Brunnsmuseet manages over 10,000 scanned photographs.1 Most of these images originate from the museum's only completed mass digitization to date. In addition, the collections have been supplemented by small-scale digitization of archival material acquired and stored at other archival institutions, including Sätra Brunn's archive in Uppsala and the The diocesan library (Stiftsbiblioteket) in Västerås.
In this text, we take a deep dive into Brunnsmuseet's broad approach to digitizing cultural heritage, about what has worked and what are the main challenges for the future.
The mass digitization
Brunnsmuseet was formed as a working group in the fall of 2013. Already that fall, planning began for the work of implementing a mass digitization of physical image material in the form of, among other things, developed photographs, postcards and newspaper clippings. The original consisted of unstructured or, at best, arbitrarily structured image material that the Brunnsmuseet had donated or borrowed from various private individuals. The structure could be in the form of folders, boxes or other distinctive containers.
The largest part of the mass digitization was carried out in the years 2014-2016 and focused on the scanning of material. This period is so far the only one in Brunnsmuseet's history where work has been carried out by paid labour. The financier for the work was the museum's initiator Tohbbe Lindblom through his company Kreativitetsverket. The staff who worked on the scanning were part-financed through wage subsidies and other labor market policy measures, including the “Cultural Heritage Lift” (Kulturarvslyftet), an discontinued cultural and labor market policy investment.
Since the originals came from private individuals, there was no catalog of structured data on the various characteristics of the included images, such as year, photographer, people depicted, etc. This information was also usually not available otherwise in the private collections, although the scanned postcards provided some clues to at least the time period to which the depictions probably belong.
I would love to tell you about the miracle method of how we went from 0 to 10,000 images without any original catalog and still found utility and use in the scanned material. About how we were able to reconstruct and find all the data that was missing and supplement each scanned object with extensive metadata. The problem is mostly that there is no such miracle method. The fact is that much of the material that was scanned is still enigmatic, that we cannot answer the basic questions of when the picture was taken, who took the picture and what or who was depicted.
Metadata
There are different types of metadata. These could be divided into three groups. These can be referred to as technical metadata, intuitive metadata, and inherited metadata.
Technical metadata describes the object's physical characteristics, such as width and height, and information that can be linked to the digitization itself, such as when the scan was made, with which equipment/software and which file format the digital object has. In relatively modern, but analog cameras, it is often possible to get the date and time watermarked on the developed image. In modern digital cameras, we automatically receive information about not only when the picture was taken, but also a geographical location about where the picture was taken.
Intuitive metadata is the result of looking at the work and interpreting it using our senses. With sight, we can add descriptions of the type "a woman is sitting on a horse". Such a general description may seem banal, but it can be useful. For example, such descriptions can increase accessibility if they are used as alternative text for those who use a screen reader or read-out screen. Such descriptions can also increase searchability and provide greater understanding of the image's subject.
The original image on Omeka (ID: 85) (link in swedish) is mounted on cabinet card, which in itself, however, does not provide any further clues as to the detailed origin of the image.
At the same time, it is not normally possible to read out details from, for example, pictures just by looking at them. When dating, we can make an estimate, a qualified guess, which can rarely date the image closer than a decade. To know more, we need something else. The ability to interpret images has largely not changed over the years, although the knowledge available to place the image in historical context has unequivocally increased.
Inherited metadata, in conclusion, is the type of metadata that was established in connection with the creation of the image or in close connection with that event, and which is not evident from the image itself. It may be the only way to know who is the photographer, who is the woman on the horse, what day and date it is, etc. If you are really lucky, they can be supplemented afterwards, but then with an increased risk that the information that provided are incorrect because, for example, the informant has forgotten important details or may confuse people.
The technical metadata that we receive automatically from the camera has thus become better. The intuitive metadata has largely not changed, but today we have a greater knowledge base for interpreting and trying to date an image. At the same time, it is easy to lose other important metadata.
It is something that has become increasingly clear as we have worked with our collections. It is difficult to start creating historical collections. To begin recreating history long after the events portrayed. You can imagine for yourself that it is easy to name all colleagues if you took a picture at a staff party. But look at the same picture two years from now and the challenge is all the greater. Here we are rather talking about pictures fifty years ago, where we were not involved.
The knowledge paradox
Museums are knowledge institutions. Museums acquire, manage and develop knowledge about their collections. Even though Brunnsmuseet is a non-profit working life museum, we adhere to this knowledge tradition. We are and want to be a museum among others. This is also how we are perceived by others. We are expected to know a lot about our own collections. Yes, we are expected to know something more than what we tell. Knowledge that, due to a lack of resources, could not be made available to a wider public. But the knowledge must still be there. It is the very cornerstone of a museum.
We largely carry the legitimacy of other, more professional museums by also calling ourselves a "museum". But what happens if we as a museum don't know?
In the movie "Notting Hill" there is a famous scene where the busy bookseller William meets the famous movie star Anna Scott. William says that they live such different lives that a relationship between them would not be possible.
Anna then replies: "I'm also just a girl, standing in front of a boy, asking him to love her." I would like to answer much the same to our museum visitors every time someone asks for more details about an image than the (if any) caption tells. With the same warm charm as Anna Scott, I want to say something like: "I'm also just a small museum, standing in front of an audience, asking them to help me."
In a way, the roles are reversed. We hope that our visitors will tell us what it is that they see. It also works to some extent. Every year we receive tips, mainly from visitors to our physical exhibitions who can tell us things like "that was my grandfather!", "her name is Ulla", "that tool is called a flea hoe", etc. But even if the mentioned method works with some success in our exhibitions, we immediately see problems when we think large scale. Let's say we exhibit a hundred pictures in our exhibitions, that doesn't help us more in supplementing the ten thousand pictures that we don't have on display. So how should we work with these?
Of course, we can post pictures on our social media. It is something we do when there is time and need, and we also get quick answers from our dedicated and knowledgeable followers who have often worked at Sätra Brunn before. But it all has its obvious limitations. We need to actively ask questions to actively get answers. With a lot of pictures, there will be a lot of questions that need to be answered.
We are investigating the working methods...
Over the years, we have discussed a number of different proposals and made various attempts to address the knowledge gap around the type of metadata that describes the human characteristics of the image and subject (as opposed to purely technical metadata). In connection with Cultural Heritage Day 2022, we arranged a three-day event where we dedicated the entire Saturday to the hope of being able to refine our digital collections.
Unfortunately, the event was not as well attended as we had hoped, and the challenge remains. In the evaluation of the event, we realized the need to give concrete examples, to present a small selection of material to work with and to tone down the technical aspect of the general marketing.
We continue to investigate this with forms of work based on the experiences we ourselves have acquired based on our context and what we can be inspired by from other memorial institutions.
So where do we stand today?
Brunnsmuseet today has the ability to independently scan paper objects up to A3 format and distribute these on the internet. The material finds some use in our exhibitions, is published on social media and a smaller part has also been published on our Omeka, our own heritage collections website, and also shared on Wikimedia Commons. We have the competence, equipment and infrastructure to go through every step from digitization to digital availability. Technically, we would have the possibility to mass upload images, but the lack of metadata means that we have only made available just under 400 images so far.
Brunnsmuseet's work and collections are supplemented by scattered material managed by older cultural institutions, which have systematically preserved large parts of Sätra Brunn's cultural heritage. Among these, Uppsala University Library and University Archive are the most prominent. Uppsala University managed Sätra Brunn for 250 years, in the years 1747–1998.
In conclusion
In conclusion, it can be said that there is still a lot of work to be done in processing the material from our previous mass digitization. This is mainly due to the fact that we lack the essential knowledge about the objects that we have, what we refer to here as inherited metadata. The knowledge paradox means increased challenges to supplement our collections with the necessary metadata, to the extent that it is possible to supplement.
However, the goal of our business has never been to digitize as much as possible. On the contrary, it has been a greater satisfaction, and it is also natural that the greatest efforts on behalf of Brunnsmuseet have been to process, improve and disseminate such material that other cultural heritage institutions have digitized and made available, either through their own priorities or at our request. This work forms the very core of what we refer to as Open Sources Sätra Brunn.
Although non-profit museums can do a lot, there are great challenges, which to some extent may be about limited resources. But the resource that is missing does not necessarily have to be about the most obvious ones such as finance or personnel, but it can also be a lack of knowledge about one's collections when trying to reconstruct history afterwards.
The estimate of images has not been updated for a long time, but the number should be somewhere between 10-15,000 image objects.