73 How to avoid misrepresentations of data

< Back to problem stories

I am Principal Investigator in a project (PENELOPE, funded by ERC, Deutsches Museum) where the key objective is to explore what is categorised as tacit knowledge, in (ancient) weaving. We make a claim that mathematical knowledge, was itself abstracted from weaving principles in Ancient Greece. In order to explicate such knowledge as being rational and technological, we show coding, algorithms, and numbers implicit in weaving practices. Our problem for data management is that, even though we developed a lot of experiments (live coded looms, robot swarms dancing around a maypole, documenting tacit technological conversations of weavers at looms), only the interaction of all of them can eventually be understood as the point we make for weaving knowledge. Once we move forward, the object cannot solve the problem, and when placed in the public domain can end up misrepresenting what the research outcomes are. How do we avoid this situation? In order to generate the necessary insights, we get users to experience the nature of this knowledge, creating analogies through different algorithmic practices – in music, in computers, on looms. We will set up a final workshop/exhibit and make videos of such experiments as documentation. However, the data processing when weaving becomes available only when the object is in motion, in use. The information is complete only when there is actual engagement with the material objects of our project. How do we save this experiential component, which is available in the project, as data structure? If we put this into a data storage facility, we fall into the trap of creating a new graveyard for weaving knowledge. How do we avoid this trap? How can we actually mark points of ‘missing’ data connections?