We argue that efficient support for schema validation and type annotation in XQuery processors deserves as much attention as efficient evaluation techniques for XPath queries have received in the past. To this end, we describe a validation procedure that operates on an encoding of trees that has already been succesfully used for XPath location step evaluation. The validation algorithm works without the construction of finite state automata and can exploit properties of the encoding to speed up the validation of already partially type-annotated trees. First experiments—carried out using the database back-end of the authors’ XQuery compiler—indicate that this approach can indeed lead to high-troughput validation support.