Extended subject access to hypertext online documentation. Part III: The document‐boundaries problem

Because the DFT (DOCUMENT, FIND, THESEUS) online documentation system supports the hybrid goals of computer‐managed on‐demand printing of software manuals as well as the interactive retrieval of reference passages, its hypertext database is partitioned by stable document boundaries. These document boundaries solve some organizational problems to which newer hypertext systems are prone: they help disambiguate multiple hits and they promote unified editing and monitoring of passages that belong together intellectually. They make online cross‐references harder to use efficiently, however, and they create “local” contexts for keyword choices that introduce apparent inconsistencies from the perspective of global searching. Supporting document‐based along with document‐independent searches also runs a risk of invoking conflicting mental models in users. Finally, the virtual database structures of which hypertext systems boast are constrained by the size of the system's text nodes. With node sizes chosen large enough to preserve the logical structure of within‐document passages, however, this constraint actually improves a hypertext's explanatory value. © 1991 John Wiley & Sons, Inc.