Rules are collected for each structured element in a document. The rules that apply to any particular instance of an element depend on the hierarchical position of the instance within the document. The rules are collected and efficiently organized in a manner that accounts for this dependency. Once the rules are collected, the rules may be further processed. For example, there may be dependencies between calculations. The computation of the calculations is scheduled in a manner that accounts for these dependencies. Once the rules are collected and further processed, the rules are used to validate the given document.