Recent proposals for creating digital scholarly editions (DSEs) through the crowdsourcing of transcriptions and collaborative scholarship, for the establishment of national repositories of digital humanities data, and for the referencing, sharing, and storage of DSEs, have underlined the need for greater data interoperability. The TEI Guidelines have tried to establish standards for encoding transcriptions since 1988. However, because the choice of tags is guided by human interpretation, TEI-XML encoded files are in general not interoperable. One way to fix this problem may be to break down the current all-in-one approach to encoding so that DSEs can be specified instead by a bundle of separate resources that together offer greater interoperability: plain text versions, markup, annotations, and metadata. This would facilitate not only the development of more general software for handling DSEs, but also enable existing programs that already handle these kinds of data to function more efficiently.