Licensing data aggregates

(Adrien Labaeye) #1

How do we license data that emerges from the 14MMM process?
I would distinguish at least two type of data

  • mapping of the mappings (urgent: the data is online withouth license)
  • transformap (the map of alternatives)

I guess the challenges are to:

  • keep data open as possible and enable sharing/dissemination/analysis, etc.
  • avoid that some greedy actors appropriate the data as their own and then keep it closed

Pelias, an Open Source geocoder based on Elasticsearch
Field research summary for TransforMap: some insights about map commons
(Adrien Labaeye) #2

In response to proposal to use Michael Maier wrote:

Clearly not - we would like remain as open as possible, and the more
restrictions we put in the less open it is. In some definitions, NC (and
even SA) (and for some people even BY) is not Open.

Choosing a NC-license wouldn’t even allow a blogger to embed one of
our maps in an article, if he has a single ad on his website to cover
his hosting fees.

Apart from the “religious” licensing wars, when working with
OpenStreetMap, we have only two options:

  • Public Domain (CC-0)
  • Open Database License (which is like CC-BY-SA for data)

ALL others are not compatible with OSM.

(Adrien Labaeye) #3

Michael Vesely wrote:

  • Licensing is a big and open question we definitely have to discuss in Leipzig.
  • We have the licensing constraints given by OSM
  • Dogmatic discussions what is free/not free/…enough
  • And there are considerations to protect the privacy (e.g. phone numbers, email) of the initiatives we map against advertising
    companies or spammers who already offered to pay money for well
    structured, categorized and maintained addresses.

(Michael Maier) #4

ad 2nd)

avoid that some greedy actors appropriate the data as their own and then keep it closed

You cannot prevent usage of ANY data by bad-minded people. If some greedy actor wants the data, he will simply ignore any licensing conditions. Even if we require login for sensible data, if someone really wants data, they can fake a request to join, grab all data and run.
Remember: the more barriers you erect to ‘protect’ something, the more possible participants you will scare off.

(Michael Maier) #5

I’ve started a wiki page on the topic “Licensing” - maybe we could gather helpful information here:

(Michael Vesely) #6

I know the religious approach to have everything always public. And as posted before, we try to be as open as possible without neglecting the risks. And we post all suitable info to OSM anyway. So no need for any licensing discussion there.
But still we are not in utopia (yet ;)) and we got requests from professional spammers to buy the well maintained and categorized addresses of initiatives. And others raised reasonable concernes not willing to be mapped because of these and other issues. E.g. Adrien was asked that kind of things in his session at Degrowth.
This is why we are investigating the possible risks to continue that discussion based on a little more knowledge about the actual risks and concerns people may have in order not to scare off people by ignoring their concerns.

(Michael Vesely) #7

@wiki so we have splintered that very discussion to a third media (trello, discourse=foyer and now a wiki, am I missing mailing list and, …)? That is the way to scare off people like me!
A wiki may be (not shure about that) the way to publish the result of the discussion but should not be used to preempt it.

(Michael Maier) #8

Yes, the Wiki should be the place where the results should be in the long term.
But also the reasons that lead to this or that license, and an overview to the license thematics.

And - someday we anyway had to start this Wiki-page, I just did that now.

And for Trello - Trello was never meant to be a discussion Platform, but some people are haveing a hard time using the right tool (discourse) for the right job :wink:

(Richard) #9

I’ agree, NC is not an option. no one like it and no one need it.
I prefer CC BY-SA

Our vision, our politics
(Jon Richter) #10

I’m just throwing in the possibility of using different contemporary licencing schemes. More info at:

Bauwens sais a reciprocity licence still needs at least a year, so the Peer Production Licence might be worth investigating.

(Jon Richter) #11

(Jon Richter) #12

So, here’s the deal: by far the MAIN impact of a CBRL on such things as text, photos, and related media will be that it becomes a no-Wikipedia license. A CBRL will have negligible effect on capitalists, will not likely provide substantial income to the commons, and almost everyone it will affect will be all of us who wanted to build something FOR the commons using both Wikimedia resources and these other resources, and the license-incompatibility will make it illegal. In other words, I think the most likely affect of CBRL is to directly hurt the commons and have no other impact.

To be clear, the NC clause in Creative Commons does exactly this harm. Thus, CBRL will not be WORSE that the NC licenses (EDIT: it will be worse in practice because of even further incompatibility, being that CBRL will be incompatible with both copyleft licenses used by GPL software and by Wikimedia and others and also incompatible with existing NC material), but CBRL only attempts to fix the problem of lacking clarity about what-is-commercial and fails to address the arguably primary problem with NC: incompatibility.

via Aaron Wolf on Loomio

(Jon Richter) #13

Hi thinkers and tinkers,

before we delegate the licensing question to lawyers, you can now have a read of:

(Jon Richter) #15

The title of the linked article indicates it as Commons-Based Reciprocity Licences.

Edit: @alabaeye Should we also turn this original post into a wiki post and compile all references from below at the top? Funnily engouh, people are actually reading this stuff.

I also have to add the

(Michael Maier) #16

I add a new consideration:

(Gualter Barbas Baptista) #17

Would it not be possible and more reasonable to offer the end users/APIs the possibility to choose among the different copyleft/copyfarleft/PD licenses? Then we could make clear (or design the interface in a way) that if the POIs are to be available on OSM, then only a smaller selection of compatible licenses is available.

I personally have problems with PD, because making something available for free (very similarly to as in beer, although for non-exclusive uses) means offering the hegemony something for free. Commercial interests (or parasite-like people/institutions) are probably using data everyone contributes more quickly, efficiently and profit-like than the groups that TransforMap aims at - and hindering/reverting the effect TransforMap envisions to have on society.

Non-public, commons data
(Michael Maier) #18

Everything is possible, as we are not even in the design phase :wink:

It would also be possible to give different licenses to different properties of the data! An example:

  • location, address, name and description are PD (for use in OSM)
  • contact person data are CC-BY-SA-NC and are stored elsewhere
  • Photo is CC-BY

A website will gather this data from different websites, and can display the appropriate License for each property besides it in a popup.

In this scenario we wouldn’t clutter up the map with thousands of contributor names, as the data displayed in the overview is ODbL (OSM), and the details can be displayed with their respective license.

I love that idea :smile:

(Jon Richter) #19

I’m highly critical of explicit non-commercial licences. Also, with conformance to data privacy legislation, we aren’t allowed to publish personal data without former personal consent anyway.

(Michael Maier) #20

That was only an example …
You know I’m not a fan of NC either :wink:

(Josef Kreitmayer) #21

Currently we are developing a data-contributor agreement for the SSEDAS project (details about the SSEDAS project will be published soon here in discourse). The following are discussions how to get a good license agreement for them going. As we want to make as much data as possible osm-compatible, the license agreement will be a public domain one.