Designing technology for people

When planning for the Heinz Nixdorf Institute began over 30 years ago, it was clear that the seven endowed professorships would include a department of Computer Science and Society. To this day, it is undisputed that students, especially engineering students, should have a general grounding in ethics and knowledge of the effects and impact of their products.

However, the question of what kind of research methodology and research subject such a department of computer science should have remains largely open to this day. The heterogeneity of social problems and the associated different competence requirements across different disciplines made it difficult to develop a coherent self-image.

When this position was filled in 1992, the establishment of a suitable research approach was on the agenda. Instead of focussing on the almost immeasurable and diverse effects of computer science on society, the research perspective was focused on the concept of interactions between computer science systems and their development and application environment. Why are interactions so important? For many engineering disciplines, technology embodies materials science, i.e. the scientific characteristics and properties that are important for the design of technical devices and systems. The behaviour of the material is independent of the expectations and assumptions of both users and developers.

This is different in computer science. Its "material", from which the systems are constructed, is text. Symbolic descriptions are used on the one hand for communication between users and developers and for modelling the problem area, and on the other hand for formulating machine operations whose formal execution is used to control complex processes in technology, business and administration. As part of the modelling process, this material is imprinted with a wide range of assumptions that relate to human behaviour to a far greater extent than is the case in other engineering fields. In these cases, the behaviour of an IT system is no longer independent of human behaviour and this changes with the use of the software, the associated learning processes and other changes in the operational environment. As a result, the model is no longer adequate and has to be adapted - new versions are created. These interactions need to be recognised and understood at an early stage so that they can be taken into account as far as possible when developing a version.

A decisive difference to the term "impact" is that interactions have a retroactive effect, i.e. they have an influence on the IT systems and must therefore be appropriately recognised when designing them. Thus, a department within computer science was established that focuses on those effects that can be influenced by means of computer science. As a result, the research was primarily concerned with the question of how incorrect assumptions in the modelling and development of systems can be avoided from the outset as part of a hypothesis-driven technology design or how development processes can be designed in such an open way that errors can serve as a means of knowledge and thus flow into a continuous further development process - for example through evolutionary system design. The change of denomination to "Contextual Informatics" was intended to emphasise the associated focus on interactions in the context of production and use with regard to research activities.

In order to be able to focus on the corresponding interactions, we chose the support of knowledge work by computers as a subject area, as it manifests itself particularly concisely in use (software ergonomics), learning (e-learning), cooperative work (CSCW) or even today in the digital humanities (e-humanities). In all these processes, where the focus is on modelling the support of human mental activities and not on modelling the cognitive processes themselves, conceptual and methodological foundations are needed in order to derive and test design hypotheses.

This requires a two-stage approach. Firstly, technical potentials must be identified on the basis of a contrastive conceptualisation. Contrastive does not mean comparing humans and machines as if they were functionally equivalent symbol-processing systems, but rather looking at the differences in the processes that make them what they are (evolution vs. construction). The aim is to replace human weaknesses and support their strengths. On the other hand, such a view must be linked to questions of practicability in the context of use, because potentials that are not designed to suit the situation or are not understood in use cannot develop any added value. Here, aspects such as suitability for everyday use, sustainability, universal availability and expandability or modifiability take centre stage. From an ethical perspective, it has always been crucial for us to increase the number of usage options and improve the implementation of the respective technical solutions.

This contextual perspective has enabled us to design and implement solutions that are clearly different from the "mainstream": instead of relying on interactive learning computers in the 1990s, we developed infrastructures that promote learning; instead of propagating the use of high-quality teaching materials on CD, we were the first in Germany to use the WWW in courses. We constructed real and virtual learning and working spaces that supported co-operation as well as hybrid laboratories. We deepened the experience we gained - mostly in interdisciplinary cooperation - in the context of a wide variety of fields of application, ranging from regional educational networks and in-company training to historical research on world cultural heritage and the creation of historical-critical digital music editions. Despite the many associated successes, the department seems to have outlived its usefulness, at least in terms of research. On the one hand, many research questions are now considered in the context of the respective application scenarios, such as "social media" (algorithm watch, socioinformatics). On the other hand, neighbouring disciplines, such as business informatics, are also increasingly turning to empirical approaches for hypothesis-driven technology design.

Regardless of how this development continues, as part of our activities at the Heinz Nixdorf Institute, we have built up technical infrastructures that have far outlasted our research purpose and are still in use today in the region (bid-owl), in the city (Lernstatt Paderborn) or, until recently, at Paderborn University (koaLA).