The talking paintings were an educational interactive museum installation designed in 20 weeks for the Dutch Open-Air Museum in Arnhem, the Netherlands. Originally, the museum was looking for a way to increase the engagement between young adults (20 – 25 years old) and the old houses that are on display in the museum. After a closer examination, the team came to the conclusion that the real problem was, in fact, the overuse of information signs. Those distract from the real strong points of the institution: education by immersion.

​

Therefore, the team of four decided to replace these signs with interactive pictures of hypothetical residents. Each of these paintings should be able to hold a conversation with the visitors about their daily lives and other information provided by the museum. The visitor could choose one of three possible topics and then “discuss” it with the person in the frame. It would be possible to switch between topics if the resident dropped a certain piece of information, to give the illusion of a real conversation. Two video prototypes were completed, a third text-based prototype was finish before the project was over.

​

For this project, I was the main narrative designer. The writing, the story research and the character creation was all done by me.

At first, all the information provided by the museum was ordered into a timeline, to understand what events happened during the time frame a resident should give information about. This was also done to prevent an character talking about an event that they couldn't possibly know about yet. Then four characters were created, each living in one house. The character sheets were created in the following order: a nurse from the 1950s, a factory worker from the 1910s, a weaver from the 1870s, and a postman from the 1970s. You can view all character sheets here.

​

Now, the real work on the narrative happened. Again, I started with the story of the nurse. My approach was to first write down a conversation between the character and a placeholder interviewer, that would be replaced by the visitor of the museum later. This linear dialogue was split up into nine topics. Then, these topics needed to be connected to each other in a meaningful way. Each paragraph inside of these topics was numbered and then got three questions attached, that would lead to different paragraphs and topics. To test this, the text was implemented into a Twine prototype. A more refined version of this approach was used to write the stories of the other two characters.

​

If you want to read about this process in more detail, check out the article I wrote for our study-internal wiki here.

​

You can play the text version of the 1950s nurse story here, the story of the 1910s factory worker here, and the story of the 1970s postman here.

​

Down here you can see the flow charts used to determine which questions lead to which passages.