Putting a collection online means turning physical objects into files that stay findable, licensed and readable for decades – the scan is only the visible part. Most of the work sits in metadata, rights clearance, shared delivery standards and long-term preservation. Each stage feeds the next, so a collection that skips one of them tends to end up technically online but practically unusable.
What putting a collection online actually involves
Museums, libraries and archives usually treat digitisation as a pipeline rather than a single project. Five stages do most of the work.
- Selection: staff decide what is worth the cost, based on condition, demand, rights and whether the original is too fragile to keep handling.
- Capture: flatbed scanners, copy stands, camera rigs and 3D scanners produce high-resolution files under controlled lighting, with colour targets so colours can be checked later.
- Catalogue and metadata: every file is tied to a record describing what the object is, where it came from and who may reuse it.
- Delivery: a smaller web-friendly copy is published through an image server, so pages load quickly without exposing the master file.
- Preservation: masters are stored, copied and checked so they survive hardware and software changes.
Capture quality is limited by the decision made before it. A file delivered at a lower resolution than the object deserves cannot be improved afterwards, which is why institutions keep an uncompressed master and generate access copies from it.
Metadata is what makes a scan findable
A file with no description is effectively invisible: search engines, library catalogues and research tools all depend on structured text attached to the image. Cataloguers record the title, date, place, creator, materials, dimensions, provenance and rights status, then attach terms from controlled vocabularies. Those vocabularies matter because two institutions may describe the same object differently – one as a “bowl”, another as a “vessel” – and a shared list keeps both records discoverable. The Getty vocabulary programme maintains structured lists for object types, geographic names and artist names that institutions reuse instead of writing their own: getty.edu.
Two further pieces of plumbing matter. Persistent identifiers give each object a permanent address even if the website is rebuilt, and crosswalking standards let a record exported from one system be imported into another without losing its meaning.
Rights and licensing decide who can reuse what
Copyright status is assessed object by object, because the rights in a photograph, a manuscript and a recorded interview are held differently. The result is expressed as a rights statement and, where possible, as a machine-readable label so a website or a researcher can tell reuse at a glance.
| Rights status | What it usually allows |
|---|---|
| Public domain | Reuse without permission, including commercially |
| Open licence such as CC0 or CC BY | Reuse under the licence terms, usually with credit |
| In copyright, rights holder identified | Reuse needs a licence or permission from the holder |
| Rights unclear or orphaned | Often displayed with downloads restricted while research continues |
Institutions also use take-down and notice procedures so a mistaken rights call can be corrected without removing the record entirely.
Delivery and preservation: standards that keep files usable
Delivering images at scale is easier when institutions agree on a common description format. The International Image Interoperability Framework defines how an image server publishes an object and its images, so many viewers, annotation tools and comparison interfaces can read the same collection without a bespoke integration for each one: iiif.io.
Why storage is not preservation
Keeping a copy on a server is not the same as preserving it. Preservation adds scheduled checksums to detect silent corruption, multiple copies on different media and in different places, written plans for migrating formats as software ages, and documentation detailed enough for someone else to understand the files years later. The Digital Preservation Coalition publishes guidance and training for exactly these routines: dpconline.org.
Web pages are part of the collection too
Galleries and national libraries increasingly archive websites as well as objects. Legal deposit rules in some countries extend to digital publications, crawlers capture pages into archive files, and the resulting collection is what keeps an old government page or a defunct news site citable.
The bottom line: a digitised object is useful only when the file, its description, its rights statement and a preservation plan travel together. Scanning is the cheapest and most visible step; the standards and record-keeping around it are what keep a collection usable after launch.