Entity clarity: keep the journal, publisher and authors distinct

1. Give each entity its own role

Article A is the work; Journal J is the serial in which it appears; Publisher P is the publishing organisation; and the authors are people responsible for the work. An author affiliation is another relationship. One name field cannot safely substitute for all of them.

Schema.org supplies vocabulary for authors, publishers and article relationships to a journal or issue. That vocabulary can express a distinction already established by the publication records. It cannot decide which journal or person is correct when the underlying evidence conflicts.

Reference: Schema.org: ScholarlyArticle

2. Reconcile the visible page, markup and deposit

In EC64, the public page names Journal J and Publisher P correctly, but the structured data places P in the journal-name field. A deposit draft uses an unexplained abbreviation of J. The author field includes an identifier copied from a name search. The coordinator creates a field-level record before editing any surface.

On small screens, scroll the table horizontally, or focus it and use the arrow keys.

3. Identify the serial without turning its identifier into a quality badge

The ISSN International Centre describes an ISSN as an identifier for a serial publication and its title; it does not guarantee the quality or validity of the content. Use the authoritative record to check the journal identity and relevant medium or title history, rather than treating a number as endorsement.

For EC64, the journal editor supplies the approved identity record and explains the abbreviation in the deposit draft. A matching short title on a search page is only a lead. No ISSN is invented or assigned to Journal J in this exercise.

Reference: ISSN International Centre: what an ISSN identifies

4. Do not join people by a name or email match

A similar name, email address or affiliation does not establish that two records describe the same person. Ask for the applicable authenticated identifier or verification process and retain any unresolved conflict. A site account ID is also not interchangeable with a scholarly person identifier.

ORCID distinguishes the source that added an assertion from the origin of the connection being asserted. Inspect that provenance for the particular item. An identifier is not a blanket certificate for every affiliation or work claim, and a sparse public record is not by itself a reason to invent missing information.

Reference: ORCID: source and assertion origin

5. Preserve bilingual names and organisation relationships

The current ROR fields represent names with language and name-type information, and express relationships between organisations separately. A translated label or acronym is not the same thing as a parent/child relationship. Use the organisation's identity evidence rather than assuming two similar strings must be one institution.

In EC64, the Arabic affiliation wording remains a separate unresolved item until the responsible source explains the relationship. A copyeditor may improve readability without changing the claimed institution, but must not silently replace it with a better-known organisation. Keep the original value and the reason for any agreed correction.

Reference: ROR: current fields and subfields

6. Check each corrected surface and retain its observation time

Prepare changes only for fields whose evidence is resolved. Compare the updated public page, its structured data and the deposit record separately. A prepared deposit correction is not a successful agency update, and an updated page does not establish that a search service has refreshed its result.

Google's guidance says correct structured data does not guarantee a search feature. EC64 therefore ends with an editorial correction record, not a promise of AI recognition, citations or ranking. Record what was checked, by whom and when; keep the author identifier and affiliation unresolved until their own evidence arrives.

Reference: Google: general structured-data guidelines

Questions for a practical final check

• Are the article, serial, publisher, person and affiliation distinct?
• Does every changed field have an evidence source and owner?
• Are name, email and account matches treated only as leads?
• Are bilingual variants and organisation relationships supported?
• Are source correction and downstream refresh independently verified?