Namespaces

An unnecessary solution for many problems Namespaces provide unique identifiers for element and attribute names, making it easier to avoid conflicts between vocabularies. At present though, and into the foreseeable future, a substantial number of applications don't need element-by-element or attribute-by-attribute precision for naming - they settle happily for a rough idea of what the document type is.
Will the controversies ever end? Namespaces have been plagued with controversy since the beginning - attribute declarations vs. processing instructions, scoping, the legitimate and illegitimate use of URIs as identifiers, odd philosophical questions about whether namespaces are 'languages' or even 'things'. The 1300+ combative messages on the xml-uri@w3.org mailing list for May and June are just a small part of the story.
Missing at the beginning, incompletely implemented since Because namespaces arrived after XML 1.0, and because validating parsers do not readily accept changes to namespaces prefixes, namespace support is still considered separate from XML procesing. The W3C clearly wants developers to start using namespaces, but not everyone is listening. Microsoft's implementations, with their odd treatment of the 'html' prefix and the requirement that namespace-affecting attribute list declarations use #FIXED, haven't helped either.