[okfn-discuss] How important is the right to publish open data?

[Jeni Tennison]

Hi,

As some of you will be aware, ODI recently took the Open Data Certificates [1] into beta and started getting them being looked at by a wider audience. I'm looking for some advice about how to handle fuzziness in rights in published open data.

One of the questions in the certificate that we've had some pushback about is the question "Do you have the right to publish this data?" where the answers are 'yes', 'no' and 'I don't know'. If the publisher says 'I don't know' then they're asked if they generate the data themselves, or collate it from sources that are all open data, and if the answer to that is no they're asked to consult a lawyer to enable them to answer 'yes' or 'no'.

If publishers answer that they don't have the rights to publish the data then they don't get a certificate. The rationale for that is that you can't license data that you don't have the right to license, so the data can't be truly open. Also it's important for reusers to have some level of guarantee that the source of data that they're using isn't going to suddenly disappear as a result of a take-down notice. We want to make publishers stop and think about the rights in the data, and prevent them misleading reusers about that potential risk.

On the other hand, there are many datasets out there that are published as open data where the right of the publisher to publish that data isn't clear. The main examples happen when someone republishes data (in a different format, or combined with other data) that wasn't clearly licensed in the first place. These are useful datasets, which self-identify as open data and in many cases the odds of a take-down notice being issued (or complied with) are slim.

This argues for another answer to the question "Do you have the right to publish this data?" -- 'probably'. If someone answers like that then we could ask them to provide a description for reusers of the potential risks in reusing the data, though I have a feeling that they wouldn't want to do that for fear of it highlighting the legal grey areas that might expose them to challenge.

My concern is that 'open data' that is published without an unambiguous right to publish introduces uncertainty into the wider open data ecosystem: all datasets and apps that incorporate that data are then similarly tainted with that uncertainty, and it will prove hard to unpick if we find that we need to later.

What do you all think?

Cheers,

Jeni

[1] https://certificates.theodi.org/
--
Jeni Tennison, Technical Director   theODI.org
+44 (0) 7974 420 482                @JeniT