Selected writing · Digital preservation
Organisation determines technology.
The 2023 FORGE paper makes one claim and repeats it three ways: technical persistence is what you get when personnel and funding persist inside a workable organisation.
Why it exists
Preservation is usually described as a technical problem.
Repositories are judged by their formats, migrations and certifications. All of that matters. But the paper, written with Gunter Vasold and Elisabeth Steiner, starts from a different observation: the systems that survive are the ones whose organisation survives.
The argument is not that technology is unimportant. It is that technical debt is often organisational debt wearing a technical costume.
The argument
Three theses, one direction.
-
Change
Every used system must change, because assumptions change, not only technology.
-
Degrade
Every used system degrades; uncontrolled change destroys the ability to change.
-
Conway
Every used system takes the shape of the organisation that builds it.
The first thesis makes software a social construct. A system is not a product delivered once; it is a set of decisions that are renegotiated. If changeability is not in the architecture from the start, long-term operation fails. Preservation infrastructure is the hardest case, because its promise is duration.
The second thesis is about entropy. Technical debt cannot be avoided, only minimised. A team that does not systematise change loses the ability to make it. The end state is not a crash; it is a system nobody dares touch.
The third thesis is Conway’s Law applied to infrastructure. Organisational change creates technical change pressure, sometimes abrupt, more often creeping. Ignored pressure produces workarounds, and workarounds are how an infrastructure stops matching its requirements.
The paper’s practical section is about people. Technical teams are small, so the bus factor is low. Decisions should be made in the group; younger members prevent blindness, older colleagues act as a corrective. Flat hierarchies and an open argument about architecture work. Improvised fixes produce dissatisfaction and conflict. Stakeholders outside the team set the resources, so the exchange with them is part of the engineering.
The conclusion is an aphorism worth keeping: personnel and financial persistence in an adequate organisational structure produce technical persistence.
What holds in 2026
More true now, not less.
Project funding is shorter, staff turnover is normal, and AI has added a new source of change pressure. Every one of those is an organisational variable.
The paper’s weakest point is its evidence: it argues from experience, not from a sample. That is honest, and it is also why the claim needs to be tested in more institutions.
What I would add: handover as a first-class artefact. A repository that cannot be handed over has not been preserved, whatever its certification says.
And a fourth thesis, carefully: documentation is an organisational artefact. It ages when the organisation changes, not when the software does.
The record
The deposit is the conference paper: three theses, one aphorism. The German title is the deposited one.