Google, Bing and Yahoo! have recently joined forces to work with schema.org to agree on a convention designed to improve understanding of the structure of different types of data (e.g. books, films, recipes) on websites so they can be shown in search engine result pages.
There are a large set of new data types described, including:
- Embedded non-text objects: AudioObject, ImageObject, VideoObject
- Place, LocalBusiness, Restaurant …
- Product, Offer, AggregateOffer
- Review, AggregateRating
There are currently 3 standards in use (referred to by Google as rich snippets): Microdata, RDFa and Microformats. Microdata has been chosen for use at the core of schema.org, but it is supplemented by additional ‘vocabulary’. The other formats will still be supported by the major search engines, but I suspect are unlikely to be developed much further and may eventually be phased out.
It’s worth being aware that in terms of HTML validation:
- Microdata is a HTML5 specification and will not validate as HTML4.
- RDFa will only validate within XHTML documents.
- Only Microformats will validate as standard HTML4.
Google provide a rich snippets testing tool to check the accuracy of data (note this tests for Microdata, RDFa and Microformats)
Schema.org provide a handy guide to getting started, that points you in the right direction.