An SGML document consists of text that is marked up with descriptive tags that specify the function of a given element within the document. As a formal language construct, an SGML document can be parsed against a document-type definition (DTD) that unambiguously defines what elements are allowed and where in the document they can (or must) occur. This formalized map of article structure allows the user interface design to be uncoupled from the underlying database system, an important step toward interoperability. Demonstration of this separability is a part of the CORE project, wherein user interface designs born of very different philosophies will access the same database. NOTES: (6) The CORE project is a collaboration among Cornell University's Mann Library, Bell Communications Research (Bellcore), the American Chemical Society (ACS), the Chemical Abstracts Service (CAS), and OCLC. Michael LESK The CORE Electronic Chemistry Library A major on-line file of chemical journal literature complete with graphics is being developed to test the usability of fully electronic access to documents, as a joint project of Cornell University, the American Chemical Society, the Chemical Abstracts Service, OCLC, and Bellcore (with additional support from Sun Microsystems, Springer-Verlag, DigitaI Equipment Corporation, Sony Corporation of America, and Apple Computers). Our file contains the American Chemical Society's on-line journals, supplemented with the graphics from the paper publication. The indexing of the articles from Chemical Abstracts Documents is available in both image and text format, and several different interfaces can be used. Our goals are (1) to assess the effectiveness and acceptability of electronic access to primary journals as compared with paper, and (2) to identify the most desirable functions of the user interface to an electronic system of journals, including in particular a comparison of page-image display with ASCII display interfaces. Early experiments with chemistry students on a variety of tasks suggest that searching tasks are completed much faster with any electronic system than with paper, but that for reading all versions of the articles are roughly equivalent. Pamela ANDRE and Judith ZIDAR Text conversion is far more expensive and time-consuming than image capture alone. NAL's experience with optical character recognition (OCR) will be related and compared with the experience of having text rekeyed. What factors affect OCR accuracy? How accurate does full text have to be in order to be useful? How do different users react to imperfect text? These are questions that will be explored. For many, a service bureau may be a better solution than performing the work inhouse; this will also be discussed. SESSION VI Marybeth PETERS