Huge collections of linked documents can now be efficiently stored. However, full online access and electronic publishing through reuse of document parts require sophistication and precision in queries. Such a query facility is only possible through the inclusion of appropriate semantic information. Manually adding such information to multigigabyte document sources is daunting for technical writers. Our approach aims at making this task feasible by exploiting a conceptual schema of the enterprise. The result is an integrated schema — one that covers the traditional information system of the enterprise as well as the information that exists solely in the world of documents.