GNOSIS(ID:8094/)for General Natural-Interaction Oriented Operating System for Information Services Hybrid of Prolog and MUMPS in Winfried Boeing, Treuhanseminar, University of Cologne, D-5000 Koeln, West Germany Related languages
References: External link: Online copy bib: @article{62161, author = {Winfried Boeing}, title = {Intelligent information retrieval with proto-Gnosis}, journal = {SIGBIO Newsl.}, volume = {9}, number = {4}, year = {1987}, issn = {0163-5697}, pages = {19--21}, doi = {http://doi.acm.org/10.1145/62158.62161}, publisher = {ACM Press}, } Extract: Overview About the GNOSIS project GNOSIS is an abbreviation for General Natural interaction Oriented Operating System for Information Services An initial prototype of GNOSIS, called proto-GNOSIS, consists of a language interpreter for MUMPS (ANSI/MDC x.11.1 - 1984) and language extensions taken from PROLOG. Proto-GNOSIS has been developed by the MUMPS System Laboratory in Nagoya, Japan, and it is part of a GNOSIS project. The objective of the project is the specification of an applications profile for a 5th generation language, preferably based on international standards like those of ISO and ANSI. The language objectives should be similar to the objectives of MUMPS; the language should be portable and transferable and independent of the type of computers or the methods of implementation The concept of the GNOSIS project follows the objectives of the X/OPEN Group. The formation of the X/OPEN Group represents a major initiative by an international group of suppliers to create a free and open market for software by establishment of a Common Applications Environment. Therefore, a portability guide with the chapters SYSTEM V, C-LANGUAGE, OTHER PROGRAMMING LANGUAGES (COBOL, Fortran), DATA MANAGEMENT (ISAM) and SOURCE CODE TRANSFER BETWEEN MACHINES has been published by the X/OPEN Group. Beyond this, the GNOSIS project also intends to provide portability, with a special interest in interfacing industrial standards, like the UNDX system call, CODASYL's Screen Management System (SMS), the Graphics Kernel System (GKS), the Manufacturing Automation Protocol (MAP), and the Technical Office Protocols (TOP). The University of Cologne has been using proto-GNOSIS for some months now. In 1983, the Treuhandseminar at the University developed a retrieval system for literature using MUMPS. The system is used by students and academic staff. A detailed overview of its capabilities was given in 1985 At this stage, we intend to use GNOSIS for a more sophisticated approach to literature search. This should provide benefits to the students, such as: ? shorter connect time with the host computer ? increased hit rates of the source documents ? more qualified and relevant choice of literature A common portable environment could also provide the students with easier access to the system In the following article, we will clarify what intelligent information retrieval (IIR) means. Then we will provide an in depth description of the use of IIR systems for access to public databases Finally we will point out how GNOSIS works. Extract: Intelligent Information Retrieval Intelligent Information Retrieval Public databases. One main area of open systems interconnection is the access to databases. American companies are in a leading position in the world in this area; for example. DIALOG has more than 400 databases In Germany and in Europe, online connections to hosts are not in as high demand as in the US. A German program called JOUR-FIZ II screened the online information and documentation retrieval for journals and newspapers The project was supported from 1982 to 1984 by the German Government. The results revealed that an extended system with source documents has economical advantages. Internally this provides competitive advantages; externally the stored information can be sold. Consequently, a good in formation retrieval system is required. The benefit of a computerized retrieval system is the possibility of combining multiple criteria for searching for information. But today most computer applications are not flexible enough to operate without predefined structures for information retrieval Single items and words are limited to functions they are related to. For example, in conventional retrieval languages like the Common Command Language for European hosts, it is necessary to mark the word AND in a specific way if it is a searched string like "ANDY" and not an operand, as in "Data AND Management". An application is needed which has less formal structures and is easy to use. Extract: MUMPS Language Extensions Provided by GNOSIS MUMPS Language Extensions Provided by GNOSIS The PROLOG features of GNOSIS are used to describe the problem. Usually a description consists of three components. First, names and structures of objects involved in the problem. Second, names of relations that are known to exist between the objects. And third, facts and rules describing these relations. GNOSIS uses facts and rules. A GNOSJS program consists of a list of logical statements, either in form of facts such as: The book "Trusted Computer System Evaluation Criteria" from the DoD is ORANGE or in the form of rules such as: You will get help if you press the button ESCAPE and you enter an expression for support. Then deductions are possible. Given the facts: Support expressions are "???" and "HELP" or any abbreviation of these. the system can deduce that ESCAPE and "H" is a cry for help. The user can even give a goal to the program like: Find every possible combination of help characters. GNOSIS enhances the MUMPS language for the description of a problem with two new literals, a list literal and a predicate literal. List literals are used to describe and evaluate a group of literals. This group is tree-structured, and it can consist of values of variables, subscripts or expression atoms. Predicate literals are used as a blueprint for all given relations, like #paper("author",["Uchida","Suzuki","Smith","Wakai"]) New GNOSIS functions support the management of the literals, including the following: ? $LAPPEND appends a list of literals ? $LEXTRACT extracts one or more terms from the list ? $LFIND finds literals from the list ? $LLENGTH counts the number of literals in the list An unary operator "&" supports list processing in GNOSIS. It is used to list expression atoms; for instance X and Y are literals from a list, X=["database"] and Y=["management"]. Then the unary operator creates anew literal, &(X_Y) =["databasemenagement"]. A PROVE command is the execution of predicate calculus, the so-called inference machine. PROVE sets the $TEST variable to 1, if facts are found in the MUMPS knowledge database, for instance PROVE [#paper] is true. |