Of course, Curiosity is very good at extracting information represented by means of Microformats.
In short, Microformats are a set of standards targeted at assigning semantic information to (x)html elements by means of css classes.
For example, the hCard microformat is a 1:1 representation of the vCard standard for encoding information about people: a vcard will be identified by an html element whose class is vcard, and the fn field will be identified by an enclosed element whose class is fn. So, in Curiosity, the xpath for extracting fn data could be like this:
//*[contains(@class, 'vcard')]//*[contains(@class, 'fn')]
The page showing a list of contacts used in a previous example, has been written using hCard (you may have noticed the little hCard button at the bottom of the page): so, you may want to define slots xpaths accordingly to hCard. This way, you will have robust slots which won't be strictly dependant upon the page internal structure, and moreover you will have re-usable xpaths.
The sample curiosity.xml shipped with the default Curiosity distribution, has a source named HCARD aimed at extracting hCard data: by opening its Source Details form, you can add a Cloned Web Source with the url of the about page of this site - which is encoded too using hCard.
In the Curiosity distribution it is also defined a provider for converting hCard data in vCard format. Nothing prevents of writing every sort of xsl - for example for converting hCard to ldif format.
Curiosity is shipped also with a source defined for hCalendar - a microfomat for representing iCalendar data.