Research paper

Download the research paper (PDF)

Evidence cut-off: 5 September 2026.

Archaeological records may be the surviving evidence

Excavation changes or destroys the original stratigraphic context. Field registers, photographs, plans, samples, databases and interpretive decisions can therefore become the only durable record of relationships observed in the ground. Their preservation is part of archaeological stewardship.

Digitisation and preservation are different

Scanning creates a digital representation; preservation keeps information intelligible and authentic across hardware, software and organisational change. Files on one drive or website can disappear through media failure, obsolete formats, lost credentials or institutional closure.

Plan before data creation

A data-management plan should define responsibilities, file naming, identifiers, formats, metadata, version control, quality checks, rights, access categories, storage, deposit and cost. Retrofitting these choices after a project ends is expensive and often cannot reconstruct missing provenance.

Preserve context, not isolated objects

A photograph becomes evidence when it is linked to a site, trench, context, feature, object or sample, creator, timestamp, scale, orientation and processing history. Relational links among registers, drawings, laboratory results, publications and physical collections should survive export from the original software.

Metadata enables discovery and interpretation

Descriptive metadata supports discovery; administrative metadata records rights and custody; technical metadata identifies format and software; preservation metadata records fixity and actions; structural metadata describes relationships among files. Controlled vocabularies and stable identifiers improve multilingual and cross-repository search.

Use preservation-friendly formats

Open, documented, widely implemented and validation-friendly formats generally reduce risk. Archive the highest-quality preservation master and derive smaller access copies. Original files may also be retained when they preserve significant functionality, alongside exports suitable for future migration.

Integrity needs active checking

Checksums can show whether a bitstream has changed, but they do not prevent loss. Repositories need multiple independently managed copies, routine fixity checks, audit logs, controlled permissions, tested recovery and monitoring for format obsolescence. Backup alone is not a preservation programme.

Persistent identification and citation

Stable identifiers should resolve to landing pages even when storage locations change. Dataset citations credit creators and allow readers to reproduce claims. Versioned releases, change histories and links between data, reports and corrections keep scholarly use traceable.

FAIR requires stewardship

The FAIR principles ask that data be findable, accessible, interoperable and reusable. Accessibility may include authenticated or mediated access rather than universal download. FAIR describes qualities of data; it does not by itself guarantee long-term preservation, justice or openness.

Access must be ethically tiered

Exact coordinates can expose vulnerable sites to looting. Archives may also contain personal information, sacred or restricted knowledge, images of human remains and data governed by source communities. Public metadata can coexist with redacted, embargoed or permission-based files when restrictions and decision processes are documented.

Communities need authority as well as visibility

Digitising community heritage without consent can reproduce extraction. Governance should include consultation, culturally appropriate labels, language access, attribution, correction mechanisms and authority over sensitive material. Reuse licences cannot erase moral rights or living relationships.

Multilingual discovery is core infrastructure

South Asian names vary across scripts, languages, transliterations and historical spellings. Archives should retain original-script forms, language tags and alternative labels while mapping them to stable entities. Search interfaces must not make English transliteration the only route to discovery.

A minimum viable public archive

Begin with a collection register, unique identifiers, open metadata, rights statements, durable master files, checksums, redundant storage, documented deposit workflows and persistent landing pages. Add APIs, IIIF images, linked vocabularies and 3D access where resources and evidence needs justify them.

Sustainability is institutional

Digital preservation requires recurring staff, storage, security, governance and succession planning. Deposit agreements should state who can act if a project or institution closes. Trusted-repository assessment, transparent policies and preservation reports make continuity auditable.

Evidence assessment

Strong: Digital archaeological records require active preservation, metadata and provenance; storage or scanning alone is insufficient. Strong: Persistent identifiers, sustainable formats, fixity checks and redundant copies reduce distinct preservation and reproducibility risks. Probable: Open, linked and multilingual metadata substantially improves discovery and responsible reuse across fragmented collections. Contested: The appropriate balance between public access and restrictions for sensitive places, knowledge and human remains must be decided case by case with affected communities. Unsupported: The assumption that uploading files to a website guarantees their authenticity, intelligibility or long-term survival.