Glossary
Retained source copy
A retained source copy is a preserved version of material used as evidence for a claim or decision.
Provenance and citation
A retained source copy preserves the material as it was available when a claim was assessed. It may be a document, page capture, file or other record, accompanied by acquisition time and source location. The copy matters because a live page can change or disappear. Preservation should respect access rights and retention obligations; keeping everything forever is neither necessary nor always permitted.
In practice, the retained copy is connected to the exact evidence span and the claim it supported. A checksum can help detect byte changes, while metadata records where and when the material was obtained. If a publisher later corrects the source, the earlier copy still explains the original decision and the correction can be assessed as new evidence.
For a defensible record, the ability to inspect the source as seen at decision time is stronger than relying on a current URL. Content addressing identifies a version by its bytes, while chain of custody documents handling and integrity over time. A retained copy supports review but does not settle whether the source statement was accurate.
Example
A Halden Therapeutics AG (fictional) decision cites a dated notice copy; when the live notice changes, both versions remain distinguishable.
What to keep and in what form
A retained source copy is the bytes of a source as they were when read, plus the metadata needed to interpret them: the address, the retrieval time, the response headers, the method of retrieval and a digest. For web pages, the standard container is the WARC format, described by the Library of Congress, which stores the request, the response and the capture metadata together and is used by web archives. Rendered text or a screenshot may also be kept, but a text extract is a derivative and should not be mistaken for the source: the same page can extract differently with different tools. For PDFs and other documents, keep the original file and its digest.
Public archives such as the Internet Archive's Wayback Machine can supplement a private copy, and the Memento protocol in RFC 7089 standardises how to ask for the version of a web resource as it was at a given time. They should not be the only copy, since an archive may not hold a page, may hold it in a different state, or may remove it. The copy that supports a decision is best kept by the organisation that made the decision.
Why pages change under a citation
Web content is unstable. The Pew Research Center found, in a 2024 study of pages and links, that a substantial share of pages that existed in 2013 were no longer accessible a decade later, and that links on current pages also commonly led nowhere. Even pages that remain reachable are edited: a registry corrects an entry, a company rewrites a statement, a regulator withdraws a letter. A citation of a live URL is a pointer to whatever the page says today, which may differ from what supported the claim. The article on preserving evidence when a cited source changes walks through the procedure for capture and for handling a later change.
A policy decides what to retain: every source behind a decision, or only those above a materiality threshold; for how long, as set by the retention schedule; and in what storage, with access controls appropriate to the content. Copyright and terms of use may restrict redistribution but generally do not prevent retaining a copy for internal evidence; where a licence is relevant, the question should be settled with counsel before capture and not after.
A worked example
An analyst cites a regulator's notice about Halden Therapeutics AG (fictional) in a decision memo. The capture step stores the page as a WARC record, a PDF print and a text extract, records the URL, the time and the SHA-256 digest of the response body, and writes the digest into the memo. Six months later the regulator updates the notice and removes one paragraph. A reviewer reading the memo opens the retained copy by its digest, finds the paragraph, and sees that the decision was reasonable on the information then available. The live page now shows something different, and the reviewer can see both and the date of the change. Without a copy, the memo's basis could not have been shown.
Related terms and boundaries
The copy is what an evidence span points into and what content addressing names by its digest. It answers the problem of link rot, supports a chain of custody, is governed by a retention schedule, and is the base for point-in-time reproducibility. It records what a source said, not whether it was right.
Common misreadings
- Keeping only a mutable URL.
- Preserving content without acquisition metadata.
- Ignoring lawful retention limits.
Appears in
- Document citation vs sentence citation
A document citation identifies an artifact; a sentence citation locates the specific statement supporting a claim.
- Citing at the sentence
Bind each assertion to the exact source passage, its locator, reading date and retained copy.
- Detecting source drift
Compare current source content with the retained cited version and reassess changed support.
- Retaining source copies
Preserve the material actually read, subject to rights, and bind its citation with a content digest.
Found an error on this page? See corrections for what counts as one, how to report it and the dated log.
Enquiry
Where does "Retained source copy" bear on a decision of yours?
This entry sits under provenance and citation. If a decision in your organization turns on it, send the entity concerned, the sources you fuse to describe it and who may read which of them. We reply with what a first record would cover.
Or write to [email protected] in your own words.