Proceedings Citation Guide

Link Rot: Archiving the Web Pages and PDFs You Cite

Web addresses stop working over time, and a broken link can make a citation useless. This guide explains why it happens and how to keep a reference usable.

Last updated 6 November 2026

On this page
  1. Why links break
  2. Prefer stable identifiers over plain URLs
  3. Archiving a page with the Wayback Machine
  4. Archiving PDFs and data you cite
  5. Citing an archived copy and using accessed dates
  6. A short workflow for writers
  7. Common mistakes
  8. Frequently asked questions
  9. Sources and further reading
  10. Related topics

In brief

  • Link rot means a URL no longer leads to the content you cited.
  • DOIs and other persistent identifiers resist rot better than ordinary URLs, but only when their owners keep them updated.
  • The Wayback Machine's Save Page Now saves one page per submission and does not save outlinks.
  • Archive a copy when you cite a web page, then keep both the original and the archived address.

Why links break

Link rot is the gradual loss of working links. A page can vanish, move or change so much that it no longer supports your claim. That last case is sometimes called content drift.

Common causes include:

  • A site redesign changes its address structure.
  • A project ends and its server is switched off.
  • A file is replaced by a newer version at the same address.
  • A domain lapses and is reused by someone else.
  • A page moves behind a login.

Workshop pages are especially exposed. Event websites are often built for one year and then left alone. Cite the published proceedings rather than the event site whenever you can. Persistent identifiers help too; see DOI, ISSN, ISBN, URN and ORCID.

Readers meet link rot in two ways. A dead link returns an error or an unrelated page. A drifted link still opens, but the content has changed since you cited it. The second kind is harder to notice, so record what the page said on the day you read it, for example by saving a copy or noting the date and the key passage.

Plan for rot early. It is much easier to archive a page the day you cite it than to hunt for it years later.

Prefer stable identifiers over plain URLs

A DOI resolves through https://doi.org/ to the current location of the object. The DOI Foundation says persistence is a function of organizations, not of technology, so the system only works while the registrant keeps the link up to date. Even so, a DOI is usually safer than a copied web address.

Some hosting services also state their own stable forms. CEUR-WS.org gives each volume a number and a URN, and the ACL Anthology advises using the DOI on a paper's page. Use the form the service itself recommends. If none exists, use the cleanest URL you can find, without tracking parameters or search-engine redirects.

Archiving a page with the Wayback Machine

The Internet Archive's Wayback Machine offers a tool called Save Page Now. You enter a URL on the Wayback Machine page, submit it, and receive a permanent address for the saved copy at once. The copy includes images and style sheets.

Know its limits, which the Internet Archive lists on its help page:

  • It saves one page per submission and does not save outlinks.
  • It cannot crawl a whole site.
  • Some sites prohibit crawling, so their pages may not save.
  • Certain security settings can make saving fail.

After saving, open the new address and confirm the content is what you cited. Check that the figures, tables and key sentences are present. A blank or error capture is useless as evidence.

Archiving PDFs and data you cite

PDFs on personal or project sites are fragile. If you may reuse a PDF, you can deposit your own work, or a copy you have the right to share, in a repository. Zenodo, run by CERN with OpenAIRE as a partner, registers a DOI through DataCite when you publish a record. It also states that a published record's identifier cannot be changed, and that a new version is a separate record with its own identifier, so a cited version stays the same.

Do not upload other people's papers unless their licence allows it. For your own accepted manuscripts, check the publisher's self-archiving terms first; see Copyright and Licences for Proceedings Papers. Your institution's repository may be the simplest compliant option.

Other archiving services exist, including some run by libraries and universities. We did not verify their terms for this page, so read each service's own rules before you rely on it.

Citing an archived copy and using accessed dates

Styles differ, so follow the one your venue requires. The general approach is the same. Give the original details first, then the archived address. Many styles also use an accessed or retrieved date, because web content can change. Harvard-type styles often require one, and APA and others use it mainly when content is expected to change.

A fictional example, in a plain author-date form:

Doe, J. (2023). An example title of a fictional web page. Example Project. https://example.org/page. Archived at https://web.archive.org/web/2023/https://example.org/page (accessed 1 March 2024)

The addresses above are placeholders. When you use a real archived address, copy it exactly from the archive page. For style details, see How to Cite Workshop Proceedings in APA 7th Edition.

A short workflow for writers

  1. Look for a DOI or other stable identifier first.
  2. If you must cite a web page, save it with Save Page Now and note the date.
  3. Check the saved copy opens and contains your evidence.
  4. Record the original URL, the archived URL and the access date in your reference manager.
  5. Test every link again just before you submit. Your reader will click them, so check them yourself.

Reference managers can store a link and a note, but they do not archive pages by themselves. Keep your own list as a backup.

Common mistakes

  • Citing a search-result address that expires.
  • Citing the event site when a published volume exists.
  • Saving a page without opening the copy.
  • Archiving a login page or a cookie banner instead of the content.
  • Dropping the original URL once an archive link exists.

To check a reference, see How to Verify That a Citation Is Real and Correct. For more frequent errors, read Ten Common Mistakes When Citing Proceedings.

Frequently asked questions

What is link rot?

It is the loss of working web links over time, when pages disappear, move or change.

Is a DOI immune to link rot?

No. A DOI works as long as its registrant keeps the target address up to date, but it is usually more reliable than a plain URL.

Does Save Page Now save a whole site?

No. The Internet Archive says it saves one page per submission and does not save outlinks.

Should I include an accessed date?

Follow your style. Many styles ask for one for web content that may change, and Harvard-type styles often require it.

Can I archive a paper I do not own?

Only if its licence or the publisher's policy allows it. Otherwise cite it and link to the original host.

Sources and further reading

  1. Internet Archive — Save pages in the Wayback Machine (2026)
  2. DOI Foundation — What is a DOI? (2026)
  3. CERN / Zenodo — About Zenodo (2026)
  4. Zenodo — Zenodo records and versioning (2026)
  5. CEUR-WS.org — CEUR Workshop Proceedings (2026)
  6. ACL Anthology — ACL Anthology FAQ (2026)