In this paper I present the reasoning behind the development of a new end-to-end publishing system for academic writers. The story starts with investigating digital preservation of word processing documents. What file formats are suitable for long-term preservation of text? I believe that the answer is a high-quality structured XML format like DocBook XML or TEI. The next question is how do we get word processing documents into that format without incurring a prohibitive cost? Conversion is possible, but it requires human intervention at some point. It would be far too expensive to have archivists editing every document by hand on ingest, so how can we get authors to do the necessary work, particularly as most academics aren't at all interested in digital preservation of their work? The answer I propose is to offer them more than just an archiving solution. Instead of just getting preservation, they get a full end-to-end digital publishing solution, the digital scholar's workbench, tailored to their needs for document interoperability, collaboration and publication in multiple formats… oh, and they get preservation too.