Integrating Correction into Incremental Validation

Many data on the Web are XML documents. An XML document is an unranked labelled tree. A schema for XML documents (for instance a DTD) is the specification of their internal structure: a schema is a tree grammar, and validating a document w.r.t. a schema is done by a running of a tree automaton. Given a document, valid w.r.t. a DTD, and a sequence of updates (insertions, deletions and replacements of subtrees), we first recall how we incrementally check the validity of the resulting document. Indeed, updating a valid document requires re-checking the parts of the document concerned by the updates. Next, the core of the paper is a method to correct subtrees for which the re-validation fails: if the validator fails at node p, a correction routine is called in order to compute corrections of the subtree rooted at p, within a given threshold. Then re-validation continues. When the tree traversal is completed (i.e. all updates have been considered), the corrections generated by each call to the routine are merged, and different correction versions for the resulting document are proposed to the user. The correction routine uses tree edit distance matrices.