Identifying logical elements by modifying a source document using marker attribute values

Methods and apparatus, including computer program products, implementing and using techniques for processing an electronic document. A first electronic document including a plurality of logical elements is obtained. Each logical element has a logical type and associated content with a visual appearance. A second electronic document is generated by associating a marker attribute value with a plurality of logical elements in the first electronic document and converting the first electronic document with the associated marker attribute values through a document conversion process. The marker attribute values are a value of a marker attribute and the conversion process preserves the association of the marker attribute values and the logical elements. The marker attribute values are then used to identify logical elements in the second electronic document.