While Ian was correct in stating that you can create vocabularies by yourself without going through the Microformats process, you cannot call it a “standard” – and that’s really what we’re talking about with both Microformats and RDFa – creating standards so that communities can inter-operate.
The problem that the Microformats community has right now is that they treat both syntax and semantics as an atomic part of each Microformat when they should really be separated. Each Microformat can have a different parser model as well as vocabulary, which means that a parser that is specific to each Microformat has to be created. There is no standard parser model for Microformats, there is no test suite to program against. One of the core computer language issues that is usually tackled first, how to express the language in a uniform manner, has not and will not be addressed for Microformats. This is by design, and it is not without consequences.
We spend a ridiculous amount of time in the Microformats community attempting to shoe-horn a semantic expression language into HTML4 and XHTML only to find that organizations reject our efforts due to accessibility issues.
RDFa solves the problem of a semantic expression language for XHTML1.1, XHTML2, HTML4 and HTML5. After all, it is merely a collection of attributes that you can sprinkle throughout your documents to express semantics. The method works in any language that has element properties. The great thing about RDFa is that you can use the same expression mechanism for ANY vocabulary. It solves one of the largest issues the Microformats community has – a uniform expression language for semantic data.