7

I've seen several attempts to define open data and open datasets.

One much cited is Tim Berners Lee's 5 Star crieria for linked open data. The Wikipedia page on Open Data is also often referenced.

While most of these definitions list several useful and important requirements of often datasets, I do not think they are not yet to be considered authorative. I also think they are rather incomplete.

For open data to be useable, it needs to come with provenance, and it should be accompanied with some sort of policy document that says something regarding its maintenance (or lack thereof) and procedures for resolving disputes and correcting errors. While just dumping the data in the public domain seems to be all some "open data" people want, if you actually care about the data being usable for serious work, having it in the public domain without ownership/accountability/maintainability defined, is probably the last thing you want. You want freedom (to reuse and to share), yes. But you also want transparency regarding who is in charge of the data and who provides QA (even if the answer is "no-one").

The closest thing I've come to addressing these concerns is the ODDC Assessing open data supply report. This is a working document, and it doesn't adress all my concerns, but it is a good start.

However, the ODDC report is a scoreboard, not a definition. And it doesn't address provenance, granularity, the role of providers as interpreters (by means of selection or aggregation), etc. just to name a few of what I regard as omissions.

My question is basically: Does there exist an authoritative definition of an open dataset that takes these type of considerations into account?

If the answer to this is "no", my follow-up question: Does there exist some sort of organization, community or international NGO that attempts to create this type of definition? I.e. something that people such as myself can get involved in to move such an (hopefully to-become authoritative) definition forwards?

(In the world of Free and Open Source Software, there are - for instance the the the Open Source Initiative and the Free Software Foundation, who attempts to supply such an definition for the term "Free and Open Source Software" - or rather "Open Source Software" and "Free Software" - since they can't even agree on what to call it. I hesitate to cite these organizations as examples, because the long going rivalry and partisanship between them is of course not something I would like to see replicated in the open dataset community. But I think their heart and their aspirations are in the right place, while the way they carry out their mission has potential for improvement).

  • Not an answer to your question, but be wary of conflating linked data and open data; they're orthogonal to one another, and data can be neither, one, the other, or both. See opendata.stackexchange.com/questions/522/… for more discussion. – Andrew Pendleton May 23 '13 at 17:17
  • I am aware of the discussion, but IMNSHO, data must provide linked semantics to be considered "open" (they're pretty useless without it). And while we have an eight times upvoted answer here that say they're orthogonal to one another, I am not so sure about that ... (I am OK with the part that says that not all linked data are open, but not vice versa.) – user135 May 23 '13 at 17:25
  • That opinion is reasonable, but far from universal, and not shared by many people/organizations that consider themselves players in the open data space. If the goal of this question is to arrive at a definition of open data that can achieve community consensus, I don't think that can be a requirement. I would also dispute your characterization of non-linked open data as useless -- I make use of it every day. It all depends on what your intended use is. – Andrew Pendleton May 23 '13 at 17:46
  • Just to make sure: The goal of this question is not to arrive at definition (that's not realistic, and SE discourages discussion). The goal of this question is to identify the players in the definition game, and the definitions that are in play. I've got some good pointers already, but I hope for more. – user135 May 23 '13 at 18:50
13

Yes, there is: the Open Definition defines openness for data (and content). The Definition was produced in 2005, heavily based on the Open Source Definition, and revised minimally since.

The key part of the Open Definition states:

A dataset [work] is open if its manner of distribution satisfies the following conditions, which simultaneously delimit the characteristics of a suitable open license:

1. Access

The work shall be available as a whole and at no more than a reasonable reproduction cost, preferably downloading via the Internet without charge. The work must also be available in a convenient and modifiable form. The license may require the work to be available in a convenient and modifiable form.

Comment: This can be summarized as 'social' openness - not only are you allowed to get the work but you can get it. 'As a whole' prevents the limitation of access by indirect means, for example by only allowing access to a few items of a database at a time. An example of 'reasonable reproduction cost' is the cost of a blank DVD required to distribute a complete database.

2. Redistribution

The license shall not restrict any party from selling or giving away the work either on its own or as part of a package made from works from many different sources. The license shall not require a royalty or other fee for such sale or distribution.

3. Reuse

The license must allow for modifications and derivative works and must allow them to be distributed under the terms of the original work.

Comment: Note that this clause does not prevent the use of 'viral' or share-alike licenses that require redistribution of modifications under the same terms as the original.

...

7. No Discrimination Against Persons or Groups

The license must not discriminate against any person or group of persons.

Comment: In order to get the maximum benefit from the process, the maximum diversity of persons and groups should be equally eligible to contribute to open knowledge. Therefore we forbid any open-knowledge license from locking anybody out of the process.

Comment: this is taken directly from item 5 of the OSD.

8. No Discrimination Against Fields of Endeavor

The license must not restrict anyone from making use of the work in a specific field of endeavor. For example, it may not restrict the work from being used in a business, or from being used for genetic research.

Comment: The major intention of this clause is to prohibit license traps that prevent open material from being used commercially. We want commercial users to join our community, not feel excluded from it.

Comment: this is taken directly from item 6 of the OSD.

(Disclosure: I helped draft the first version of the Open Definition and have helped curate it since along with other members of the Open Definition Advisory Council)

  • 2
    The recently released Project Open Data syncs with this entirely - project-open-data.github.io/principles. – Gray B. May 24 '13 at 3:59
  • The one line summary is helpful to include here: “Open means anyone can freely access, use, modify, and share for any purpose (subject, at most, to requirements that preserve provenance and openness).” – Steve Bennett Oct 8 '16 at 7:54
4

I think your question has two distinct aspects:

  1. What might be considered to be an "Open" dataset?
  2. What are the best practices for publishing data to the web?

To some extent these are orthogonal, e.g. there are best practices for publishing data that don't specifically relate to its licensing.

I think there are already clear answers to the first part. There are some excellent attempts to define what open-ness means in the context of data. The material on opendefinition.org is very good and has some good guidance on open licensing.

Answering the second involves more factors and personally I don't think this has much to do with "open-ness". Publishing data to the web for others to use means that you have to think about:

  • Sustainability: will the data be available over the long term?
  • Quality
  • Provenance
  • Means of access
  • Timeliness

The importance of these questions varies depending on what you want to do with the data. E.g. build a quick hack vs building a business.

The Open Data Institute is working on an "Open Data Certificate" that addresses these questions and much more. You can see this work in progress at:

http://theodi.github.io/open-data-certificate/

The certificate assesses data on multiple levels, giving a scoring and steps for further improvement. IMHO this is a good, nuanced way to address these wider issues.

2

There is also the Open Database License: http://opendatacommons.org/licenses/odbl/summary/

The ODbL states that:

You are free to share, create and adapt the database, as long as you attribute, share-alike and keep thee database open.

The ODbL is used by the Open Street Map project. In the OSM Wiki they have some resources about the ODbL and why they use it: http://wiki.openstreetmap.org/wiki/Open_Database_License

1

The following resources are mainly about parliament/government open data, but I guess much of it applies to general open data.

1

I don't think there is an authorative definition. And there may never be one. Issues such as access, degree of (technical) openness, raw versus cooked data and whether data must be available at source for it to be considered open are some of the contested issues I have already encountered. I resort to the summary Open Definition as the most useful:

"Open data is data that can be freely used, reused and redistributed by anyone – subject only, at most, to the requirement to attribute and sharealike." http://opendefinition.org/

Once the detail is added, the contestation begins.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy