By now you should have a good idea what your data looks like and how copyright works in your jurisdiction. To put this theory in to practice, the next step is to pick one or more datasets, this can be an entire collection or a subset of a collection. To do this, we have developed a couple of guidelines to help you with the selection. 

If you completed the previous task you should have a list of all the data of your institution. Now we're going to divide the sets in 4 categories:

1. Metadata
2. Content in the Public Domain
3. Content of which you own the copyright yourself
4. Content to which third parties own the copyright. 

Who owns Copyright?

Before we continue to divide the data into the aforementioned categories, it's a good idea to take a look at who actually owns copyright in some standard scenarios. It's important to know that your copyright has nothing to do with ownership. So if you don't usually own copyright to the collections that you manage. The main rule is that the creator/author has the copyright. However there are some exceptions possible. What if someone creates a work in service of someone else or for a company/institution? You have to make a distinction whether the work is created within the terms of a contract of employment or rather as a freelance project. If the work is created by an employee, copyright is automatically transferred to their employer, unless otherwise stipulated in the contract. If you are working with freelancers or other third party creators, it's the opposite. They keep their copyright unless these are explicitly transferred to your institution. 

If your institution is regularly working with freelancers or third party contractors for the management, maintenance or digitization of collections, it might be a good idea to make sure copyrights are always transferred so they can be governed centrally. 

In situations where the original author/creator has passed away and copyright doesn't end with his passing, it's automatically transferred to his heirs, in which case you'll have to get permission of all the heirs to make use of the work. 

It gets even more complicated, because the original author can also transfer his copyright or grant (exclusive) licenses for his work. In this case he usually keeps some properties of his copyright (mostly the so-called moral rights in case they exist in your jurisdiction), but the majority of rights then belong to someone else. Transfer of copyright is a wide spread practice in the world of book publishing and also for audio and video productions. 

Metadata

According the logic in the previous paragraph, copyright on metadata will usually belong to your institution, unless you make use of freelancers or third parties to create metadata for your collection. If this is the case, your should take a look at these individual contracts to determine wether there's a transfer of copyright or not. 

If your institution owns the copyright you'll have no legal problems when making the metadata available as open culture data. Therefore, metadata sets are usually a good starting point for institutions who want to open up their data. 

Public Domain

Works in the Public Domain are no longer protected by copyright. In this case everyone can make use of the work without permission. Actually you can say that Public Domain is the default setting for creative works and that copyright creates a temporary exception for exclusive exploitation of the work. If your collections, or parts of it, consist of works in the Public Domain, opening it should pose few problems. If your institution makes Public Domain works available, you should make it clear that copyright is not applicable to end users. A perfect way to do this is by using the Creative Commons Public Domain Mark. More about this tool in the next task!

When you live in the EU you can use Europeana's Public Domain Calculator to determine wether your works are still copyright protected or not. If you live outside of the EU, you should consult your copyright savvy peers or try to build upon your newly acquired knowledge about copyright protection terms in your own jurisdiction. 

Works of which you own copyright

Depending on your jurisdiction this scenario will be more or less likely. In most cases however, institutions won't own copyright to the collections they manage. Exceptions might be works that were commissioned by the institution, photographs taken by the institution (not necessarily if faithful reproductions of 2D works.), texts made by the institution, cases where artists, heirs or organisations donate a collection including copyrights to the institution, etc. Even if you own the rights to some of your (sub)collections, these works probably don't make up an extensive part of your collection. They still could be useful for reuse by others tough, so consider including these works in your open data policy. 

Since you own the copyright, it's up to you to decide wether you apply a free culture license or not. Off course, we suggest you do. Get your works out there and see them getting used in ways you could never imagined!

Works of which a third party owns copyright

The categories above pose little to no problems when it comes to opening up your data. Unfortunately, most of the time collections will turn out to be in this last category. Here's where the theory gets simple but in practice it might prove to be very difficult to make these works available for reuse. The general rule is that you need permission from the rights-holder to make a work available to the public. As we explained in the first paragraph of this task, the rights holder can be any of a number of people ranging from the original creator, to the heirs, or a professional publisher. If you want to open your collection consisting of works where there's still copyright involved but you don't own it, it's up to you to get explicit permission for every work in the collection.

In some cases you might be able to strike a deal with a collecting society to grant a general license to use the works in exchange for a sum of money, but this technique isn't really tried and tested. 

The figure above gives a visual representation of what you should accomplish with this task. The goal is to divide your data in each of the four categories and then make an assessment of which sets would be easiest to open up. 

Make a list of your datasets and give them a green, orange or red light according to how many problems you see in attaching an open license to the set.