Loading field map
Digital curation is the active management and preservation of digital information over its entire lifecycle, from conceptualization and creation through active use, long-term access, and—when appropriate—deletion or transformation. It is a field of practice and research within information science that addresses a central problem: digital objects are fragile, formats become obsolete, storage media degrade, and the contexts that make data meaningful can be lost. Without deliberate intervention, most digital information becomes inaccessible within years or decades, even when it remains physically intact.
The term "curation" was borrowed from museums and archives, where it implies selecting, organizing, and maintaining collections with an eye to their future value. Digital curation extends this idea to the full range of digital materials—research data, government records, cultural heritage objects, software, personal archives, and the outputs of scientific instruments—and to the entire span of their existence, not just their final disposition. The field is distinguished from digital preservation, its close relative, by this lifecycle orientation. Preservation typically concerns the technical and organizational measures needed to keep digital objects usable over time; curation encompasses those measures but also includes the intellectual work of appraisal, selection, annotation, and ongoing management that determines what is kept, how it is described, and how it remains meaningful.
Digital curation emerged as a distinct concern because digital information behaves differently from analog information. A printed book can survive for centuries on a shelf with minimal intervention; a digital file requires a functioning storage medium, a readable file format, compatible software, and a hardware platform capable of running that software. Each of these layers is subject to failure, and the failure of any one renders the object inaccessible. This is often described as the problem of technological obsolescence, but it is more accurately a cascade of dependencies: media degrade, devices disappear, formats fall out of use, and the semantic context—the documentation, metadata, and tacit knowledge that explains what the data means—is frequently lost first.
The field's central questions follow from this fragility. What digital materials are worth keeping, and who decides? How can the authenticity and integrity of digital objects be demonstrated over time, especially when they must be migrated to new formats or systems? What metadata is necessary for future users to understand and trust a digital object? How should curation decisions be made when resources are finite and the volume of digital information is vast? And how can the costs and responsibilities of long-term stewardship be distributed across institutions, disciplines, and generations?
A further set of questions concerns the relationship between curation and use. Digital objects are not static artifacts; they are often computational, interactive, or dependent on software environments that themselves change. A dataset may be continuously updated, a video game may require a specific operating system, a scientific simulation may need particular hardware. Curation must therefore decide not only what to preserve but what form of the object is worth preserving—the original bits, a normalized version, a functional emulation, or a documented description of what the object did.
The practices that feed into digital curation have deep roots in libraries, archives, and museums, where selection, description, and preservation have long been core responsibilities. However, the specific challenges of digital materials began to receive systematic attention only in the late twentieth century, as institutions realized that their growing holdings of electronic records, databases, and digital publications were at risk. Early work in the 1970s and 1980s focused on machine-readable data files, particularly in the social sciences, where data archives such as the Inter-university Consortium for Political and Social Research (ICPSR) developed procedures for documenting and preserving survey data. National archives and standards bodies also began confronting the problem of electronic records, leading to early frameworks for managing records that existed only in digital form.
The term "digital curation" itself came into wider use in the early 2000s, particularly through the work of the Digital Curation Centre (DCC) in the United Kingdom, founded in 2004. The DCC and related initiatives gave the field a name, a research agenda, and a professional identity, distinguishing it from the narrower technical focus of digital preservation. The growth of data-intensive science, often described as e-science or cyberinfrastructure, was a major driver: researchers in fields like genomics, astronomy, and climate science were producing datasets so large and complex that their management required explicit attention, and funding agencies began requiring data management plans as a condition of support.
The development of the field has been shaped by several influential frameworks. The Open Archival Information System (OAIS) reference model, standardized in the early 2000s, provided a common vocabulary for describing the functions of a digital archive, including ingest, archival storage, data management, access, and preservation planning. The OAIS model is not a curation framework per se, but it gave institutions a way to talk about their responsibilities and to design systems that could interoperate. The DCC Curation Lifecycle Model, developed in the mid-2000s, mapped the full range of curation activities from conceptualization through disposal, emphasizing that curation is continuous rather than a set of discrete steps. The Data Curation Profiles and similar tools attempted to capture the specific needs of different research communities.
Digital curation is not organized around a single paradigm or a sequence of rival schools. It is better understood as a field held together by a shared problem, with different traditions emphasizing different aspects of that problem. These traditions overlap, borrow from one another, and often coexist within the same institution. Nevertheless, several distinct approaches can be identified.
The oldest and most institutionally grounded approach to digital curation comes from archives and records management. This tradition is organized around the concepts of provenance, original order, and authenticity. Its practitioners are concerned with the legal and evidential value of records—documents that serve as evidence of business transactions, government actions, or personal activities. The central question is not simply how to keep digital objects usable but how to ensure that they remain trustworthy as records: that they are what they claim to be, that they have not been altered, and that their context of creation is documented.
This tradition developed methods for appraising records (deciding which have enduring value), for capturing records in ways that preserve their structure and relationships, and for documenting the chain of custody through which records pass. It has been particularly influential in government and corporate settings, where regulatory requirements demand demonstrable authenticity. The archival tradition tends to be cautious about transformation: it prefers to preserve original digital objects in their native formats, even when those formats are obsolete, and to maintain the metadata that links records to their creators and functions. Its limits become apparent with born-digital materials that do not fit traditional record categories—databases, websites, social media, or scientific data—where the notion of a discrete "record" is difficult to apply.
A second major tradition grew out of the sciences and the data archives that serve them. This approach is organized around the concept of the dataset as a research output with potential for reuse. Its practitioners are concerned with data quality, documentation, and interoperability—ensuring that data can be understood and used by researchers other than those who created it. The central question is how to make data "FAIR": findable, accessible, interoperable, and reusable, in the acronym adopted by a 2016 set of guiding principles that have become widely influential.
This tradition emphasizes metadata standards, data citation, and the development of disciplinary repositories where data can be deposited and shared. It tends to be more willing than the archival tradition to transform data into preferred formats, to normalize it, and to prioritize future usability over original form. It is closely connected to the open science movement and to the policies of research funders, which increasingly require data management plans and data sharing. Its limits appear when data are complex, sensitive, or poorly documented: the FAIR principles address discoverability and accessibility more than long-term preservation, and the assumption that data can be separated from the software and hardware that produced them is often problematic.
A third tradition focuses on the technical mechanisms for keeping digital objects usable. This approach is organized around the concept of the digital object as a sequence of bits whose interpretation depends on hardware, software, and documentation. Its practitioners develop and evaluate strategies for media migration, format conversion, emulation, and bit-level preservation (the maintenance of exact copies of the original bits). The central question is how to ensure that a digital object can be rendered and understood at some future time, given that the original technical environment will not persist.
This tradition has produced the most concrete and testable results in the field. It has developed checksums and fixity checking to detect bit rot, the Open Preservation Foundation's format registries and risk assessments, and emulation frameworks that recreate obsolete software environments. It has also produced the significant insight that preservation is not a one-time event but a continuous process of monitoring and intervention. Its limits are equally clear: technical preservation cannot solve the problem of meaning. A perfectly preserved bitstream is useless if no one knows what it represents, and an emulated environment is of limited value if the documentation explaining how to use the object has been lost.
A fourth tradition, more recent and more diffuse, draws on the humanities and on scholarly communication. This approach is organized around the idea that digital objects are cultural and intellectual artifacts whose value is not fixed but is continually renegotiated. Its practitioners are concerned with the interpretive labor of curation—the decisions about what to collect, how to describe it, and how to present it to different audiences. The central question is not only how to preserve digital objects but how to curate them in ways that shape their meaning and use.
This tradition has been influenced by museum studies, by the digital humanities, and by critical information studies. It emphasizes the social and political dimensions of curation: who gets to decide what is preserved, whose stories are told by the archives we build, and how the categories and metadata we use reflect particular worldviews. It has also engaged with the problem of "dark data"—the vast quantities of digital information that are collected but never used—and with the ethical questions raised by preserving data about people who did not consent to its long-term retention. This approach is less concerned with technical solutions than with the conceptual frameworks that guide curation decisions, and it is often critical of the assumption that curation is a neutral, technical activity.
These traditions are not mutually exclusive, and most practitioners work across them. A research data curator in a university library might use the OAIS model (from the archival tradition) to design a repository, apply FAIR principles (from the data science tradition) to develop metadata, employ format migration and fixity checking (from the preservation engineering tradition), and make selection decisions that reflect the values of the research community (from the curatorial tradition). The field's textbooks and training programs typically present all of these perspectives as complementary.
The relationships among the traditions are nonetheless marked by productive tensions. The archival tradition's emphasis on authenticity and original form can conflict with the data science tradition's preference for normalization and transformation. The preservation engineering tradition's focus on technical mechanisms can seem indifferent to the social questions raised by the curatorial tradition. And the curatorial tradition's critical stance can appear impractical to practitioners who must build working systems. These tensions are not signs of incoherence; they reflect the genuine complexity of the field's subject matter. A digital object is simultaneously a physical artifact, an intellectual work, a legal record, and a social construction, and no single tradition captures all of these dimensions.
Digital curation today is an established but still evolving field. It has professional associations, degree programs, conferences, and a substantial body of research literature. It is practiced in national libraries and archives, in university libraries and data repositories, in government agencies, in museums, and in the private sector. The field's institutions include national initiatives such as the Digital Curation Centre in the UK, the National Digital Stewardship Alliance in the United States, and similar bodies in other countries, as well as international standards organizations and disciplinary data networks.
Several durable challenges define the current landscape. The first is scale: the volume of digital information produced each year vastly exceeds the capacity of any institution to curate it, and the field has not yet developed convincing strategies for selection at scale. The second is the problem of software and computational environments: an increasing proportion of digital objects are not static files but dynamic systems—websites, applications, simulations, machine learning models—that cannot be preserved as simple bitstreams. The third is the question of sustainability: long-term preservation requires ongoing financial commitment, and the institutional models for funding curation over decades or centuries remain fragile. The fourth is the ethical dimension: as digital curation becomes more capable, questions of privacy, consent, and the right to be forgotten become more pressing.
The field has also been shaped by the rise of artificial intelligence and machine learning, which both create new curation challenges and offer new tools for curation. AI systems can assist with metadata generation, format identification, and the detection of sensitive information, but they also produce outputs—models, training data, and the processes that generate them—that are difficult to preserve and document. The relationship between human judgment and automated curation is an active area of debate.
Digital curation remains a field whose practitioners are acutely aware of the provisional nature of their work. Every preservation strategy is a bet about the future: that a format will remain readable, that a metadata standard will remain meaningful, that an institution will remain funded, that future users will care about the objects being kept. The field's intellectual contribution is to make these bets explicit, to develop methods for improving their odds, and to keep open the possibility that future generations will find value in what we choose to keep.