detail and functionality.

But I’d like to highlight a more specific problem GDS has created for some data publishers.

Individual DGU records now display less metadata. Full GEMINI metadata files have been removed.

In particular GDS has changed how it displays the terms of re-use for datasets.

Most records now have a simple link to the latest version of the Open Government Licence.

Usually that’s fine. But in many dataset records GDS has effectively changed the terms of re-use by removing or editing the use constraints submitted by the data publisher.

Let’s look at some examples.

• NIEA’s Waste Site Licences & Exemptions dataset was previously licensed under v2 of the OGL. GDS has updated that to v3.
  

• Leeds City Council’s Adopted Highways dataset was previously licensed as Creative Commons Attribution, with a link to the licence. GDS has changed that to “cc-by” with no link or explanation for users.
  

• BaNES’s Contaminated Land Register dataset was previously licensed under the INSPIRE Public Sector End User Licence. GDS has changed that to “Licence: None”.
  

• Environment Agency’s Consented Discharges to Controlled Waters with Conditions dataset was previously licensed under EA’s Conditional Licence. GDS has changed that also to “Licence: None”.
  

• Scottish Government’s Data Zone Boundaries 2011 dataset was previously published with an attribution statement that acknowledged the intellectual property rights of Ordnance Survey. GDS has removed that statement, placing Scottish Government in technical non-compliance with its OS licence.

What should data publishers do if they’re not happy with DGU’s presentation of their datasets?

• Many government bodies now run their own open data sites. This is an increasingly affordable option for publishers who want to make sure users have access to the canonical version of their metadata.
  

• Whenever possible, licensing information and other key documentation should be packaged in the same download as the data itself. 
  

• If the dataset is in a format that doesn’t support additional information, such as an unzipped CSV file, make sure the licensing and attribution is included in the Summary or attached as a separate resource to the DGU record.

There is more to come for data publishers. GDS has also been working on changes to the publishing tool that harvesters records into DGU, with launch of a “new improved service” planned for June.

Note: The reference to DGU dumps in the above post was amended on 13 April after confirmation that these are still being published following the move to the new platform.