My worst practical problem with XML is that it is kryptonite to doing conflict resolution in version control systems. I've eased my pain with one such document

My worst practical problem with XML is that it is kryptonite to doing conflict resolution in version control systems. I've eased my pain with one such document by translating the document elements to a directory full of files, each of which has the name of its unique identifier attribute and the contents of which looks something like === attributes === a=b === contained elements === id_1234 id_4567 === associated text === This is the text associated with this element Those one-line tags which are syntactically distinct from attributes or identifiers serve as anchors for the diff program to "do the right thing". For this application, I can omit whitespace text elements, and there is at most one #text element per containing element, so I can conflate them. Were there more than one, I could generate id's for the contained text elements. Are there better ways of doing this? Tools?