*************************************************************** ****************** WELCOME TO SGML NEWSWIRE ******************* *************************************************************** * * * To subscribe, send mail to sgmlinfo@avalanche.com * * * * (Please pass along to interested colleagues) * * * *************************************************************** IEEE EXPERT: KNOWLEDGE-BASED DOCUMENT FILING ============================================ For all you 'tekkies' out there: You may be interested in reading about the hard details of a report on artificial intelligence in text-based information retrieval. "Information retrieval, formerly bound to some equivalence between a text's wording and meaning, is moving toward the use of logical inference, knowledge bases, and rule resolution. These techniques augment a system's ability to identify the semantics of a document beyond its textual content. Our approach to document modeling and classification is based on representing knowledge about documents' roles... [The Kabiria project] is a knowledge-based filing and retrieval system designed around the notion of documents as objects embedded in a rich procedural and domain context. The context describes a document's semantics by taking into account the activities in which it is used and the domain rules that justify its existence in the office environment. Kabiria provides a homogeneous environment for classifying documents according to standard types, for filing documents according to organization and domain-dependent relationships, and for retrieving documents according to the knowledge embedded in the classification scheme. Retrieval involves both querying and browsing activities, which also support exploration of the structure of the document base within the specific office and domain context. The sidebar on pages 36-37 frames out work within the context of related projects dealing with document modeling and intelligent retrieval." The sidebar referred to in the last paragraph above focuses on Office Document Architecture (ODA), but draws comparisons with SGML: "International committees are making great efforts to draw up standards that will enable different systems to share a common understanding of documents' structures. The result is a multipart standard commonly referred to as the [ODA], which distinguishes between a document's contents and its structural characteristics, that is, between its logical and layout structures.... Some relationships can be drawn between ODA and SGML... The biggest difference between the two is that SGML makes no direct provisions for describing the layout of documents or for defining any form of content other than text. Layout characteristics can be described by attributes given in the markup tags, but these are not in any way part of the standard. Similarly, SGML provides a mechanism for including different types of content, but they are also not part of the standard." Excerpts taken from: IEEE Expert, "Knowledge-Based Document Filing," Silvano Pozzi and Augusto Celentano, October 1993. ************************************************************** * SGML NEWSWIRE LIST MANAGER * * * * Linda Turner * * Corporate Communications * * Avalanche * * 947 Walnut Street * * Boulder, CO 80302 * * sgmlinfo@avalanche.com * * linda@avalanche.com * * Vox: (303) 449-5032 * * Fax: (303) 449-3246 * **************************************************************