The hCard spec says: <http://microformats.org/wiki/hcard#Property_Notes> The 'n' property is OPTIONAL if any implied 'n' optimization rules are in effect. suggesting that it is mandatory otherwise, while the cheatsheet: <http://microformats.org/wiki/hcard-cheatsheet> lists it as: OPTIONAL, but MUST NOT occur more than once The later also says of 'adr': # At least one child-property MUST be present in adr. # If an adr-child (e.g street-address) is present it will ONLY be considered part of the adr data IF it is inside an adr. but makes no such stipulations for 'n' and its children. I certainly never use "n" for places: <foo class="fn locality">Birmingham</foo> and parsers (Operator, Tails, Tails Export, X2V) seem to have no problem with them - other than trying to optimise place names into given-, family- and nick- names; an issue I've raised previously: <http://microformats.org/discuss/mail/microformats-discuss/2007-December/011169.html> (aka: <http://tinyurl.com/223qzo>) <http://microformats.org/discuss/mail/microformats-discuss/2008-January/011182.html> (aka <http://tinyurl.com/2c2yvr>) and which is awaiting acceptance or otherwise of my proposed solution. There are also difficulties using n-children where the make-up of the 'fn' is not known in advance; consider a database of low-granularity names with properties like: John Smith-Brown Mary Anne Jones Ai Ki Chi so sites like Wikipedia use: <foo class="fn">[name of indeterminate structure]</foo> which again causes no problems to parsers. Should the spec be reworded? -- Andy Mabbett