Unfortunately, there is no 100% foolproof process for determining how to validate an arbitrary XML document. If you are receiving a document, you should not leave choosing the validation mechanism to a remote party (e.g. downloading a DTD using its document-specified URI). Doing so opens your application to, at the very least, a potential denial-of-service attack. A validation mechanism may not even be specified in the document: W3 XML Schema (XSD) does not require it; RELAX NG does not seem to support such a mechanism. Then there are some XML documents that just don't have a schema of any form.
Nevertheless, there are times when you need to inspect a document to find out what it is. Most commonly, support is required for multiple versions of a document, where the structure and validation mechanisms change over time.
Note: when talking about validation, this post is not referring to whether the XML is well formed or not. Any XML parser should be able to check the syntax. This is about external constraints imposed on the document structure via a schema, DTD, etc.
Posts
