IATI data licensing (for unencumbered data use!)

(Andy Lulham) #1

@stevieflow’s recent datastore discuss post raised some issues about IATI data licenses that we should aim to resolve, in order to facilitate unencumbered and unfettered data use.

For context, the licensing section of the publishing guidance on the (old) IATI website says:

The TAG secretariat, in collaboration with a small group chaired by the World Bank and including an open data/intellectual property lawyer, produced a set of recommendations for licensing. The IATI Licensing Standard, agreed in February 2011 at a meeting of signatories and the IATI Steering Committee, is that information published through the IATI standard should be licensed under an open license. It is a set of principles that must be adhered to, rather than a prescriptive set of terms and conditions.

(Emphasis mine. NB it looks like the above paragraph was lost in the move to the new iatistandard.org site, and so it’s missing from the corresponding page. Thanks to @markbrough for this institutional knowledge!)

So, here are four proposed action points aimed at providing a consistent approach to IATI data licensing:

  1. Make data licenses mandatory for all new datasets in the registry (i.e. make it a required field, and remove the “Not specified” option). @amys has already created a ticket for this.
  2. Improve the licensing guidance on the IATI site so that it is (i) self-consistent and clear that IATI data is open data (see: the contradictory wording that @siemvaessen flags on discuss) and (ii) provides a bit more explanation of why e.g. non-commercial (or closed more generally) is problematic.
  3. Fix existing “closed” and “no license specified” data in the registry, by contacting publishers and asking them to select an open license. I guess if these aren’t updated after some pre-defined notice period (h/t @stevieflow), these datasets would either need to default to an open license or be removed in order to be consistent with the IATI Licensing Standard.
  4. Remove the closed license options from the “choose a license” dropdown on the registry.

How does that sound? Tagging people who commented on licensing in the datastore thread: @stevieflow @Herman @yohannaloucheur @bill_anderson @siemvaessen @David_Megginson

(Herman van Loon) #2

Hi Andy,
Sounds great! Imo these 4 points nail the license issue. Completely agree that we shouldn’t accept closed licences as a part of the standard. IATI is an open standard.

(Yohanna Loucheur) #3

Thanks Andy for this very useful summary and list of actions! Agree as well.

(Steven Flower) #4

Excellent, thanks @andylolz. I agree - this is a very practical list to implement

I see some difficulty here, as we are actually saying that non-open data would no longer be accessible via the IATI Registry? This should therefore involve a period of notice (as with the deprecation of version 1) to enable these publishers to schedule their action (if any)?

(Andy Lulham) #5

+1. Have amended the original post.

IATI Datastore - what data should go in?
(SJohns) #6

Just thinking this through for CSOs. Anything that makes it super-simple to do the right thing, for example making the license field mandatory and only making it possible to only choose an open license is very welcome. Looking at the dashboard, losing the ‘licence not specified’ would affect 1395 activities from 443 organsations, and losing the non-open options would affect 54 activities from 23 organisations.

I wonder if we should go further and reduce the number of options on the dropdown list (currently 10 excluding ‘not specified’) to just the 6 open data licenses that meet the open definition requirements plus one category that allows governments to show their data licensed under an OGL or similar.

Reason is that 2929 activities are currently licensed under the license “OKD Compliant::Other (Attribution)” and there are several more ‘Other’ options.

Given that licenses are there for data users to check how the data is allowed to be used, having undefined/vague license information is not very useful.

And for civil society data providers it’s already hard to understand the difference between CC and ODC - throw in an option for ‘Other’ and they will go for that by default (to see what I mean, see the list of organisations using the “OKD Compliant::Other (Attribution)” option.

What do you think?

Also, what about having an element or attribute in the Standard which defines which open data license is attached to the data? For example as an attribute of iati-activities? Then the license is right there in the data, the same way you can embed a CC license in other open works.

I’m not a technical person so it may have been considered before and thrown out! @markbrough might know.

(Andy Lulham) #7

This is great – thanks @SJohns! I had no idea about the license page on the IATI dashboard – thanks for sharing!

Agree that undefined/vague license information is bad! If an “other” license is chosen, there would ideally be a way to specify it by linking to or embedding some license text. Fortunately, all 9 of the open licenses available in the Registry dropdown declare open definition compliance (see the “ OKD Compliant::” prefix in this list) so that’s definitely something.

I’m certainly in favour of pushing publishers towards choosing well-known open license options, such as the Creative Commons suite.

However, I think it could be a problem to remove the “other” open license options, because a publisher isn’t necessarily choosing the data license for the first time when they select it in the registry. They might be publishing some already-licensed data, so need a way to select the existing license.