Link, Screenshot, or Full Capture? Choose the Right Reference Fidelity

Each format preserves different evidence

A link is the smallest reference. It preserves the route back to a live source and may be all you need for a stable document. Its weakness is dependence: the page can change, move, require access, or disappear. A bookmark preview can make the link easier to recognize, but the preview does not remove that dependence.

A screenshot preserves visible pixels at one moment, viewport, and state. It is strong evidence for composition, spacing, color, or a transient interface. It does not preserve interaction, responsive behavior, text outside the frame, underlying assets, or source code. Dropmark’s support documentation explicitly distinguishes a normal link from a screenshot or image grab, illustrating why the formats should not be conflated.[1]

A fuller capture aims to preserve more of the source material. Web archives demonstrate that this is a process, not a bigger screenshot: collection scope, crawling, replay, access, and preservation all introduce constraints. The Library of Congress explains that archived sites may be incomplete and that rights and access conditions still matter.[2]

Use the lightest fidelity that can support the decision

For a reading list, keep the link and a note. For a button transition, record the source plus motion or state details. For a layout likely to change before critique, add a dated screenshot. For evidence that must remain available over time, evaluate a lawful archival method and a backup plan.

• Link: source discovery, citation, or planned revisit.
• Preview: recognition and triage while the live source remains authoritative.
• Screenshot: a particular visual state, viewport, or moment.
• Fuller capture: long-lived access where scope, storage, and permissions have been considered.

Keep provenance independent from the file

Whatever you save, keep the original URL, title, publisher or creator, capture date, and a description of the relevant state. The W3C Web Annotation model separates a body—the comment or annotation—from its target, which is a useful mental model: your interpretation and the source it describes are related but not interchangeable.[3]

Use truthful state names in the interface and in exports. ‘Saved’ can mean only that a URL record exists. More precise labels such as linked, previewed, and captured help collaborators understand what may disappear and what has actually been preserved.