How to license a dataset


#1

Dear all,

I’m involved into an H2020 project that has collected and generated quite a lot of data.
We published the dataset into a set of dedicated git repository.
We are using the datapackage format to share the (meta-)data.
We add also a REDME file that try to be informative with the main description od the data sources used and the main elaborations.

You may found an example here:

We would like to specify the licence, we know how to do it following the datapackage specifications:

What it is not clear to me is how to properly specify the license in the README, do you think that add a dedicated section like in the below section, is enough?

# License
[Creative Commons Attribution 4.0 International](https://spdx.org/licenses/CC-BY-4.0.html)
Copyright 2016 - 2020, Hotmaps contributors
SPDX-License-Identifier: CC-BY-4.0

Do you have better ways?


#2

Hi @zarch I normally do the following. It is a bit redundant as you can see but should maximize the chances of being picked up by common license filters.

OPEN LICENSE

Copyright (c) 2000-2018 Bill Frog <bill.frog@open.isp>

SPDX-License-Identifier: CC-BY-4.0
License-Text: https://spdx.org/licenses/CC-BY-4.0.html

Creative Commons Attribution 4.0 International License
This work is licensed under a Creative Commons CC BY 4.0 International License.

Some technical points to note:

  • I would avoid markdown markup
  • the “(c)” is important for license tooling
  • providing emails addresses is optional

The concept of Hotmap contributors is rather vague. Normally one lists individual authors by name or pseudonym.

The OKI datapackage JSON file also supports legal fields. There are currently two proposals in the openmod community for a minimum set of mandatory fields. I don’t have the details at hand but someone more knowledgeable may well post.

My experience is that most of the advice out there from organizations like the FSFE and OKI on creating license notices is inadequate. HTH.


#3

You can find the OEP metadata string (v1.4 still under development) here:

It is a JSON string following the frictionless datapackage format. It is slightly different from other selections.