Digital Preservation of Academic Web Publishing
Academic publishing has a long memory and the web has a short one. A journal issue, a working paper series or a project archive is cited for decades, while the software that presents it is typically replaced every five to seven years. The gap between those two timescales is where scholarly material disappears — not through deletion, but through a platform change that nobody planned an exit from.
Citation is a technical requirement
When a researcher cites a URL, that address becomes part of the scholarly record. It is no longer the publisher's private routing decision. Any migration of an academic site therefore has an obligation that a commercial redesign does not: every previously cited address must continue to resolve, either directly or through a permanent redirect.
In practice this means the identifier scheme should be designed to survive the CMS. Paths that encode the platform — a query string with a numeric post identifier, a directory named after a plugin — will not survive. Paths that encode meaning, such as a volume, issue and article slug, will.
Where persistent identifier services are available they should be used, but they are not a substitute for stable URLs; a resolver that points at a dead address preserves the reference and loses the resource.
Preserve the document, not just the page
Academic sites carry a large tail of attachments: PDFs, datasets, supplementary material, conference programmes and slide decks. These are frequently the most-cited items and the most likely to be lost, because they do not appear in a page-by-page visual review of the new site.
Enumerate them from a crawl and from the file system, preserve their paths, and check their content type headers after migration. A PDF served as plain text is technically present and practically broken.
Metadata belongs with the document: author, date, licence, language and version. Losing that turns an archive into a folder of files.
Freezing is a legitimate outcome
Not every academic property should be migrated into an active CMS. A completed project site, a past conference or a closed journal series has no future editorial activity. For these, the correct outcome is to convert them to static files, keep the URLs identical, and host them cheaply and indefinitely.
Freezing has real advantages: no software to patch, no dependency updates, no security surface, and a hosting cost near zero. It also removes the recurring temptation to redesign material that should simply remain as published.
The decision rule is straightforward. If the content will change, migrate it. If it will not, freeze it. If it duplicates something active, consolidate and redirect. If it has no value at all, archive a copy and retire it deliberately rather than by deleting a domain.
Institutional memory and interdisciplinary studies
Sites associated with interdisciplinary European research — reviews, working paper series, project outputs — often accumulate across several funding cycles and several suppliers. The result is an estate with inconsistent structures and no single owner, which is why so much of it becomes unreachable when hosting contracts change.
The remedy is unexciting and effective: one inventory, one identifier policy, one standard platform, and a named person responsible for annual link integrity checks. Everything else in preservation follows from those four.
Technical references
The documentation below covers the procedures behind this article in far more operational detail than an editorial piece can.
- static versus dynamic hosting analysis
- static site conversion tooling
- open data conversion case study
- website portability checklist
- modelling structured content
- multilingual publishing setup
- backup and recovery practice
- long-term hosting considerations
- keeping references resolvable
- migrating content into a CMS
- structured metadata for documents
- conversion documentation library