Reference formats
BibTeX Metadata Problems That Break Citation Workflows
Understand why valid BibTeX syntax can still produce incorrect references, missing identifiers, duplicate records, and inconsistent rendered citations.
Syntax validity is only the first check
A BibTeX entry can parse successfully while its title, author, year, venue, pages, or DOI describes a different work. Braces and entry keys control how a record is written; they do not establish scholarly identity.
This distinction matters when a bibliography is assembled from multiple exports, hand-edited records, or AI-generated suggestions.
The fields most likely to cause trouble
Review title capitalization, author separators, journal versus booktitle, year values, page ranges, and DOI normalization. Conference papers, preprints, translations, and online-first articles often use different fields for related versions.
Custom macros and non-standard fields may render correctly in one tool and disappear in another. Keep warnings visible and inspect the normalized model before generating a final bibliography.
- Use a stable entry key, but do not treat it as a scholarly identifier.
- Normalize DOI values and remove resolver prefixes only when your workflow expects it.
- Check journal and booktitle fields for the intended publication type.
- Compare year and pages against a provider record.
- Detect duplicate or near-duplicate entries after normalization.
A better BibTeX QA pipeline
Parse the file into a common citation model, report syntax and field warnings, detect duplicates, and only then verify the underlying work against scholarly providers. Keep parser output separate from provider evidence and any AI suggestion.
This order prevents a clean-looking export from hiding an identity problem and makes it clear which fixes are deterministic and which require human review.
把這套工作流程用起來。
在 CiteWise 工作區中使用有來源支持的證據驗證參考文獻。