From methods to practice: On the sustainability of archival software in research settings

Adrian Demleitner
Tanzmatten 4, 2502 Biel/Bienne CH
orcid 0000-0001-9918-7300


This presentation is concerned with the sustainability of technological creations that accompany and support our research in the digital humanities.

These can be software, data structures, trained models for machine learning, and others. Since these artefacts are an essential part of the knowledge we produce, we need to ensure their long-term survival alongside our other outputs, such as academic writing. Technological artefacts are often dependent on complex systems and environments that are neither robust nor can function without maintenance.

How do these premises shape our methods in the creation of said artefacts? How can we balance prototyping and sustain outputs in the long term? Alternatively, what could be a sustainable approach to technological artefacts in the digital humanities?

This presentation draws on my longstanding practice as a software developer. I want to touch on these topics by showing how these thoughts and problems influence the design and creation of digital archival software within the project «Participatory Knowledge Practices in Analogue and Digital Image Archives».


prototype, software, archives, methods, participation, sustainability

0. Introduction

Before I start I want to give you a little bit of context on myself. I have a background as a software developer with a good 20 years of experience and within this function I am part of the research project Participatory Image Archives, or PIA for short.

Although being a “techie” I studied design and just recently finished my masters in design research at the University of Arts in Bern. That has an influence on my research practice as I’m often reliant on design theory and research methods. Design research as I practice it is close to anthropology and, next to using ethnographic methods, often has a focus on sociotechnical and material issues.

There is a certain archetype of artefacts in science-fiction that I’m fascinated with. They are often portrayed as geometric sculptures in rugged landscapes, hinting to space-faring folks and civilisation long ago. They are usually some kind of archive, holding knowledge or giving warning about some dangers. And, here comes the science-fiction part: They always seem to work properly after millennia without any maintenance or bug-fixing.

In the following minutes I want to talk about this temporal sustainability, especially in relation to prototyping, one of the more ubiquitous methods in practice. I want to take up the challenges surrounding prototyping in research and how they relate to the project that I’m a part of. To end this presentation I will propose participatory processes as a way of approaching sustainable prototyping. There will also be a trajectory from exciting theory to mundane practice in this talk.

1. Prototypes

Let us begin with framing prototypes. I’ll speak form a position where artefact usually means prototype and prototype usually means software.

By their nature, prototypes are highly dynamic and fragile objects (Asante-Agyei, Manfredi, and Erickson 2022). This is expressed in two aspects. They are never quite fixed in time-space or finished as they are iteratively worked on. They are also dished out rather hacky without much thought on technological sustainability. That leads to all kinds of data rots and software entropies due to failing dependecies, costly bugs, broken environments and more. Prototypes become usually unusable quicker then well architectured software.

In relation to my research setting I can identify three different approaches to defining prototypes, which are entangled.

A simple and straight forward definition of a prototype would include producing something, with the help of which we can test a thesis, concept or process.

This makes them epistemic objects. Through their coming into being we can ask and sometimes answer all kinds of pragmatical questions. Prototypes as epistemic objects are approximations of something, enabling us to evaluate what is missing. Like a laboratory setup for other disciplines, prototypes become essential materialities for our research, without which we could only dabble in the conceptual. The technological and prototypical artefacts that we create are an indispensable support to our research and epistemological strategies. They contain knowledge that can’t be expressed otherwise and thus makes their longterm survival crucial.

Susan Leigh Star’s boundary objects move prototypes into social spheres. She proposes that technological artifacts build networks of relations with us. And, most importantly, that they poses their own kind of agency and are important actants, figuring as mediators in collective uncertainity management (Asante-Agyei, Manfredi, and Erickson 2022). Star’s boundary objects have communicative qualities, in that all actants can inscribe their individual interpretation into the prototype while maintaining a communal and collective identity.

An approach that incorporates these three main aspects is the approach of design researchers Bofylatos and Spyrou. In their view prototypial artefacts are essential parts of a processes of design semiosis.

“Design semiotics is the study of meaning in the specific field of design professions (industrial design, graphic design, architecture, among others) addressing the way in which designers and users make and share meanings by means of artefacts.“ (Mendoza-Collazos 2018)

An artefact then is not only a thing we produce to test a thesis, but is seen as an assemblage together with designers and users, which is well expressed in this model by Luna Gasparini. Here we see the artefact in an integrated loop together with its human actants.

At this moment our prototypes become object and method at the same time. It is a thing that we manifest for a specific goal, but it is also a communicative process that we have with our peers through this prototype.

“The artifact […] is thus to be conceived as a place where a “common mind” is shaped: a place of shared behavioural and cognitive habits, beliefs and opinions, compared and contrasted viewpoints or values, the place where shared meaning resides.” (Bofylatos and Spyrou 2017)

2. Contextualising PIA

In the following I would like to give form to this theoretical excursion by rooting it in PIA. This part here is taking from our funding application and states our general intention.

The common goal […] is to create a platform for participatory digital research on archival images and their metadata. This will involve, first, the development and evaluation of visual interface tools for contextualizing individual images and entire collections, and second, the implementation of an overall prototype.

I’ve already talked about prototypes and want to expand on the topic of participation now.

Participation as a concept and term is often contextual and subject to unspoken, possibly unconscious or semi-conscious assumptions (Bethmann, Hilgenböcker, and Wright 2020). In our project we often work with this model by Wright to locate where we would see a function or process of our platform.

As an institutionalised archive we are moving around closer to the lower middle of this model. Our reasoning for the current research project includes that moving higher might be beneficial for the archive, as well as participants. To that end we have different questions and prototypes we pursue.

We’re coming together from three different disciplines; Cultural anthropology, digital humanities and design research. What unites as all is the imaginary of a prototype. The problem arises from the what this prototype is, has to be or could be. The idea of it is a boundary object, that we are in constant discourse over. The challenges that arise from our organisational structures and needs can be directly linked to the definitions of prototype I carried out earlier.

We have the need for epistemic objects by our PhD candidates. We are also heavily reliant on prototypes as boundary objects that are mediating actants in our research project. And last but not least, we need to combat rots and entropies and in a best case scenario create something of lasting value, that survives the limited lifetime of this project.

3. Proposing participation as a strategy for sustainable prototypes

I wanted to give you these brief impression on the structure of PIA, because it substantially influences our methods and practice. We’re interested in prototypes, which are at method and output at the same time. This creates the problem of practice versus imaginary, which I have the feeling is something we all experience. Different actors within the project need or expect different things.

We try work through these differences by participatory approaches. As an interdisciplinary team, we are in constant exchange and discourse through which we try to let the other disciplines take part in our own. This is basically a communicative approach, in which shared vocabularies are created and knowledge is stabilised inter-disciplinary. On a more abstract layer, this can also be seen as attempts of creating interfaces between team-members and disciplines. These interfaces enable translation of discipline- and practice-specific processes and knowledge.

My thesis is, that the same approach to participation within the team as well as participants is also benefitial to ensure prototype sustainability. [This is an important point, so let us recap quickly. I introduced prototypical artefacts as important mediating objects that enable us to produce, share and stabilize knowledge. Knowledge that can’t be expressed otherwise. I also mentioned that, despite their importance, they are highly dynamic, instable objects that fall apart rather quickly.]

What I’m proposing now is to expand on the approach of seeing prototypes through design semiosis. Understanding them as mediating devices enables us to use communicative strategies. This should theoretically create access and participation in the prototype and leading to sustainable prototypes.

The key point is, to embed or weave them tighter into the networks that they are already part of. Ensuring epistemological participation stabilizes the objects, or at least snapshots of it, makes them accessable over longer time and opens up interfaces that enable translation of the knowledge captured within the object, now and later.

I isolated three different, but entangled rough strategies under the umbrella of using participatory processes as ways of make our prototyping more sustainable and stable. They expand on the approach by Bofylatos and Spyrou.

A real engineer once told me that one should always use the weakest technology to solve a problem. By weakest this person meant the easiest to implement and maintain, extend or replace. They usually come with little dependencies and added complexities. This approach is mainly geared against rots and entropies of any kind, trying to prevent that a prototype is becoming unusual through external factors too quick. I don’t have a concrete example for this, but I try to incorporate these thoughts in all my technological decisions. I’d rather use boring technologies and do mundane manual work instead of letting some magic framework take over that will introduce all kinds of problems within the next few years. Using the weakest technology lowers the expectation towards technical expertise and creates more access and participation on that level.

The second approach tries to expand on the first one. If prototypes are boundary objects that help share and stabilise knowledge within the research-team, then the choice of technologies need to reflect that. In our project we started out with a postresql database for it’s extended functionalities, but then moved to sqlite to reduce complexity. In cases where several team-members have to work with database layouts and content in a more programmatically way, but can’t code, we rely on CSV, which is a simple text file formated in a certain way. This might sound straight forward, but under the given factors in this presentation it is a concious choice.

No participatory approach would be complete without documentation. Documentation is everything and still neglected, even by me if I may be honest. In PIA we attempt to have a four-layered documentation model that resides on a spectrum from having technical literacy to wanting to be informed about the technological possibilities. The four layers in that order are

I want to lead you through these 4 layers with an example of documenting an API.

Readers of the technical docs need intimate knowledge and skills to understand what is in front of them. This level is often accessed by developers. Generally they need to know, how to solve a specific problems. Guides cater for the some public, but are often written linearly and introduce the reader to the scope, functionalities and abilities of a technology.

Project examples and data explorations are examples of what is possible. They mix accessible language and tech speak and enable a heterogenoues reader-group to get to know what is feasable of and how to solve certain problems. It is the best entry point for people at the beginning of their programmers journey and is geared towards strengthening coding literacy.

The public facing entry points ensure, that the documentation is found through a variety of public search queries, making it’s existence known. If sharing and collaboration is important to an institution having public facing entries is important.

This spectrum of documentation tries to involve as many relevant stakeholders as possible and to pick them up at their own level of expertise.

These are three strategies which I deem relevant to what I’m doing right now. There are certainly more and better ones and I’d be interested in hearing from you how you let people participate in your prototyping processes.

4. tl;dr

Let’s recap quickly to wrap this presentation up. I introduced prototypes not only as little experiment, but as an important epistemological object for our research. I followed with looking at prototypes through Star’s boundary objects, making them important nodes of exchange in our networks of research actants. This notion is expanded upon through Bofylatos and Spyrou approach of seeing prototypes as mediating objects in a process of design semiosis.

Given my remarks in this presentation, I want to propose participatory processes as beneficial for the sustainability of our prototypes. The more accessible an artefact is the more people can passively or actively participate in this design semioses. It is akin to treat the prototype as a communicative device that lets all participating actants share and stabilze their knowledge, not only in the prototype, but through it as well.

As part of PIA, and personally, I’m highly interested in producing things that last longer then the mere runtime of a research project. Digital technologies haven’t yet proven to be able to deliver the needed stability for that. Instead of concentrating on building massive and cryptographic monolith, I see much potential in creating artefacts that can be translated and transformed. The way to do that is to make them as accessible as possible through participatory approaches.

Asante-Agyei, Charis, Louise Manfredi, and Ingrid Erickson. 2022. “Unfolding the Future: Prototypes as Epistemic Objects in Innovation and Collaboration Work,” February.

Bethmann, Andreas, Elke Hilgenböcker, and Michael Wright. 2020. “Partizipative Qualitätsentwicklung in der Prävention und Gesundheitsförderung.” In Prävention und Gesundheitsförderung, edited by Michael Tiemann and Melvin Mohokum, 1–13. Springer Reference Pflege – Therapie – Gesundheit. Berlin, Heidelberg: Springer.

Bofylatos, Spyros, and Thomas Spyrou. 2017. “Meaning, Knowledge and Artifacts, Giving a Voice to Tacit Knowledge.” The Design Journal 20 (sup1): S4422–S4433.

Leigh Star, Susan. 2010. “This Is Not a Boundary Object: Reflections on the Origin of a Concept.” Science, Technology, & Human Values 35 (5): 601–17.

Mendoza-Collazos, Juan. 2018. What Is Design Semiotics?

Rheinberger, Hans-Jörg. 1997. Toward a History of Epistemic Things: Synthesizing Proteins in the Test Tube. Stanford: Stanford University Press.