Why web exhibits disappear
A printed catalog sits on a shelf and stays readable without help. A web exhibit does not. It depends on a server, a domain name, a content system, plugins, fonts, embedded videos and sometimes outside services. If any of those stops working, the exhibit can break, even if no one decided to remove it.
Common causes of loss are mundane. A hosting plan lapses. The person who maintained the site leaves. A software update changes how pages display. An embedded video service closes or changes its links. Archiving is the work of reducing these dependencies before they fail, ideally while the people who built the exhibit are still around.
Decide what “archived” will mean
There is a range of outcomes, and it is worth choosing one on purpose.
- Keep it live, unchanged. Cheapest in the short term, but it keeps all the maintenance burden and security concerns of a running site.
- Convert to a static copy. Pages are saved as plain HTML, images and styles, with no database or server-side code. This usually lasts the longest and costs the least to host.
- Keep a reduced version. A landing page, a summary, key images and a downloadable file, with the full interactive version retired.
- Keep only a record. Screenshots, a text version and documentation, held in the institution’s records.
For many small projects, a static copy plus a record of the content is a sensible middle path. Exhibits built around complex interactivity, such as a 3D scene or a map with live data, may need a reduced version instead, since the interactive parts are the hardest to preserve.
Make a complete inventory first
Before changing anything, list what the exhibit contains and where each piece lives. This typically includes:
- The text of every page, including captions and any hidden descriptive text for images.
- Original images at the best available quality, not just the web-sized versions.
- Audio, video, transcripts and captions.
- Any embedded content hosted elsewhere.
- The site’s structure: page titles, order and the links between pages.
- Rights information for each item, including any permissions that expire.
The rights point deserves attention. Permissions granted for a limited run may not cover indefinite display. Check what was agreed for each image, particularly where there were unclear ownership questions such as orphan works, and remove or replace items that cannot stay.
Prefer simple, widely supported formats
Files last longest when they are in formats many tools can read. Plain HTML and CSS, standard image formats, text in a plain or openly documented format, and common audio and video types are generally safer than proprietary or very new ones. Keep the highest quality masters you have; understanding image resolution and file formats helps when deciding which version is the true original and which is a derived copy.
Where possible, bring embedded outside content in-house. A video hosted on a third-party service can be saved as a file alongside the site, with its captions. If that is not allowed by the terms or rights, note the original link and keep a transcript and a few still frames so the content is at least understood.
Test the copy, not just the original
After producing a static copy, open it with no connection to the old system. Click through every page. Check that images, captions, downloads and internal links work, and that nothing quietly loads from the old server. Open it on more than one browser and on a phone. Try it with screen readers if you can, since structure that depended on the old content system sometimes breaks in conversion.
It is also worth asking someone who did not build the exhibit to look through it. They will notice missing context that the team no longer sees.
Write the documentation
Technical files are only part of an archive. A short plain-language note should record:
- The exhibit’s title, dates, purpose and intended audience.
- Who created it and who holds the rights.
- How it was built, in broad terms, and which parts were custom.
- What was changed during archiving, and what no longer works.
- Where all the files are stored and who is responsible for them.
This document is a close relative of the project handover notes written at launch, and it is often the most valuable item in the whole archive. In a few years, people will have forgotten why certain decisions were made.
Records of the objects shown also matter. Good collection metadata attached to the images and text means the material can be found and understood on its own, even if the web pages themselves are eventually lost.
Plan where it will live and who looks after it
Store more than one copy, in more than one place, and check them from time to time. A single copy on a staff member’s computer is not an archive. Institutional repositories, a records system or a trusted web host are better homes. Decide who will check the archive, how often, and what triggers a review, such as a change of domain or hosting.
A public address that stays the same is valuable. If the exhibit moves, set up redirects from the old address so existing links and search results still reach it. If the show is retired entirely, a short notice page explaining that and pointing to the archived copy is kinder to visitors than an error message.
Accept that some loss is normal
Not every feature will survive. Animations, search boxes and interactive maps may have to be simplified or replaced by recorded demonstrations. That is acceptable if the text, images and meaning remain accessible. A modest archive that is complete and readable is better than an ambitious one that stops working within a year or two.