Writing the Web while disconnected

The World Wide Web has exploded as a medium for sharing information in a relatively simple, easy-to-use, and appealing manner, Technology for accessing Web resources has naturally evolved to support disconnected viewing of cached pages normally residing on servers around the world, and much attention has been given to providing this capability. The Web was originally conceived, however, as a medium for both information sharing and collaborative work, for creating and updating information, as well as for reading. Very little work has considered how to support the update and creation of material while disconnected, for propagation back to the appropriate servers upon reconnection. This article discusses issues arising in such support, compares the problem to that in disconnected file systems, and describes an approach for disconnected updates, as demonstrated by our Caubweb' " technology for disconnected interaction with Web-based resources. he World Wide Web has exploded as a medium for sharing information in a relatively simple, easy-to-use, and appealing manner. A staggeringly diverse collection of information is " published " on an ever-growing number of Web servers around the world, and the relative ease of use for the typical access client (a Web brows-er) has resulted in a very broad deployment of the technology, in terms of both the user population and the context of use. For example, businesses now regularly Web-publish corporate documents for internal use, while also making available large amounts of information externally to both their partners and their customers. Universities and libraries disseminate knowledge via the Web. Business transactions increasingly occur on the Web. Grandparents view baby pictures via the Web. Elementary school classes share science projects with peers on other continents. Workgroups create and process shared information spaces via the Web. The latter example is particularly interesting to us. While most of the current use of the Web is strictly for viewing information created by others, without changing or adding to it, the Web was also initially envisioned as a medium for col-laborative creation and sharing of information [l]. In this sense, the Web has missed some of its anticipated potential. The current client model is heavily oriented toward passive interaction with the information, and most clients and servers d o not take advantage of the aspects of the information exchange protocol (Hypertext Transfer Protocol, HTTP [ 2 ]) that could support more active interaction. Another exciting trend in information access is the use of …