Glossary
RDFa
RDFa (Resource Description Framework in Attributes) is a W3C spec for embedding metadata in HTML attributes. Supported by schema.org and Google but rarely used in modern web SEO.
RDFa is the third schema.org format alongside JSON-LD and Microdata. It uses HTML attributes like `vocab`, `typeof`, `property`, and `resource`. Example: `<div vocab='https://schema.org/' typeof='Person'><span property='name'>Jane</span></div>`.
RDFa is more powerful than Microdata in some edge cases (cross-document linking, multiple vocabularies on one page) but those edge cases rarely apply to typical SEO use. For 95% of structured-data needs, the three formats are functionally equivalent.
Google parses all three formats. The choice between them is operational, not SEO: JSON-LD wins on maintainability and framework support; Microdata and RDFa both require attribute-level markup. Most new projects choose JSON-LD.
When RDFa appears: legacy sites (especially academic / government / library sites) often use RDFa because it predates JSON-LD adoption. Some semantic-web research projects still prefer RDFa because it can express more complex relationships than schema.org alone supports.
If you're starting a new SEO project in 2026, you should use JSON-LD. RDFa is a 'know what it is when you see it on someone else's site' format rather than something to actively choose.
Example
A library catalog uses RDFa to mark up books, authors, and publishers using a custom vocabulary alongside schema.org. The RDFa lets external research tools query the data; the schema.org parts let Google show book details in SERP. JSON-LD couldn't have expressed the custom vocabulary as cleanly.