Current landscape is characterised by dispersal, fragmentation and unresolved tensions

Digital heritage archives today exist in a highly dispersed and heterogeneous ecosystem. The landscape comprises data repositories, regional heritage portals, large-scale aggregators, project-specific portals, domain-specific databases, and institutional heritage information systems — with limited interoperability between them. A significant and under-acknowledged portion of digital heritage data remains on local servers and old computers, outside any formal repository infrastructure.

Gap between aims and actions

In the contemporary heritage digitisation practice, one of the major challenges is that the direction of attention is typically directed to produce digital artefacts for difficult to anticipate future use. A typical way of thinking is to try to guess at the present what are the future needs regarding the technical accuracy of digital (re)presentations and their documentation and develop techniques and procedures that would makle it possible to preserve the digital artefacts and documentation into the future. If any outspoken aims besides preservation are formulated, they are often abstract such as societal cohesion, well-being, competitiveness and healthy cultural identity. What remains repeatedly unspelled is what practical steps needs to be taken to link the distant goals and present practices of heritage digitisation.

Well-known structural weaknesses

At the same time as heritage digitisation suffers from a gap between goals and actions, the sector is held back by three interconnected structural weaknesses that all have been recognised for a long time ago. First, repositories remain dispersed, with no common infrastructure in place for long-term digital archival storage. Second, standards are either lacking or inconsistently applied — shared metadata schemas exist in principle, but coherent implementation across institutions is far from the norm. Compounding this, there is no shared understanding or recognised understanding of quality, leaving users without reliable means to assess the reliability, completeness, provenance, or general goodness or badness of digital heritage datasets.

As a result no comprehensive digital archival infrastructure for long-term storage of digital and digitised heritage exists. A lot of work has been done and is on-going to remedy the situation but as whole, we are not there yet. The reasons are largely structural: adequate infrastructure demands sustained investment, yet the recurring costs of digital preservation have never been systematically built into project funding models or institutional budgets.

Infrastructure as condition, not feature

A critical conceptual shift is needed in how we think about digital heritage systems. Susan Leigh Star’s observation that “it’s infrastructure all the way down” is instructive: digital heritage practice does not simply involve a user meeting a screen, but a user meeting a complex, layered infrastructure. This infrastructure shapes what can be found, what can be used, what can be sustained — and equally, what disappears. Treating infrastructure as an afterthought rather than a foundational concern has added to the sector’s fragility similarly to projectification, or the organisation of heritage work into discrete, time-bounded, externally funded projects. It is amplified by digital platforms designed around project cycles rather than long-term institutional stewardship. The consequence is a steady accumulation of data whose long-term care has neither been planned nor resourced.

The nature of the problem is in the dichotomy between digital logic and heritage complexity

The core tension in digital heritage is in that the digital is the art of either/or whereas heritage is the art of the both and beyond. Digital systems require formalisation, discretisation, and standardisation. Heritage practice, by contrast, is built on multiplicity of interpretation, contextual meaning, incompleteness, and the irreducibly analogue nature of cultural objects and practices.

Heritage data is characteristically marked by several qualities that set it apart from more standardised forms of information. It is highly heterogeneous, ranging from small and relatively simple datasets to highly complex, multi-layered assemblages. It is also epistemically diverse, having been produced across mutually different disciplinary cultures, each with their own norms of evidence and documentation. Beyond this, heritage data is deeply contextually embedded: its meaning depends on provenance, on practice, and on the interpretive community in which it was created and used. Finally, it is variably complete, and crucially, the degree of its incompleteness is often neither communicated nor even fully knowable.

Digital systems, for their part, present their own distinctive set of challenges. Unlike physical objects, whose deterioration tends to be gradual and visible, digital systems are fragile in ways that demand active, ongoing care simply to sustain what already exists. Access to them is technology-dependent, relying on rapidly evolving technological stacks that can quickly become obsolete. They are also selectively accessible — straightforward to ingest in certain contexts, yet difficult to bring into use in others. And structurally, they tend toward rigidity, alternating between miscellaneous and highly organised modes with little consistency in how content is arranged or governed.

Emerging data risks of homogenisation and data colonialism

Repeated attempts to solving the challenges in the heritage field without adeuately addressing the dichotomy between digital logic and heritage complexity risks to intensify rather than to help solving the problems in the long run. Two major threats of forcing heritage to follow digital logic are spelled homogenisation and data colonialism.

Homogenisation, the flattening of diverse heritage practices into a single dominant model, is encouraged by large-scale aggregators and shared metadata standards that were designed with particular epistemic cultures in mind. The multiplicity of expression is not necessarily a sign of sloppiness but rather than of a multiplicity of meanings  also in heritage data. When that multiplicity is suppressed in the name of interoperability, something essential risks to be lost.

The heritage sector has abundant standards, yet a question remains whether we really know how to use them. Research on data-documentation standards shows that standardisation reshapes not just the outputs of heritage work but the processes that produce them, often in ways that are invisible to practitioners and policymakers alike. Blind systematisation is lethal to heritage data. The push to make data “clean” can strip away significant parts of contextual information including paradata — the contextual and procedural information about how data was produced  — that makes it interpretative. in parallel, as Börjesson and colleagues have shown, in spite of the apparent advantages of systematisation, it is crucial to ask whether data can be made too clean.

A parallel risk concerns what scholars have called data colonialism: the question of whose networks, whose categories, and whose infrastructures come to define what counts as heritage and how it is preserved. Decisions about infrastructure are never merely technical — they are also decisions about power, about whose data is made accessible, on whose terms, and in whose interest. It is not a question of what is often discussed in terms of bias – that data is not neutral or completely representative of everything – but rather about failed attempts to acknowledge and make it transparent, and defend those in the margins.

Toward sustainable digital heritage practice

Unlike often highly technical heritage digitisation discourse might seem to suggest, sustainable digital heritage practice is not primarily a technical problem. In the long run, it is an information problem of passing on and keeping alive tangibles and intangibles of the past and the present letting them evolve so that they make sense in the future. At the same time, at the present it is a problem of values, governance, and culture. While the inherent wickedness of the problem means that there is no one recipe that would ensure success across all conceivable heritage contexts, there are apparent ingredients that would benefit of being taken more seriously by every heritage institution, funder, and policymaker.

Data is not enough: paradata is needed

Information is increasingly a process issue. Data without paradata — the contextual information about the practices, decisions, and conditions that produced it — is of limited value. Heritage data literacy must therefore go beyond technical skills and subject knowledge to encompass an understanding of the practices underpinning data creation, curation and use. Infrastructure design must make room for the documentation of process, not just outputs.

Frameworks that avoid dead ends

Heritage institutions need frameworks — not rigid prescriptions, but flexible structures that help practitioners navigate choices without foreclosing future options. Given unknown future technologies, access patterns, and user communities, decisions about formats, platforms, and standards should be evaluated for their reversibility and their long-term implications, not just their immediate cost-efficiency. The goal should be to avoid digital dead ends that lock data into systems that cannot be sustained.

Harmonisation with humility

The digital heritage future requires harmonisation — but harmonisation that actively cares for diversity rather than suppressing it. Standards and shared infrastructures are necessary but must be designed with epistemic humility: an acknowledgement that the categories and structures built into systems carry assumptions that may be culturally specific, temporally limited, or otherwise partial. Certification and quality frameworks should reward contextual richness and to a reasonable extent even difficult to manage messiness, not penalise it.

Formalisation versus flexibility

The tension between formalisation and flexibility is not resolvable by choosing one over the other. Rather, what is needed is institutional and technical arrangements that can hold both. The arrangements need to be formal enough to enable sharing and sustainability, flexible enough to preserve the interpretive richness that gives heritage its value. The real gap in digital heritage is not between object and text, but between the object and its formalised representation — and that gap must be acknowledged, documented, and managed rather than papered over.

Switching perspective

Finally, solving the discrepancy with desired futures and current actions requires a shift of perspective. Rather than trying to predict what is needed in the undetermined and fundamentally unknown and unknowable future, it is necessary to switch perspective. We need to decide when we would like to see something specific to happen, mentally shift to that point in time and look backwards to imagine what steops would need to be taken to advance from the current state of affairs to the desired future and design future actions on the basis of those steps.

Recommendations

What then could be recommended? Heritage institutions and policymakers should consider at the least the following:

• Invest in shared, sustainable digital archival infrastructure as a public good, with long-term funding models that go beyond project cycles.
• Develop and recognise certificates of quality for digital heritage repositories that account for paradata richness, contextual documentation, and long-term access planning — not only technical compliance.
• Reform documentation standards to capture process, not only outputs, embedding paradata requirements into funding conditions and professional guidelines.
• Design interoperability frameworks that preserve diversity, ensuring that aggregation and harmonisation do not systematically disadvantage minority, indigenous, or non-dominant heritage traditions.
• Promote data literacy in the heritage professions that encompasses understanding of documentation and information use as generative practices, not just technical skills or subject expertise.
• Establish governance mechanisms that address data colonialism [#Couldry2019a] risks, ensuring that decisions about infrastructure, access, and categorisation are made transparently and accountably.
• Switch perspective from looking forward to an unknown future to look back from a desirable future to this day, to determine what steps need to be taken to make the desired future.

Conclusion

Sustainable digital heritage practice is the global digital future in the making — not as a metaphor, but as a concrete institutional and technical challenge that demands immediate attention. The choices made today about infrastructure, standards, documentation, and governance will determine what of our shared cultural memory survives, who can access it, and on what terms. The challenge is not to make heritage fit the logic of digital systems, but to build digital systems capable of holding the complexity, diversity, and meaning of heritage. A digital heritage future worth having is one built on harmonisation that cares for diversity, and on humility before an unknown future.

Acknowledgements

This text is based on a keynote lecture presented on May 26, 2026 at the Digital Heritage Summit 2026 organised under the auspices of the Cypriot Presidency of the Council of the European Union and the Cypriot Deputy Minister of Culture by the UNESCO Chair on Digital Cultural Heritage at the Cyprus University of Technology, in partnership with Heritage Malta in Limassol, Cyprus.