Testing XML Documents for Information Content

More and more applications are generating XML documents as their primary or secondary outputs because XML is much easier to parse than traditional formats. However, simple string comparison is often too naïve to properly test XML documents. Instead, they must be evaluated for information content. XML has many syntactic options that make testing output more difficult than testing traditional, less rich formats. Elliotte Rusty Harold explores the challenges and pitfalls of testing XML documents and, even more importantly, offers suggestions of what to check and what to ignore. He presents valuable tools for testing XML-parsers, schemas, DTDs, canonical XML, and XPath--and discusses automating tests by writing xUnit test cases that use XML APIs such as DOM to compare the actual to the expected output.