M M d a t-i a; 0 7 Ji eä V i t - j ä f . ^ t\ ^ < t . "i 1 < ' H - 1 _ ■Si i ^ ^ Jrv p j I - Volume 17 Number 3 November 1993 ISSN 0350-5596 and Informatics <, L,. I - • y ■ ■ fÄ L, ,, , " - v * ,--"'S 1 --- - ^ —' - ■■ i.'''-'- ^ "" I i _X 1 , fj-' — 'i" ^ J ^^ 'i,- » V r r- -j ' - 4,1 ,>;■. ", f ! ! ^^ " Profiles r: Hubert Dreyfus ^^ - - r^ ^ r -ri,!-»- ^ -'»"Zf ' S - , , , ^ i "Ut m mm Principia Cybernetica Project f\':Telecommumcàtiohs in Slovenia? l'iil-i: ' ,^ > ' V - ■ vi. u, t , -, , äs ^ r ^ ^ . I , S.1. i'. - , 4 t 1-, ^ 3 * ^ t 4- ■i —I ^ - , T" fSž The Slovene Society ^formatìka, Ljubljana, Slovenia f Informatica An International Journal of Computing and Informatics Subscription Information ^ . Informatica (ISSN 0350-5596) is published four times a.year in Spring, Summer, Autumn, and Winter (4 issues per year) by the Slovene Society Informatika, Vbžarski pot 12, 61000 Ljubljana, Slovenia. The subscription rate for 1993 (Volume 17) is - DEM 50 for institutions, - DEM 25 for individuals, and ~ DEM 10 for students plus the mail charge. Claims for missing issues will be honored free of charge within six months after the publication date of the issue. BOeX Technical Support: Borut Žnidar, DALCOM d.o.o. Stegne 27, 61000 Ljubljana, Slovenia. Lectorship: Fergus P. Smith, AMIDAS d.o.o., Cankarjevo nabrežje 11, Ljubljana, Slovenia. Printed by Biro M, d.o.o., Žibertova 1, 61000 Ljubljana, Slovenia, Orders for subscription may be placed by telephone or fax using any major credit card. Please call Mr. R. Mum, Department for Computer Science, Jožef Stefan Institute: Tel, (+386) 61 1259 199, Fax (+386) 61 219 385, or use the bank account number LB 50101-678-51841. According to the opinion of the Ministry for Informing (number 23/216-92 of March .27, 1992), the scientific journal Informatica is a product of informative matter (point 13 of the I tariff number 3), for which the tax of traffic amounts to 5%. y. Informatica is published in cooperation with the following societies (and contact persons): ; ' Robotics Society of Slovenia (Jadran Lenarčič) i; Slovene Society for Pattern Recognition (Franjo Pernuš) j Slovenian Artificial Intelligence Society (Matjaž Gams) Slovenian Society of Mathematicians, Physicists and Astronomers (Bojan Mohar) Referees: D, Abel, M. Bohanec, M. Bonač, M. Drobnič, B. FilipiC, J. Fiimkranz, M.A. Jenkins, A. KaraliC, M. Kljajič, M. KovaCiC, J. LenarčiC, B, Likar, B. Litow, D. Mladenič, L Mozetič, M, Preäem, M. Rihar, B. RobiC, I. Savnik, L Sega, M. Tamiz, G. Trinidad, O. de Vel. The issuing of the Informatica journal is ßnanciälly supported by the Ministry for Science and Technology, Slovenska 50, 61000 Ljubljana, Slovenia. ^ ............' ' f I ' . - .......... ^ - , . - ' — An IntexTiational Journal of Computing and Informatics Contents: Editorial Program of Informatica . 213 Profiles: ILL. Dreyfus . 214 An 0(\/iV) Token Based Distributed Mutual Exclusion Algorithm . Kia MaUki Niki .Pissinou. Yelena Yesha 221 Quality of Decision Rules: Empirical and Statistical Approaches ' Ivan Bruha Sylva K očkova . 233 Logos of the Informational , Anton P. Zelezhikar 245 The Information Technology (IT) for Enabling Disabled People - A Strategic Agenda - On-Kwok Lai 267 A Combined Algorithm for Testing Implications of Functional . Mirko Maleković 277 Electronic Networking for Philosophical Development in the Principia Cybernetica Project Francis Heylighen Cliff Joslyn 285 To the Internal Representation Debate 294 Global Modernization Plan of Telecomrhunications Infrastructure and Services in Slovenia (MTIS) r : Gorazd Kandus Iztok Tvrdy 297 Reports and Announcements 314 ^ . - v^i EDITORIAL PROGRAM OF INFORMATICA For the readers and editors of Informatica it may be important to. know which criteria of successfulness (scientific significance) of the journal in the international context are we pursuing. Which are the criteria of excellence for editors i and what Icind of activity do we expect from our authors and readers? 1 Certainly, editors of Informatica have been chosen on the basis of their sci- I entific (publishing, research, and organizational) activity in the global and Eu- ropean environment. The choice was made intuitively considering the available i public references and personal communication. We began with publishing of 1 ' profiles of our distinguished editors (e.g., T. Winograd, J. Šlechta, and H,L. 1 Dreyfus) and will continue with provoking interviews. But, in the coming pe- I riod we should appreciate to have the evidence concerning, for example, the j ten most important articles in the last ten years for each of our editors. Such I an overview would be extremely interesting for our readers, for it would show the complexity of themas pertaining to the fields of computation, artificial intelligence, robotics, informatics, and the revived area of cybernetics. The second important goal of Informatica is to get a firm position in the so: called citation index and reviewing ambience (e.g. data bases such as SSCI, SCI and TCI, and special reviewing journals). Since Informatica covers different, in an interdisciplinary manner connected scientific domains, the citation index Ì for an author could be multiplied over several disciplines, bringing him or her a higher score (citation value) of publications. This perspective of disciplinarily \j interwoven regions (e.g., computing, mathematics, philosophy, robotics, cyber- I netics) concerns both editors and authors and would represent an advantageous ; challenge in promotion and circulation of the journal. ( My advice to the editors, authors, and readers of Informatica is to use as \ much as possible citations from Informatica, even of the papers belonging to the author himself/herself, but published in another scientific journal. Informatica will coUect such data at the end of every fiscal year and use them in a competition for financial support of its donators, The publishing program of Informatica remains as it was traced in the previous Editorials. We are an international interdisciplinary scientific journal, joining several areas of science and philosophy—concerning computation, informatics, robotics, mathematics, and cybernetics. We foster some new disciplines belonging to the field of the informational, for instance, the theory of chaos, photonics, quantum dynamics, social informatics, economy theories, new formalisms, etc. Our goal is not to become a journal with the so-called hard-disciplinary orientation, which are nowadays numerous and still popular, but do not consider the phenomenon that is already happening on the global scene: a changing paradigm of sciences and their future significance for the mankind. Simultaneously, we have to observe the globally relevant projects which may change the nowadays understanding substantially in the direction of the cy-berhetical, formalistic, and technological. AU these projects could importantly improve the scope of the informational. I expect the help and understanding of the international community of Informatica, building up a new opportunity for publishing of interdisciplinary achievements in the broad realm of informatics. —Anton P. Železnikar, Editor-in-chief PROFILES The introduction of the position and the work of professor Hubert L. Dreyfus (e.g., his curriculum vitae in a factical form), the most famous criticizer of the computer and artificial intelligence conceptualism and philosophy (hopes, beliefs, technology), is anything but a simple task. To write down an analytical version of his curriculum vitae would take a substantial effort concerning the analysis and interpretation of the po: sitions, attitudes, and work of professor Dreyfus. It would be possible right now to pick out some of his significant achievements, but this would make the whole story shortened and deficient. From informative point of view, I believe very much, the readers of Informatica will appreciate to have as complete as possible data of his activity on this and the next pages. After a time and by the help of additional sources, I wiU make an attempt to write an exhaustive analysis and interpretation of the listed data and evaluate the significance of them for the philosophy, practice of computation, artificial intelligence, and informatics in the coming epoch. Hopefully, the reader will be able to get his/her own overview of professor Dreyfus' academic, philosophical, and scientific activity from the listed data. Hubert L. Dreyfus Education B.A. Addresses 1116 Sterling Avenue Berkeley, CA 94708 (415) 841-9926 FAX: (415) 642-4164 SS #335-28-6107 Department of Philosophy University of California Berkeley, CA 94720 (415) 642-7463 or 2722 e-mail: dreyfusScogsci. berkelsy. edu Harvard, 1951 (highest honors in philosophy) M.A. Harvard, 1952 Ph.D. Harvard, 1964 Academic Activity —Assistant, General Education (Natural Science), Harvard, Summer 1952. —Teaching Fellow, General Education (Humanities), Harvard, 1952-53,1954 and 1956. —Instructor in Philosophy, Brandeis, 1957-59. —Assistant Professor of Philosophy, M.LT., 1960-66. —Associate Professor of Philosophy, M.I.T., 1967-68. —Associate Professor of Philosophy, University of California at Berkeley, 1968-72. —Professor of Philosophy, University of California at Berkeley, 1972 to present. —Director, N.E.H. Summer Institute, University of California, Berkeley, 1980. —Director, N.E.H. Summer Seminars, University of Caüfornia, Berkeley, 1981, 1983 and 1984. — Visiting Professor, Technical University, Vienna, Austria, May-June 1986. —Director, N.E.H. Summer Institute, University of California, Santa Cruz, 1988. — Visiting Professor, Frankfurt University, Summer 1989. —Co-Director, N.E.H. Summer Institutes, University of California, Santa Cruz, 1990, 1992. — Visiting Professor, Ecole Normale Supérieure, Paris, May 1991. — Visiting Professor, Technical University, Vienna, June 1991. Honors and Awards —Phi Beta Kappa, 1951. —Harvard Sheldon Traveling Fellowship, 195354, (Freiburg). —Fulbright to Belgium, 1956-57, (Husserl Archives, Louvain). —French Government Grant, 1959-60 (Ecole Normale Supérieure, Paris). —Baker Award for Outstanding Teaching, 1966. —NSF Grant: January 1968-September 1968; Research Associate in Computer Sciences, Harvard Computation Laboratory. —Harbison Prize for Outstanding Teaching, 1968. —ACLS Grant, 1968-1969. —Guggenheim Fellow, 1985. —NEH Basic Research Grant, 1988-1989. —Yrjö Reenpää Medal, Finnish Cultural Foundation, 1991. —Phi Betta Kappa Lecturer, 1992-93. Publications Articles: —(with James Broderick) "Curds and Lions in Don Quixote", Modern Languages Quarterly (June 1957). —(with Samuel J. Todes) "The Three Worlds of Merleau-Ponty", Philosophy and Phenomenologi-cal Research (June 1962). —(with Joseph Pequigney) "The Landscape of Dante's Inferno", Italian QtiorfeWj/(Spring 1962). —"Wild on Heidegger", Journal of Philosophy (October 1963). —(with Joseph Pequigney) "Landscape and Guide: Dante's Modifying of Meaning in the Inferno", M.I.T. Publication in the Humanities, Number 66. —"Merleau-Ponty's Existential Phenomenology", M.I.T. Publication in the Humanities, Number .69. —"Philosophie aux Etats-Unis", Encyclopédie comparée U.S.A. - U.R.S.S., Laffont, Paris (November 1967). —"Alchemy and Artificial Intelligence", RAND Paper P-3244 (December 1965). (Japanese translation, September 1967.) —"Phenomenology and Artificial Intelligence", Phenomenology in America, James Edie (ed.), Quadrangle Press (1967). (Reprinted as "Do Computers Think?", in Reading for Liberal Education, Holt, Rinehart and Winston, 1967.) —"Why Computers Must Have Bodies in Order to be Intelligent", Review of Metaphysics (September 1967). (Reprinted in La Mente e le Macchine, M. Bianca, editor, La Scuola Editrice, 1978.) —"Philosophical Issues in Artificial Intelligence", M.I.T. Publication in the Humanities, Number 80. —"Pseudo-Strides Towards Artificial Intelligence", Theoria to Theory, Volume 2, Second Quarter (January 1968). —"Cybernetics as the Last Stage of Metaphysics", Akten des XIV Internationalen Kongresses fur Philosophie, Vienna (1968). —"The Perceptual Noema; The Suppressed Originality of Aron Gurwitsch", Life-World and Consciousness, ed. Embree, Northwestern University Press (1970) (reprinted in Husserl, Intentionality and Cognitive Science - see below), —(with Samuel J. Todes) "The Existentialist Critique of Objectivity", Patterns of the Life-World, ed. Edie, Northwestern University Press (1971). —"Phenomenology and Mechanism", Nous (March 1971). —"A Critique of Artificial Reason", Thought (1968). Reprinted in Euman and Artificial Intelligence, ed. Crosson, Appleton - Century Crofts, (1971); in Interpretations of Life and Mind, ed, Marjorie Grene, Routledge, Kegan, Paul, (1971); and in Dialogos (February 1973). —"Sinn and Intentional Object", Existential Phenomenology, ed. Robert Solomon, Harper and Row (1973). —(with John Haugeland) "The Computer as a Mistaken Model of the Mind", Philosophy and Psychology, ed. S.C. Brown, Macmillan (1974). Translated in Portuguese (1984). —"Artificial Intelligence", The Annals of the American Academy of Political and Social Science (March 1974). —"Human Temporality", The Study of Time II, J.T. Eraser and N. Lawrence, editors, SpringerVerlag (1975). —"The Priority of The World to My World: Heidegger's Answer to Husserl (and Sartre)", Man and World (May 1975). —"The Misleading Mediation of the Mental", Philosophical Dimensions of the Neuro-Medicai Sciences, S.F. Spicker and H.T. Englehardt, editors, Reidel (1976). —(with John Haugeland) "A Husserl - Heidegger Dialogue: Philosophy's Last Stand", Heidegger and Modern Philosophy, edited by Michael Murray, Yale University Press (1977). —"Things Computers Still Can't Do (Part I)," ROM, Vol. I, No. 9, (March/April 1978). "What Computers Still Can't Do (Part II)," Creative Computing, Vol. 6, No. 1, (January 1980). —(with Stuart Dreyfus) "Inadequacies in the Decision Analysis Model of Rationality", Foundations and Applications of Decision Theory, ed. C. Hooker, Reidei (1978). —"Empirical Evidence for a Pessimistic Prognosis for Cognitive Science", Behavioral and Brain Sciences, Vol. 1, Number 1 (1978). —"Cognitive Psychology vs. Hermeneutics", Behavioral and Brain Sciences, Vol. 1, Number 2 (1978). —"Incompatible Constraints on the Input/Outputs of a Human Intentional Black Box," Behavioral and Brain Sciences, Vol., Number 2 (1978). —(with Stuart Dreyfus) "The Scope, Limits, and Training Implications of Three Models of Aircraft Pilot Emergency Response Behavior", Operations Research Center Report (February 1979). —(with Stuart Dreyfus) "The Psychic Boom: Flying Beyond the Thought Barrier", Operations Research Center Report (March 1979). Reprinted in Creative Computing (1981). —"A Framework for Misrepresenting Knowledge", Philosophical Perspectives in Artificial Intelligence, ed. M. Ringle, Humanities Press (1979). —(with Stuart Dreyfus) "A Five-Stage Model of the Mental Activities involved in Directed Skill Acquisition", Operations Research Center Report (February 1980). —(with Stuart Dreyfus) "Proficient Adaptable Response to Emergencies Caused by Identifiable Malfunctions: Contrasting Training Implications of Two Proposed Models", Operations Research Center Report (February 1980). —"Existentialism," "Phenomenology," "Martin Heidegger," "Edmund Husserl," "Karl Jaspers," "Soren Kierkegaard," "Gabriel Marcel," "Maurice Merleau-Ponty," "Franz Rosenzweig," "Max Scheler," in Funk and Wagnalls' New Encyclopedia, (1980). —"Les Ordinateurs peuveat-ils etre vraiment inteUigents?", Critique, Les Philosophes Anglo-saxons par eux-memes, (Aout-Septembre 1980), —"Holism and Hermeneutics", The Review of Metaphysics (September 1980). Reprinted in Hermeneutics and Praxis, ed. Robert Hollinger, University of Notre Dame Press, (1985). —"Dasein's Revenge: Methodological Solipsism as Unsuccessful Escape Strategy in Psychology," Behavioral and Brain Sciences, Vol. 3 (1980) —"Knowledge and Human Values: A Genealogy of Nihilism", Teachers College ižeconf (Spring 1981). —"From Micro-Worlds to Knowledge Representation: AI at an Impasse", Mind Design, edited by John Haugeland, Bradford Books (1981); reprinted in Readings In Knowledge Representation, Ed. Ronald J. Brachman and Hector J. Levesque, Morgan Kaufman Publishers, (1985). —(with Piotr Hoffman) "Sartre's Changed Conception of Consciousness; From Lucidity to Opacity", Library of Living Philosophers on Sartre, ed. P.A. Schilpp, Open Court Publishing Co. (1982). —Response to John Searle's "Meaning", Colloquy 44, The Center for Hermeneutical Studies in Hellenistic and Modern Culture, The Graduate Theological Union and The University of California, Berkeley (October, 1982). —"How We Behave", Michel Foucault Interview, Vanity Fair (November 1983). —"Between Techné and Technology: The Ambiguous Place of Equipment in Being and Time", Tulane Studies in Philosophy, Tulane University, Volume XXXII, (1984). French translation, "De la techné à la technique" in Martin Heidegger, ed. Michel Haar, Cahier de L'Herne, Number 45, Editions de L'Herne, Paris, France (1983); reprinted in Livre de Poche. —"Beyond Hermeneutics: Interpretation in Later Heidegger and Recent Foucault", Hermeneutics: Questions and Prospects, edited by Gary Shapiro, University of Massachusetts Press (1984). Interpreting Politics, Michael T. Gibbons, ed., Basil Blackwell, (1987). —"What Expert Systems Can't Do", Raritan (Spring 1984). —(with Stuart Dreyfus), "Putting Computers in Their Proper Place: Analysis versus Intuition in the Classroom", Columbia Teachers College Record, Vol. 85, No. 4 (Summer 1984). Reprinted in Run: Computer Education^ 2nd. Ed. by Harper and Stewart, (1985), and in The Computer in Education, A Critical Perspective, Douglas Sloan ed., published by Teachers College Press, (1985). Reprinted as "L'ordinateur à sa place: Analyse ou intuition dans la s alle de classe", in Le Temps de la Reflexion, ed. by Marc Froment-Meurice, Gallimard, (October 1985). —"H n'y a pas d'ordinateur intelligent", Psy- chologies. No. 13, (Juillet-Aout 1984), Paris, France. ^(with Stuart Dreyfus), "Mindless Machines: Computers Don't Think Like Experts, and Never Will", The Sciences, The New York Academy of Sciences (November/December 1984). —(with Stuart Dreyfus), "Skilled Behavior: The Limits of Intentional Analysis", Phenomenolog-ical Essays in Memory of Aran Gurwitsch, ed. Lester Embree, The University Press of America, Washington, B.C. (1984). —"The Role of the Body in Intelligent Behavior", (Chapter 7 of What Computers Can't Do) reprinted in Philosophy, Technology and Human Affairs, ed. Larry Hickman, Ibis Press (May 1985). —(with Stuart Dreyfus) "From Socrates to Expert Systems: The Limits of Calculative Rationality" Philosophy and Technology II: Information Technology and Computers in Theory and Practice, edited by Carl Mitcham and Alois Hüning, Boston Studies in the Philosophy of Science Series, Reidel, (1985); reprinted in Technology In Society, Vol. 6, (1984), and in Oblong, Department of Architecture, University of California, Berkeley, (Summer 1986). Rewritten and presented to the American Academy of Arts and Sciences, Bulletin, Vol. XL, Number 4, (January 1987). Revised and reprinted in Interpretive Social Science, A Second Look, Paul Rabinow and WiUiam M. Sullivan, editors, University of California Press, (1987). Revised and translated in French as "Laportee philosophique du connex-ionism", Introduction aux sciences cognitives, under the direction of Daniel Andler, Collection Folio/Essais, Editions Gallimard, (1992). —"Merleau-Ponty's Conception of Psychopathol-ogy", German translation in Leibhaftige Vernunft, eds. Alexandre Métraux and Bernhard Waidenfels, Fink Verlag, (1986). —"Why Studies of Human Capacities Modeled on Ideal Science Can Never Achieve Their Goal", Rationality, Relativism and The Human Sciences, ed. J. MargoEs, M. Krauz and R.M. Burian, Mar-tinus NijhofF Publishers, (1986). —(with Paul Rabinow), "What is Maturity? Habermas and Foucault on 'What is Enlightenment?'", Foucault: A Critical Reader, ed. David Hoy, Basil Blackwell, (1986); reprinted in Critique, (Aout-Septembre, 1986). Finnish transla- tion in tiede & edistys, (2/1986). —(with Stuart Dreyfus), "Why Computers May Never Think Like People" Technology Review, Vol. 89 No. 1., (January 1986); reprinted in Current, Number 283, (June 1986); Computers in the Human Context: Information Technology, Productivity and People, Tom Forester ed., MIT Press and Basil Blackwell, (Spring 89). —(with Stuart Dreyfus) "How to Stop Worrying about the Frame Problem even though It's Computationally Insoluable", The Robot's Dilemma: The Frame Problem in Artificial Intelligence, ed. Z. Pylyshyn, Ablex Publishing Corp., (1987). Revised and reprinted as "Coping with Change: Why People Can and Computers Can't", Wo steht die Analytische Philosophie heutel. Wiener Reihe, Band 1, (Spring 1986). Reprinted in LOGOS: Philosophic Issues in Christian Perspective, Vol. 7, (1986), Santa Clara University. —(with Stuart Dreyfus) "Putting Computers in-Their Place", Social Research, Volume 53, Number 1, (Spring 1986). —(with Stuart Dreyfus) "Why Expert Systems Don't Exhibit Expertise", IEEE-EXPERT, Volume 1, Number 2, (Summer 1986). —"Menschliche Sprache versus Computer Code", Technik und Gesellschaft, Memoranden des Forschungsinstitut für Technik und Gesellschaft der Technischen Universität Wien, Austria, (Fall 1986). —(with Stuart Dreyfus), "Why Skills Cannot Be Represented by Rules" ,j4(it;ances in Cognitive Sci-^ enee, ed. N.E. Sharkey, Elis Horwood Limited, (1986). Reprinted as "The Mistaken Psychological Assumptions Underlying the Belief in Expert Systems", in Cognitive Psychology in Question, ed. Alan CostaU and Arthur StiU, Harvester Press, (1987). , . ^"Foucault's Therapy", PsyCritique, (Fall 1986); reprinted in Wiener Reihe: Die Philosophen und Freud, Band 3, as "Foucault versus Freud" (1989); "Foucault's Critique of Psychiatric Medicine" in The Journal of Medicine and Philosophy, Center for Ethics, Medicine, and Public Issues, Baylor College of Medicine, Houston, Texas, (1987). —(with Stuart Dreyfus), "Competent Systems: The Only Future For inference-Making Computers", Future Generations Computer Systems, North-Holland, Vol. 2, Number 4, (December 1986). \ —"Misrepresenting Human Intelligence", The Tradition of Philosophy, ed. Hariison HaU and Norman Bowie, Wadsworth, (Fall 1986); reprinted in Thought, A Review of Culture and Idea, Ford-ham University Quaterly, (December 1986); Artificial Intelligence: The Case Against, Rainer Born, ed., Groom Helm, (Winter 1987). —(with Jane Rubin) "You Can't Get Something for Nothing: Kierkegaard and Heidegger on How Not to Overcome Nihilism", Inquiry, Vol. 30, Nos. 1-2 (March 1987). —"Husserl, Heidegger and Modern Existentialism", The Great Philosophers, An Introduction to Western Philosophy, edited by Bryan Magee, BBC Books, (Fall 1987). Portuguese translation: "Husserl, Heidegger e o Existencialismo Moderno", Os Grandes Filo'sofos, Editorial Presenca, (1989). —(with Stuart Dreyfus), "Making a Mind vs. Modeling the Brain: AI Back at a Branchpoint", Daedalus, (Winter 1988). Reprinted in The Artificial Intelligence Debate, M.I.T. Press, Stephen Graubard, ed., (Summer 1988), In Alternative Intelligence: Artificial Intelligence at a Crossroad, ed. Massimo Negrotti, Springer-Verlag, London, (February 1989). In The Philosophy of Artificial Intelligence, edited by Margaret A. Boden, Oxford University Press, (1990). —"Husserl's Epiphenomeaology", PersptctiveB on Mind, Herbert R. Otto and James A. Tuedio, eds,, D. Reidel Publishing Company, (1988). —"Artificial Intelligence: The Problem of Knowledge Representation", VIA 9, Re-Presentation, Journal of the Graduate School of Fine Arts, University of Pennsylvania, (Spring 1988). —"Socratic and Platonic Basis of Cognitivism", Artificial Intelligence & Society, Forli Issue, Volume 2 Number 2, K.S. Gill, ed.. Springer-Verlag London, (April-June 1988). Translated into Italian as "Si pu'o accusare Socrate di cognitivismo?", Nuova Civiltà delle Macchine, Rivista trimestrale di analisi e critica, Nueva Eri, Anno VI, Nu. 1/1 (21/22), (June/July 1988). Published with the title: "Socratic and Platonic Sources of Cognitivism", in Historical Foundations of Cognitive Science, pp. 1-17, J-C-Smith, ed., Kluwer Academic Publishers, (1990). Reprinted with the title: "Is Socrates to Blame for Cognitivism?" in Artificial Intelligence, Culture and Language: On Education and Work, Bo Go x :ranzon and Magnus Florin, eds., Springer-Verlag, (1990). —(with Jerome Wakefield), "From Depth Psychology to Breadth Psychology: A Phe-nomenological Approach to Psychopathology", Hermeneutics and Psychological Theory: Interpretive Perspectives on Personality, Psychotherapy, and Psychopathology, Stanley B. Messer, Louis A. Sass, & Robert L. Woolfolk, eds., Rutgers University Press, (1988). —(with Stuart and Renee Dreyfus), "Skillfully Coping with the World: Man vs. Machine", Museum Studies Journal, (1988). —(with Stuart Dreyfus), "On the Proper Treatment of Smolensky", Behavioral and Brain Sciences, (1988). —"Artificial Intelligence: The Prob- lem of Knowledge Representation", Encyclopedic Philosophique, ed. André Jacob, Presses Univer-sitaires de Prance, (1989). —"Alternative Philosophical Conceptualizations of Psychopathology", Phenomenology and Beyond: The Self and its Language, edited by Harold A. E)urfee and David F.T. Rodier, edts., Kluwer Academic Publishers, Dordrecht, (1989). German translation in Ubergänge, Vol. 15, (1989). —"On the Ordering of Things: Being and Power in Heidegger and Foucault", Michel Foucault, Philosophe, Le Seuil, Paris, (1989). Reprinted in Heidegger and Praxis, Spindel Conference 1989, The Southern Journal of Philosophy, Vol. XXVIII Supplement, (1990). Reprinted in Michel Foucaul, Philosopher, Routledge, Chapman and HaU, (1992). —(with Stuart Dreyfus), "Towards a Reconciliation of Phenomenology and A.I.", The Foundations of Artificial Intelligence: A Sourcebook, Derek Partridge and Yorick Wilks, eds., Cambridge University Press, (1990). —(with Stuart Dreyfus), "What is morality? A Phenomenological Account of the Development of Ethical Expertise", Universalism vs. Communi-tarianism, ed. David Rasmus sen, The MIT Press, (1990). Revised and reprinted in Revisioning Philosophy, James Ogilvy, edt., State University of N^w York Press, (1992). —(with Jerome Wakefield), "Action and the First Person", John Searle and his Critics, E. Lepore and R. Van Gulick, eds., Basil Blackwell, (1990). —"Defending the Difference: The Geistes/ Naturwissenschaften Distinction Revisited", Einheit der Wissenschafìen: Internationales Kolloquium der Akademie der Wissenschaften zu Berlin, Walter de Gruyter, (1991). Reprinted in The Harvard Review of Philosophy, (1991). —"Heidegger's Hermeneutic Realism", The Interpretive Turn: Philosophy, Science, Culture, Hi-ley, Bowan, Schiisterman, edts,, Cornell University Press, (1991). —"Heidegger on the Connection between Nihilism, Technology, Art and Politics" Cambridge Companion to Heidegger, ed. Charles Guignon, Cambridge University Press, (forthcoming 1992). —"Reflections on the Workshop on "The Self"", Anthropology & Humanism Quaterly, the American Anthropological Association, (forthcoming). —(with Stuart Dreyfus), "Frictionless Forecasting is a Fiction", (forthcoming). Reviews: —Review of Langan's The Meaning of Heidegger, Philosophical Review (July 1961). —Review of Sartre's Search for a Method, Philosophical Review (October 1966). —Review of Frederick Olafson's Principles and Persons, Philosophical Review {Jxiiy 1970). —Review of Frederick A. EUiston and Peter Mc-Cormick's Husserl: Expositions and Appraisals, Nous (1979). —"The Dasein as a Whole", review of Martin Heidegger's The Basic Problems of Phenomenology^ Times Literary Supplement (September 17,1982). —"Wittgenstein on Renouncing Theory in Psychology", review of Wittgenstein's Remarks on the Philosophy of Psychology, Vols. 1 and 2, Contemporary Psychology, Vol. 27, Number 12 (1982). —"What Can We Learn from Husserl?", review of David Bell's Husserl for Times Literary Supplement, (Spring 91). —Review of Frederick Olafson's Heidegger and the Philosophy of Mind, The Philosophical Review, Vol C, No. 3 (July 1991). Interviews: —"The Mind Machines", WGBH-TV Boston, NOVA, (FaU 1978). —"Concerning the limits of artificial intelligence", Dan Stromberg, Datalogi Linkoping, Tekniska Hogskolan, Universitetet I Linkoping, (December 1982). —"The Computers are Coming", Dan Rather, CBS Evening News, (September 5-9, 1983). —"Intuition versus Calculative Rationality", California Times, California Council for the Humanities radio program, (Fall 1983). —"Artificial bitelligence", Viennese television documentary, (May 1984). —(with James Quay) "To Be or Not To Be, That Is the Question: Some Comparisons between Human Beings and Computers — An Interview with Hubert L, Dreyfus", Federation Reports, The Journal of the State Humanities Councils, Vol. VIII, Number 1 (January/February 1985). —"Mind Over Machine", Jerry Robinson, Voltage: Technology and Human Society, (Spring 1985). ^"Currents", B^A^^r-TK Artificial Intelligence debate with Pamela McCorduck, New York, (March 21, 1985). —"Intelligences Artiflcielles", Canadian Broadcasting Corporation, CBUF-FM, Vancouver, (September 6 and 13,1985). —"Sand and Lightning", KQED-TV, Sciences Notes, San Francisco, (October 16, 1985). —"Intelligence Artificielle: Qu'est-ce qui empeche les ordinateurs de penser?" Actuel, No. 72, Paris, (October 1985). —"Kultur und Gesellschaft", Jürgen Tomm, Sender Freies Berlin TV, (October 1985). —"Intelligence ArtifìcieUe: La Menace-Entretiens avec les deux plus grands spècialistes mondi-aux de l'Intelligence Artificielle", Guitta Pessis-Pasternak, Les Nouvelles Littèraires, FaU 1985. —"Artificial Intelligence and Natural Language Interfaces", Computer Chronicles, KS AM-TV, (December 6,1985). —"Artificial Intelligence", Australian TV, (May 2, 1986). —(with Stuart Dreyfus), "Mind over Machine", KPFA, (March 24, 1986). —"The Great Philosophers - Program 12: Busserl, Heidegger and Modern Existentialism", Brian Magee, British Broadcasting Corporation-TV, (June 25,1986). —"Artificial Intelligence", Jim Bürge, "Horizon", British Broadcasting Corporation-TV, (Taped: June 26,1986 - Broadcasted: Spring 1988). —(with Stuart Dreyfus), "Mind over Machine", Jan Black's Journal, KCBS 74, San Francisco, (September 5,1986). —"Mind Over Machine", The Mind's Eye series of conversations hosted by Jeffrey Mishlove, Spectrum Foundation, San Rafael, Ccilifornia, (Taped: September 29, 1986; First broadcast; June 1618, 1987). Published on videocassettes in the following programs: Computers and the Mind^ Contemporary Perspectives in Philosophy, and Mind Over Machine, Conversations on the Leading Edge of Knowledge and Discovery with Jeffrey Mishlove produced by Thinking Allowed Prroduc-tions, (Fall 1988). —"Intelligente Computer gibt es in 300 Jahren", P.M.Computerheft, N.IO, (October 1986), —"AI Conversations; Interview with Philosopher Hubert Dreyfus", PC AI, (Summer 1987). —"Über die Grenzen künstlicher Intelligenz", interview conducted by A. Locker and published in Informatik Forum, No. 3/4, (Dec. 1987). —(with Stuart Dreyfus), "Artificial Intelligence in Computers Doesn't Work", for Social Thought, KALW-FM PubUc Radio, (Taped: March 21, 1989; Broadcasted: Sep. 14, 1989). —"The Machine that Changed the World", WGBH/BBC production, (April 1992). Books —(with Patricia AUen Dreyfus) Translation with Introduction of Merleau-Ponty's Sense and Nonsense, Northwestern University Press (1964). — What Computers Can't Do: A Critique of Artificial Reason, Harper and Row (February 1972). Published in Russian, Portuguese, Serbo-Croatian. Second edition, with new preface (1979). Revised second edition in French (Flammarion, 1984), German (Athenäum Verlag, 1985), Chinese (1987), Italian (Armando, 1989). Excerpted as "Die Grenzen der Küntstlichen Intelligenz", in gdimpuls. Gottlieb Duttweiler Instituts für Entscheidungsträger in Wirtschaft und Gesellschaft, No. 2, (June 1985), Rüschlikon, Switzerland. —(with Paul Rabinow) Michel Foucault: Beyond Structuralism and Hermeneutics, University of Chicago Press (1982); second edition (1983); Pi-ench translation: Michel Foucault, Un Parcours Philosophique, Editions Gallimard, Paris, France (1984). German translation: Jenseits von Struk- turalismus und Hermeneutik, Athenäum Verlag, (1987). Published in Spanish,Japanese and Portuguese translations, (1988). —(with Stuart Dreyfus) Mind over Machine: The Power of Human Intuitive Expertise in the Era of the Computer, Free Press (1986). German translation, Rowohlt Verlag, (1987). Japanese translation, ASCII Inc., (1987). Revised paperback edition with new Preface, (1988). —Reedition of Michel Foucault's Mental Illness and Psychology with a 40-page Foreword, University of California Press, (1987). ~~Being-in-the-World: A Commentary on Heidegger's Being and Time, Division I, M.I.T. Press (1991). — What Computer (Still) Can't Do, revised third edition, with new 50 pages preface, M.I.T. Press, (FaU 1992). —Heidegger and Foucault on the Ordering of Things, University of California Press, (forthcoming). Monograph —Heidegger and Foucault on the Ordering of Things, The Hamline Review, Vol. 13, N. 2/Fall 1989. Anthologies —Husserl, Intentionality and Cognitive Science, Bradford/M.LT. Press (1982). Editor's Introduction reprinted in Grundprobleme der kognitiven Wissenschaft, ed. G. Heyer and D. Munch, Suhrkamp Verlag, (forthcoming). —(with Harrison Hall), Heidegger: A Criticai Reader, Basil Blackwell, (1992). Video-tape Lecture Series —Beyond Philosophy: The Thought of Martin Heidegger, Twelve two-hour lectures, available from the University of California Extension Media Center (1982). —Brain and Mind, Symposium, The Johns Hopkins Video Library, (1983). Edited by Anton P. Železnikar An 0{\/iV) Token Based Distributed Mutual Exclusion Algorithm Kia Makki Department of Computer Science, University of Nevada, Las Vegas Las Vegas, Nevada Niki Pissinou Center For Advanced Computer Studies, University of Southwestern Louisiana Lafayette, Louisiana Yelena Yesha Computer Science Department, University of Maryland Baltimore County Baltimore, Maryland Keywords: distributed systems, distributed mutual exclusion, critical section, distributed algorithms, communication networks Edited by: Xindong Wu Received: August 10, 1993 Revised: October 27, 1993 Accepted: October 28, 1993 In this paper; we present a toJcen based distributed mutua/ excJusion aJgoritiim for a distributed computer system of N sites. The proposed algorithm is based on timestamps and the theory of finite projective planes. It aiso makes use of a "Token Queue" which is part of the token and contains a h'st of a,il sites which are requesting the token. The algorithm is deadJock free, free from starvation, and requires no message exchange in the best case, and 4-\/W — 2 message exchanges in the worst case per criticai section execution. 1 Introduction One of the most interesting and fundamental problems in the area of distributed systems is the problem of mutual exclusion. The problem is to guarantee the integrity of a shared resource called critical section (CS) by restricting the use of such a resource to one site at a time in a distributed computer system which consists of a set of N distinct sites that communicate with each other only by sending messages over a communication network. In such systems, due to the lack of shared memory and a global clock and due to an unpredictable message delay, the task of designing a distributed mutual exclusion algorithm that is free from deadlock and starvation is much more complex. Over the past several years many algorithms to achieve mutual exclusion in. distributed computer systems have been proposed [1, 8, 6, 12, 13,14, 9, 16, 15, 11, 17, 21, 18, 19, 23, 24, 25]. These algorithms are briefly discussed in the next section. Li this paper, we present a token based distributed mutual exclusion algorithm for a distributed computer system of N sites. The proposed algorithm is based on timestamps and the theory of finite projective planes. It also makes use of a "Token Queue" which is part of the token and contains a list of all sites which are requesting the token. The algorithm is deadlock free, free from starvation, and requires no message exchange in the best case, and Ay/N - 2 message exchanges in the worst case per critical section execution. The remainder of this paper is organized as follows: In the next section we provide a brief overview of related work, In section 3, we give the preliminaries for our algorithm. The description of the algorithm is given in section 4, foUowed by the correctness of the algorithm in section 5. Section 6 discusses the performance analysis of the algorithm. We conclude with some final remarks. 2 Related Research During the last decade, a significant amount of effort has been focused on the development of the efficient distributed mutual exclusion algorithms. One of the earliest efforts in the area was by Alsberg and Day [2], They appoint a central node to manage the access to a shared resource and the sites wishing to enter their critical sections need to get permission from this central site. This solution is highly efficient in that it requires only three messages per critical section execution, regardless of the size of the system or the density of requests. However, an obvious drawback of this solution is that the central site is unfairly burdened with responsibility and with message traffic. Thereafter many distributed mutual exclusion algorithms have been proposed. In general, these algorithms can be classified into two groups [23] permission based and token based. In the following subsections we describe the characteristics of each group and briefly survey some of the representative algorithms from each group. 2.1 Permission Based Algorithms In permission based algorithms, in order for a site to gain access to the critical section, permission must be requested from some well defined set of sites referred to as its request set. The requesting site may enter its critical section as soon as it receives a reply to each one of its requests, and must send a release message to each site in its request set when it is finished, allowing them to permit other sites to enter the critical section. In permission based algorithms, in order to ensure that every site resolves conflicting requests in the same way, a unique timestamp is included in each critical section request message to create a logical consistent ordering of requests. In such algorithms, a site which is not currently requesting wiU send replies to aU requests that it receives. A site which is requesting will immediately send replies to those requests which are timestamped before its own request, and wiU delay replies to other sites until after it is finished with its own critical section. Lamport [6] was the first to design a fuUy distributed permission based mutual exclusion algorithm using logical timestamps. In his algorithm each request set is the entire network. Then, if N is the number of sites in the system, and if self-messages are not counted, the algorithm requires N - I requests, N - 1 replies, and N - 1 releases; or 3(N — 1) messages per critical section execution. Ricart and Agrawal [18] realized that if all sites must grant permission by sending replies, then the release messages are superfluous, since a reply involves an implicit release. Therefore, they have reduced the number of messages in Lamport's algorithm to 2{N - 1). Carvalho and Roucairol's algorithm [4] has further improved the number of messages in Ricart and Agrawala's algorithm by avoiding some unnecessary requests and reply messages. The idea behind this algorithm is that, if site i has not received a request message from site j, and since site i has executed the critical section last time, then site j has implicitly given permission to site i. Since site i does not have to ask permission from such a site the number of messages is reduced. They have shown that the number of messages exchanged per critical section execution in their algorithm is between 0 and 2(JV - 1). Maekawa [8] further reduces the number of messages per critical entry to cy/N (where c is a constant between 3 and 5) by imposing a logical structure on the network. Maekawa's algorithm is significantly different from the others. In his algorithm, in order for a site to enter the critical section, a site must get permission from some set of sites, Maekawa uses the theory of finite projective planes to group the sites. The interesting aspect of Maekawa's algorithm is that the size of each of these sets is \/iV, where N is the number of sites in the network, Maekawa's algorithm is prone to deadlock because a site is exclusively locked by one requesting site at a time, and requests can arrive in any order. Since our algorithm utilizes some of the concepts used in Maekawa's algorithm, a detail description of this algorithm is given in section 3, Sanders [22] developed a unifying framework for a large class of the permission based algorithms by defining the information structure that each site maintains. In this framework, when a site wants to enter the critical section, it sends request messages to a set of sites specified by its information structure and it may enter the critical section only when it receives permission from all the sites in this set. The concept of information structure Year : Discoverer # of Messages # of Messages # of Messages Best Case Worst Case Average 1978 Lamport 3(iV - 1) 3(iV-l) 3(JV - 1) 1981 Ricart and Agrawala 2(iV- 1) 2{N - 1) 2(JV- 1) 1983 Carvalo and Roucairol 0 2{N - 1) — 1985 Maekawa 3%/iV — 1991 Agrawal and Abbadi 0{logN) (N + l)/2 — 1992 Singhal 0 2{N - 1) (JV-1) Table 1. Chronological Order of Permission Based Distributed MutuaJ Exclusion Algorithms was later employed by Singhal [24] who proposed a dynamic information structure algorithm which reduces message traffic by cleverly initializing an information structure and updating it as the algorithm evolves. In fact, Singhal's dynamic information structure algorithm uses a similar strategy to reduce the number of messages as in Carvalho and Roucairol's algorithm. Singhal's algorithm requires between 0 and 2(iV — 1) messages and on the average in light traffic, requires (JV — 1) messages. More recently Agrawal and El Abbadi [1] have presented an efficient algorithm for solving this problem. The algorithm Imposes a logical tree organization on the site of the network. They show that their algorithm requires 0{logN) messages per critical section execution in the best case and 0{N) messages per critical section execution in the worst case. Table 1 lists some of the permission based mutual exclusion algorithms in a chronological order. . 2.2 Token Based Algorithms In the token based algorithms the token is a "unique and singular" message (also known as the PRIVILEGE [25]) which circulates among the sites. Ordy the site which possesses the token may enter its critical section. The various token based algorithms are distinguished by the methods for determining how a site obtains the token, and where a site sends the token when it is finished with its critical section. One of the earliest token based mutual exclusion algorithms is by LeLann [7]. In this algorithm a token circulates on a ring of sites. A site wishing to enter the critical section simply waits for the token to arrive, it captures the token from the ring, it enters the critical section and then it simply passes the token back to the ring. This rather simple scheme results in requiring one constant size message (the token) being transferred per critical section invocation when all sites are waiting to enter to their critical sections. Otherwise, if m sites are waiting to enter to their critical sections in one complete cycle of the token all of these are satisfied in N messages or on the average N/m messages per critical section invocation is required. Suzuki and Kasami [25] present a token based distributed mutual exclusion algorithm which is an improvement over the well known Eicart and Agrawala [18] permission based algorithm. Suzuki and Kasami's algorithm assumes a logically fully connected network topology and the existence of a token queue. In this algorithm a requesting site sends request messages to all other sites in the system. This guarantees that the site possessing the token will receive a request message and win enqueue it into the token queue. Each request is labeled with a sequence number, and the token message also contains an array with the sequence numbers of the last request satisfied at each site. As such, late arriving messages (i.e., requests which arrive after the requesting site has been serviced) can be ignored. In this algorithm the number of messages per critical section execution is iV" — 1 request messages plus 1 token message or N messages per each critical section entry. Their algorithm was later improved by Ri-cart and Agrawala [19]. Ricart and Agrawala's algorithm [19] also requires N messages per critical section invocation. The main difference is that there is no token queue. In recent years some very efficient token based algorithms have been proposed. One of the earliest one is due to Naimi and Trebel [17] which does not use sequence numbers. Instead, they organize the sites into a dynamic, logical, rooted tree. When a site invokes mutual exclusion, it sends its requests to the site which is possibly holding the token. When a site receives a request message, it will either forward this request message to another site possibly holding the token (if it is not requesting and it does not hold the token) or add all incoming request messages to its queue and appends the queue to the token message. Here there is no guarajitee that a request will, in one step, reach a site that is expecting the token, but a request is guaranteed to reach the token in a finite amount of time. Performance of the algorithm will depend on the average height of the tree and on the extent to which requests need to be forwarded. This algorithm requires between 0 and N messages and on the average 0{logN) messages per critical section execution. Another tree based mutual exclusion algorithm is due to Raymond [21]. In this algorithm the network topology is assumed to be a static unrooted tree (a spanning tree of the actual network) and it uses this tree for locating the token. Each site communicates only with its neighboring sites and holds information about its neighbors. In addition, each site knows which of its neighbors is on the path to the root where the token is. This site is referred to as Near. A site requests the token by sending a request message to its Near site. A sequence of request messages are sent between the requesting site ajid the site holding the token, until a request message arrives at the. site holding the token. Then the token is passed along the same path in the reverse direction. As the token passes through, the direction of the edges traveled by the token is reversed such that every path always leads to the site holding the token. An interesting characteristic of this algorithm is that each site needs to know only its neighbors in the tree. In the best case, this algorithm requires no messages per critical section execution, since the token may be available locally. Raymond shows that for a radiating star topology, the number of message exchanges per critical section is O(logN) on the average. However, in the worst case the number of messages per critical section is proportional to the diameter of the network which can be 0{N). Singhal [23] developed a heuristicaHy-aided algorithm to achieve mutual exclusion which uses state information to reduce message traffic. In this algorithm the sites maintain information about the probable location of the token and when a site wants to enter its critical section it uses a heuristic to guess which sites are likely holding the token and send the token request messages only to those sites rather than to all of the sites. The number of messages exchanged per critical section execution in this algorithm is between 0 and N and the average number of messages exchanged in light traffic is {N + 1)/2. Recently MakM et al. [12] also proposed an algorithm which achieves mutual exclusion in a distributed system. The proposed algorithm makes use of a "Token Queue" which is a part of the token and contains a list of aU sites which are requesting the token. The token and queue are sent to the first site on the token queue followed by the second and third etc. Also each site maintains a record of which site is a good site to request the token from, so that a single token request can be made in order for a site to be placed on the token queue. Periodic update messages are sent to sites which are not on the token queue to inform them whenever the good site changes. In general, the algorithm requires (JV/m) + 1 messages per critical section where m is the number of sites on the token queue. The performance of the algorithm improves from JV 1 messages per critical section in an extremely light token request environment, to 2 messages per critical section in an extremely high token request environment. Table 2 lists some of the token based mutual exclusion algorithms in chronological order. 3 Preliminaries Given a distributed system with JV sites, Maekawas's algorithm [8] divides the sites into N subsets numbered 1 to iV each containing VW sites. In fact, at least one site is common between the subsets associated with any two sites. The subsets are chosen so that each subset is represented by at least one member of every subset. Each subset is usually referred to as a "request Year Discoverer # of Messages Best Case # of Messages Worst Case # of Messages Average 1978 LeLann 1 N Nim 1982 Suzuki & Kasarni N N N 1983 Ricart & Agrawala N N N 1987 Naimi Sz Trehel 0 N 0{logN] 1989 Raymond 0 0{N) OilogN) 1989 Singhal 0 0{N) (JV -1- l)/2 Ught traffic 1991 Makki et. al. 2 N + 1 N/m+1 ■ Table 2. Chronological Order of Distributed Token Based Mutual Exclusion Algorithms set," Ri, for Ì = 1 to iV, In this algorithm a site i can enter its criticai section only when it receives permission from aU the sites in its request set iž;. Mutual exclusion is enforced because a site gives permission to only one site at any time. Formally, the subsets must satisfy the following four properties: 1. For any combination of i and < i,j < N, Ri n Äj ^ 0 2. Ri, 1 < i < N^ always contains i. 3. The size of iž,-, j Ri [, is k for any i, that is 1 Ä1 1 = I I = 1 «3 I = - I ÄJV I = fc (where N is related to A; by iV = k{k -1) +1). 4. Any j, 1 < j < iV, is contained in the k, iž,'s l 3 ■■ i = 1,2; = i 4 = 1,2 and J hj where rc is the number of training examples that are covered by the rule R and belong to the class C; rc is the numb er of examples covered by the rule R but not belonging to the class C, etc.; r = rc + rc is the number of examples covered by Ä; 2 Empirical Formulas The problem of defining rule quality remains an interesting issue in machine learning. Several attempts to define such a criterion have been made [1], [13], [4], [17]. However, the majority of these formulas represent an empirical, ad-hoc approach to the definition of rule quality. We emphasize the term 'empirical' here, since the work is based on intuitive logic and not necessarily backed by statistical or information theories. The two characteristics of consistency and completeness can be combined to yield a single number, the rule quality. The meaning of the result should be more or less understandable to human experts. The common policy is to use a weighted sum or multiyilication of the above components. (1) Rule quality as a weighted sum of the consistency and completeness has the form (named after one of its authors); QualityMìchalskì(R} = wi*cons(R)+ (1) W2 * compl(R) where wi,w2 G (0; 1) aie user-defined weights, usually + u)2 = Formula (1) is used in [1] as part of a more complex scenario which also comprises compre-hensibility and the cost of concept description. However, [1] does not state how to determine the above weights; only an 'experimental' determination of the weights is proposed. Formula (1) is also used in the CN2 unordered algorithm [5], and in CN4 with an unordered mode [8] as one of the selectable formulas. An interesting application of the weighted-sum formula for rule quality is used in the incremental learning system YAILS [17]. The weights in YAILS are not user-defined but specified by the algorithm itself: wi = 0.5 -f J + cons(Ji) W2 = 0,5 - i + cons{R) (2) These heuristic formulas follow from the results of experiments in [17] and emphasize consistency. Here, the weights are made dependent on consistency, which introduces a flexibility into the procedure: the larger the consistency value, the more influence consistency has on rule quality. (2) Rule quality as a product of consistency and completeness has the form (named after one of the authors): = cons{R) * f{compl[R)) (3) where / is an increasing function. Appropriate determination of f may, similarly to (2), prefer the consistency. The completeness works here as a factor that reflects a confidence to the consistency of the given rule. This scenario corresponds to the noisy-data processing paradigm, mentioned above. A large number of experimental tests performed in the knowledge integration system INTEG.3 [4] led its authors to the following form of the function / in the formula (3): f(x) = exp(a: - 1) 3 Statistical Formulas The empirical formulas for rule quality are easy to understand and provide quite comparative results. On the other hand, they are ad-hoc ones, without any theoretical support, Nevertheless, they may be interpreted and/or modified as reasonable schemes for defining rule quality. This paper discusses three such sources. (A) The theory on contingency tables seems to be one of these sources, since the performance of any rule can be characterized by them. Generally, there are two groups of measurements (see e.g. [3]): association and agreement. (1) A measure of association indicates a relationship between rows and columns in a 2 * 2 contingency table so that the i'-th row may be 'associated' with the j.-th column. In other words, this measure reflects an 'association' on both diagonals of the contingency table. Measures of association are described in detfdls in [3], We have found that the following two statistics may serve as reasonable schemes for the rule quality: —the Pearson x^-square statistic applied to the 2*2 contingency table: Quality^2(R) = ^ ^ ifij - fi+f+j) fi+f+j (4) _ (rc * TC — rc* rcy c*c*r*r I. Bruha et al. —the (?2-likelihood statistic or, if absolute'frequencies are used: Qualityg2(R) - (5) , , rc* K _ 1 2 + rc + In--h rc * In r*c Č * T J where logarithm here is to the base e. Both statistics are distributed asymptotically as x^ with one degree of freedom, It should be noted that the J-measure introduced in [16] and often quoted as one of the most promising measures [9], in fact, is equal to the G2-likelihood statistic divided by 2K: Qualityj{R) = Qualityg2{R) 2* K Qualitycohen{R) = (6) K *rc +K *rč-r*c — r*č K^ — r*c — f*Č or introducing consistency and completeness; Qualitycoh en (7) K + cons(R) * compl{R) — rc ^ cons{R}+compHR) _ ^^ From the viewpoint of machine learning, the above formulas work properly only if the class C is the majority class of examples covered by the rule. If not, then the rule quality symbolized by any measure of association may yield the largest values for the rule that does not cover any example of the class C. One may easily pi-ove that both statistics reach its maximum for rc = 0. (2) A measure of agreement is a special case of association that indicates an 'association' of the elements of a contingency table on its main diagonal only. The simplest measure proposed in [3] is the sum of the main diagonal: Al + /22 Cohen [7] suggests to compare the actual agreement Aagree = ^ fa with the chance' agreement t C agree = /t+Z+i which occurs if the row vari-i able is independent of the column variable, i.e., if the rule's coverage does not yield any information about the class C. The difference Aagree - Cagree is then normalized by its maximum possible value. This leads to the measure of agreement that we interpret as rule quality (named after its author): Qualitycohen{R) = -,- 1 - fi+J+i Note that the formulas (6) and (7) depend on four variables: r, c, rc^ K, or cons(Ä), compl{R)f rc, /1, respectively. Cohen's statistic provides a formula which responds to the agreement on the main diagonal of the contingency table, i.e. when both components rc and rc are reasonably large. Coleman [3] defines a measure of agreement that indicates an 'association' between the first column and any particular row in the contingency table. For the purpose of the rule quality definition, the agreement between the first column ('example from the class C") and the first row ('rule R covers an example') is the proper one. The formula follows the Cohen's statistic in principle by normalizing the difference between the 'actual' and 'chance' agreement. Hence, Coleman's formula for the rule quality is QualityCo!eman(R) = III ~ fl+f+1 fi-\- - /i+Z+i or, using the absolute frequencies, Ji + Tc — r ^ C Qualitycoiem^R) = -=- (8) r * c or, introducing consistency and completeness, Qualitycoleim.niR) = (9) K + cons{R) * compl{R) - rc K * compl^R) — rc Coleman's statistic intuitively exhibits necessary properties of a rule quality: agreement between 'rule covers' and 'example from class C\ Nevertheless, one can find, by simple analysis that Coleman's formula does not properly comprise the completeness of a rule; substituting rc by c * com,pl{R) changes (9) to Quality Coltman{R) = K + cons{R) — c has cons{R) = ^ as its estimate. The random variable 2 _ irc-r*p)'^ ^ {Tc-r*{l-p)f T *p T *{1 — p) On the other hand, Cohen's statistic is more completeness-based. Therefore, we have modified Cohen's formula in two ways to obtain these definitions: Qtiaiifi^SKiBi(Ä) = (10) QuaHtysKlB2{R) = (H) where the coefficients 2, 3 and 1, 2 have been added to normalize the statistics. (B) Information theory [15] is another source of statistical measurements suitable for defining rule quality. Following this theory, Kononenko and Bratlco [10] define the so-called information score for a classifier when classifying an unseen object; their formula can be also applied to rule quality: QualityiKJB(R) = - Ioga ^ + logj cons(R) (12) where the logarithm here is of base 2. More precisely, this formula is vaJid only if C con5(Jl) > — K which could be interpreted as a necessary condition for the reliability of a rule. (C) The vehicle of the confidence sets can be also exploited. The unknown conditional probability p = P(example of C\ rule R fires) has a x^ distribution with one degree of freedom. Hence, P(Y' (a as the basic implication pertcdning to the entity a extern U/ where informational operator Def has the meaning "informs (means) in the sense of a definition" or, simply, ''means by definition". □ The semicolon simultaneously denotes the parallelism and the end of a particular entity in a parallel array. It is necessary when entities are presented by formulas, which exceed a single line. □ Definition 4 (Informational sets) Informational sets are collections of entities (operands. formulas, and formula systems)^ which may be comprehended as informationally isolated items, where the question of their mutual informational interaction is not relevant. An informational set of elements is denoted by or A comma is the separator between isolated elements. Parallel notations of informational sets are permissible, that is, (a, ß, ß, ß, or , ■ or < • w \ w U! The comma has a similar function as the semicolon in the previous definition. □ 4 Axioms of Informing In this section we have to bring the duality of the presence (existing, arising) of an informational entity to the axiomatic surface. As a whole, in its entirety, or as an informational unit, informational entity a includes its own component of informing, marked by OTr in the functional (predicative) notation, by I{a). The informational duality of Q lies in its nature to be a form and process simultaneously, that is in the parallel nature of its operational activism and passivism within the same unity, on one side (expressions a |= and ^ a), and the explicit expression of entity informing la, on the other side. That what we have to bring face to face are verbal phrases to inform and to be informed, on one side, and the integral property of informing of entity, on the other side. In fact, we have to decide between different kinds of expression of the phenomenon of informing. If a 1= and j= a say that a informs and that a is being informed, respectively, than T^ represents the possibility of inforrning of o: in cc's entirety. In this respect, informing includes the expressive power of both a \= and \= a and introduces the informing of a as a distinguished operand within O!, that is, informational entity. Axiom 5 (Informing of an entity) [existentialism of Informing] We have the following senseful axiomatic implicational consequence: (« (« N; N a)) Entity a implies that it informs and that it is informed (informational phenomenalism); inform-ingness and informedness imply that entity has the entity, called informing Ja as an information-ally active part of the entity. □ In mathematical terms, la seems legai as a predicative function of a, that is, 1(a). It represents the systemic (phenomenal) informing of a, expressed implicitly as (a |=; 1= oc). On the other hand, operand-operator expressions a j= and 1= a as operator open entities may not fit the traditional mathematical agreement and function (mean) as a non-consistent novum. This situation becomes even more inconvenient in case of expressions of the form (•••((« N) Imagine, for instance, the following: both + a and a + may stiU represent a reasonable expression of a's additivity. But, what could be said in cases of -l-(-l-a), (a+)-!-, -|-(a+), etc. New terminology (meaning) of such algebraic open cases has to be introduced. Operator -I- always characterizes the operand, to which it belongs. Thus, in formula (-f- a), operator characterizes the operand a. Axiom 6 (Partitionism of informing) [In-clusiveness of an entity informing] Axiom =;> (la C a) says that informing of an entity implies that the informing is a part (i.e., an active part) of the entity. □ The last axiom sounds natural, so, no further discussion is necessary. Axiom 7 (Externalism of informing) [Entity Bounded Externalism of Informing] Externalism of an entity informing (informing's informingness) is closed to the entity. That is, I. (la N a) says that informing la informs its entity a in a closed way, □ Informing informs entity a. Axiom 8 (Internalism of informing) [Entity Bounded Internalism of Informing] Internalism of an entity informing (informing's informedness) is closed to the entity. That is, Xoi (a Iq) says that the informedness of an entity informing is caused in an entity-closed way. □ Informing is informed by entity a. Axiom 9 (Metaphysicalism of informing) [Informational Regularity of Metaphysicalism of Informing] As usually, J^a -^ (^Qf H ^c) Informing of an entity informs metaphysically (basic-cyclically) as any other informational entity. □ The last axiom assures that informing la can be metaphysically decomposed as a regular informational entity. Axiom 10 (Phenomenalism of informing) [Circular Bounding of Informing] Let ys introduce four implications for the specific phenomenalism of an entity informing: Informing's first mode of phenomenal circularity: la ^ ((a h la) h a) Informing's second mode of phenomenal circularity: (a h (Ja h ö)) Informing's third mode of phenomenal circularity: Informing's fourth mode of phenomenal circularity: la [Xa h (a h 1.)) The phenomenalism of an entity informing is closed to the entity itself. □ Let us interpret the last four modes of informing of an entity, which are characterized by specific cyclical ways. In the first mode, a informs Ja and, then, this process (subfoimula) as an entity informs a. While informing is directly informed by a (subformula a |= la), it informs a indirectly through the entity ct ]= 2q [formula (a 1= Xq) N III this mode, the a-cycle is evident. In the second - mode, a informs the process la t= a, in which informing Ja directly informs entity q. On the other hand, informing Ja is informed indirectly by a through the entity j= ct [formula (a |= (Ja |= o:)]. In this mode, the a-cycle is eyident too. In the third mode, entity ct informs Ja- in this mode, the metaphysical cycle of the form Ja 1= Ja is coming to the surface. In this cycle, Ja directly informs a, but entity a informs entity Ja indirectly, through the composed entity [= a. In this mode, the so-called Ja-cycle is evident. In the fourth mode, entity J^ informs the composed entity a |= In this mode, informing J^ informs entity a indirectly, but entity a directly informs its informing J^- In this mode, the Ja-cycle is evident too. Definition 5 (Informing of an entity) [Phenomenal Informing of an Entity] Informing Ja of an entity a is an a's phenomenalism, that is, an interweavement of different informational phenomena determined in the previous axioms. Formally, i-a ^Def < (JcCa; Ja h a; a h Ja, la H IQ 'Ì o h partitionism externalism internalism metaphysicalism phenomenalism 1 phenomenalism 2 phenomenalism 3 phenomenalism 4 This system of informing-(Ja)-definitional formulas can be understood to be a set of independent single formulas or a system of arbitrarily informa-tionally connected formulas. □ 5 Implicational Nature of Axioms Implication belongs to the information ally most common forms of expression. Implication means nothing else than a logical connection of informational entities, their implicitness in an informational context that always exists and informs in a spontaneously (intentionally, orient at ionaJly, worldly) interconnecting, interwoven, and cyclic way. Implication in informational axioms points to an artificial origin, which has its roots anywhere in the existing informational context and is posed by the observer or observer's intention, aim, truth, or belief. That is nothing else than a phenomenological reasoning or attitude, which pervades the today logic, sciences, and technological undertakings, By logical convention, informational implication a => ß is read in the following sense: a as an informational operand (entity, formula) implies ß as an informational operand (entity, formula). In this context, "a as an informational operand" has the meaning of the statement "it is true that a is an informational operand", which is in accordance with the implicational convention in ordinary (mathematical) logic. 6 How to Deduce Informational Consequences The informational is the most general term, which is not informationaUy particularized yet, The question is how to deduce the general and how the particular consequences (theorems) in the framework of a general and a particular informational theory. Some decomposition and composition philosophy, which concerns informational deduction, wiU be found in section 7. After that we have axiomatized and defined, the deduction of a general informational theory is a pure syntactic procedure, which never ends. There are inftnitely many general informational consequences, which can be deduced by pure syntactic means from the previous axioms and definitions. But, that does not hold for particular informational theories, which are axiomaticaUy bounded and correspond (model) particular informational cases. Thus, for instance, informational theories of discourse, time, understanding, a certain intelligence, etc. become senseful in an informational manner. Such particularized and complex informational entities are described by some initial informational formula systems, which develop and behave informationaUy. Consequence 1 (Informational Syntax) A general syntax of informational formulas follows from the previous axioms and definitions and is the following: ip informational entity: operand, formula, or formula system; any operand symbol: a, (Ö, • ■ ■, w or A,B,-- -,Z; {(p) parenthesized entity, tp j= entity externalism\ 1= ip entity intemalism; 'p\=

The listed example is in no way a non-scientific case. For instance, in quantum theory, parallel occurrences of one and the same particle (photon) can appear in different places (two places at once), traveling different routes and being detected when they interfere with one another (Quantum Magic and Quantum Mystery in [Penrose 90]). On the other hand, arising of information in mind as described by parallel-decomposing processes lies outside of the conventional doubt and is consciously evident. Similar explanation schemes can hold for the remaining cases of Rule 3 (internalism, metaphys-icalism, and phenomenalism). Consequence 3 (Parallelism of Informing) A consequence of Axiom 10 and Rule 3 is the parallelism of informing la- Thus, form («h; Nö)'^'Def Nparticularly5 Nparticularly ^^ Hparticularlyi Hparticulaily ^^^ Hparticvdarly' Nparticularly ^^ ^a / where a maintains its identity, but can inform in different (parallel) ways. Informational operators 1= particularly' Nparticularly' ' ' ' ' ^particularly differently particularized forms of the general operator h and read as 'inform(s) particularly'. This case is nothing other than an example of parallel decomposition of ct's informing. Definition 6 (A list of operands) Instead of a sequence of parallel formulas of the form the shortcut form ai,a2,---,an h ß can be used, where ai,a2, - •• ,Oin is a list of operands separated by commas. □ Definition 7 (A list of parallel formulas) A formula system j= ßi; «2 1= ^2; ■■■; On h means that formulas of the system, separated by semicolons, inform in parallel. This system of formulas can be marked, for instance, by (p and written as hmark ("l |= ßW "2 N /^2 i " ' ' 1 ßn) or also, in a parallel clear form, as By this consequence, parallel (dijferent, alternative) entities of informing, that is, ■ • • are introduced. □ The last consequence, introducing parallel occurrences of informing, which can be marlced as • • •, J^, has to be understood as a particular (particularized) parallel phenomenalism of the V Nmark ' föiNiöi; «2 h ß2] C^n^ßn The last form was already used in some previous cases. Certainly, operator [=inark can be replaced (particularized) by (in) any meaningly adequate form. □■ 7.2 Spontaneity, Circularity, Intentionality, and Alternativity as Entity-parallel and Entity-circular Phenomenalism The title of this subsection stresses that spontaneity, circularity, alternativity, and intentionality as informational phenomena is an entity-parallel and entity-circular plienomenalism. It means that these entities are interconnected, interdependent, and mutually supported in an informational (physical, biological, social, etc.) way, As observers of informational entities, we obviously consciously perceive this kind of interrelatedness. On the other side, spontaneity, circularity, alternativity, and intentionality seems to cover the most important aspects of human experience in philosophy, technology, and everyday life. Further, the four distinguished entities can spontaneously, circularly, alternatively, and intentionally resemble other significant informational entities which, may appear in an intricate way. One of the basic axioms concerning informational entity says that an informational entity informs in a spontaneous-circular or circular-spontaneous way, A special operator for this faculty of an informing entity can be introduced. Definition 8 (Operator of a circular-spontaneous informing) Let us introduce the following basic informational operator: a 9-» means a informs circularly and spontaneously The particularized general operator would look cumbersome, ^ [—circular—spontaneously for example. □ The four special operators are as follows by the next definition. Definition 9 (Operators of spontaneity, circularity, intentionality, and alternativity) We introduce the following four basic informational operators: ot means a informs spontaneously a O means a informs circularly a ^ means a informs intentionally a means a informs alternatively We can use also particularized general operators of informing, that is, ^ h^apontaneously ) circularly i <3! ^iatentionaHy ) ^ Nalternatively respectively. □ Although in the previous two definitions, the informational externalism of entity a is shown, the reader can easily complete the cases of inter-nalism, metaphysicalism, and phenomenalism by himself/herself. Consequence 4 (An implication concerning the spontaneous-circular informing) A consequence of definitions 8 and 9 concerning the spontaneoiis-circufar informing of an entity in general and its intricate particular informing s of spontaneity, circularity, intentionality, and alternativity is the following: (a // a a O; a M \ (a ^ o(0 0 o =())) where --4o(oo(;^o=()) is a notation of the so-called informational composition (operational unity or perplexity) of informational operators and Token 'o' marks the operator of operator composition. □ 7,2.1 A Structure of Spontaneity Spontaneity as an informational entity is in no way a purely chaotic phenomenon without boundaries in the spontaneity itself. Spontaneity does not mean a chaotic informational arising, but physical, biologic, discursive, social, etc. persevering, directionality, or intentionality. Simultaneously (in parallel), spontaneity as an informational entity is cyclically and alternatively structured, where alternativity can be understood as a specific form of spontaneity. Our tasi is to capture spontaneity of informational entities as a regular informational faculty, to give the phenomenon an explicit formal character. For this purpose, a special informational operator (-+) was introduced. Consequence 5 (An implication of spontaneity) Informational spontaneity means a certain circularity, intentionality, and alternativity of an informational entity. Thus, \ \ (a Oo(>-+o=s)) (a = or also, / (a / (a Oo(>-»o=})) Consequence 7 (An implication concerning general circularity) Informational circularity means a certain spontaneity, intentionality, alternativity, metaphysicalism, cyclic parallelism of an informational entity. This property of informational circularity means (ao, O a) (ct o(>-» o =ì)) ; where O o (>-» o is a notation of informational composition of informational operators O, >—and The duality of the first and the second expression is characteristic and can be observed elsewhere in the realm of the informational. □ Alternativity may be recognized as an explicit spontaneous faculty of choice among different possibilities of informing. However, circularity and intentionality may iiave their own characters of spontaneity in their circular and intentional possibilities. Thus, spontaneity stays in the background of any informational activity. 7.2.2 A Structure of Circularity Circularity (cyclicity) is one of the basic faculties of an informing entity. Circularity means recursiveness, metaphysical and other kinds of cycling, memorizing, maintaining of informational components, informational persevering which keeps a course of spontaneity, intentionality, alternativity, etc, Metaphysicalism is a basic interior mechanism of circularity. Cyclic processes can include parallel structures, representing circular-parallel forms of informing. Consequence 6 (Implications and a meaning concerning metaphysical circularity) Pertaining to metaphysicalism a |= a, there is, (aü) ==> (a ^ a); (O a) (a ^ a); (a 1= a) (a 0;0 a) where informational operator means 'means'. Obviously, {a ü) (0 a); (O a) ^ {a o) Metaphysicalism is a specific circular form of informing. □ or also, {a O, O a) I. (a o(i-i- 0 =})) f^a a; \\ a ^ a; \a:=ia JJ □ Spontaneous, intentional, and alternative circularity is either metaphysical (that is, entity-interior) or entity-exterior (concerning entities outside the entity in question). 7.2.3 A Structure of Intentionality Intentionality of something belongs to the most basic virtues of physical things, informational entities, social processes, and other informationaUy structured phenomena. In intention, there is besides linguistically known meanings always something spontaneous, cyclic, and alternative. Intentionality proceeds in the direction of, for example, a goal, belief, physical structure, organization, and process, keeping as it seems, the validity of the so-called physical laws as the most general images of minds. Thus, let us set the following frame consequence. Consequence 8 (An implication concerning intentionality) Intentionality of something implies its spontaneity, circularity, and alternativity in its fragmental and composite form. Evidently, one can set (a or also. \ {a o(o o =})) (aw) of operator-altemative-operator conventions are introduced: o(o o □ 7.2.4 A Structure of Alternativity When an informational entity informs, it informs alternatively to some extent, Witliin a kind of informing there is possible to observe the so-called counterinforming. Alternativity (alternative choice of possibilities) belongs to the phenomenalism of counterinforming. So, to a kind of informing there may exist an alternative informing. Alternativity means aJternativeness, alternation, opposition, otherness, pluralism, parallelism, severalty, succession, etc, as a unity. To each informational operator there is possible to determine its alternative type. One of the most general informational alternativity is that which concerns an entity informing. Definition 10 (Alternative informing) Let us mark by =j the alternative operator to the general operator |=. IVe have the following reading convention: a ß reads a informs ß /? =1 a reads ß is informed by a □ The alternativity of the case is in the difference between 'to inform' and 'to be informed', where objects which inform and which are informed may have different functions. This difference lies in an entity's externalism and internalism. In the first case, a first entity informs something. In the second case, a second entity is informed by something. We can agree that, e.g. intention of the first entity might be different of the intention of the second entity. In this sense, to each informational operator its informational alternate can be introduced. The following convention seems to be appropriate. Definition 11 (Alternative operators) We can define a set of alternative operators corresponding to the original ones. The following pairs h, H informs [alternatively] does not [alternatively] inform IN, HI informs [alternatively] in parallel does not [alternatively] inform in parallel h, H informs cyclically [alternatively] ^ does not cyclically inform [alternatively] Ih, HI informs [alternatively] parallel-cyclically 7^1 does not parallel-cyclically inform [alternatively] infoi -nis $pon t an eo us-circ ularly [alternatively] informs spontaneously [alternatively] 0, o informs circularly [alternatively] informs intentionally [alternatively] =t, □ informs alternatively [alternatively] Several other alternative operators can be introduced. Any particularized operator has its alternate. For example, l^particularized and particularized are cases alternative to each other. The directionality of the original and alternative operators is evident. The original operators point from the left to the right, and the alternative from the right to the left. But, there is not only the directionality, because by alternative operators their semantics in comparison to the original ones is essentially changed. The original operators express the informingness of their left operand, while the alternative ones express the informedness of their left operands. The difference between the informing and informedness, that is, between an entity's externalism and internalism is substantial, Alternativity means the distinction which arises in an expressing and impressing process, when something emits information and something accepts it. Within the scope of the discussed circumstances concerning informational alternativity we can consider the next consequence. Consequence 9 (Implications concerning alternativity) Informational alternativity of an A.P. Železnikar entity is alternatively recursive entity, is an al-ternativeness of alternativeness, which causes the alternative operator of informing. There is, {a (ta); (S=a)=^ (a=t) By this consequence, the alternative informational operators are introduced, according to Definition 11. □ According to the previous discus si oa, there is possible to agree with a general consequence pertaining to alternativity in the following way. Consequence 10 (General implications concerning alternativity) For the alternativity and alternative alternativity there is a IN a Oi O; V\a ^ \ 11= o o (o OK-.))) HI a; t— a; 0 a; \V ^ a / (ar:|;((=||o^)oo)o^a) 7.3.1 Marking Decomposition Formulas In principle, decomposition proceeds from a topical theme, e.g., a marker, basic formula, or even formula system into greater details. On contrary, composition proceeds from particular detail themes and builds up a topical theme. Definition 12 (Markers of informational decomposition) For informational decomposition A of informational entity a, that is A(a) in general, we introduce the following particularized decompositional cases: There exist various other alternative consequences. □ 7.3 Serial, Parallel, Circular, and Metaphysical Modes of Informational Decomposition We shall now develop a systematic approach to the topical modes of informational decomposition which concern the informationaUy serial, parallel, circular, and metaphysical. For this purpose we introduce a standardized symbol A which is a mark for the formula of informational decomposition in general. Different superscripts and subscripts then determine the specific cases of the mentioned decompositions of an informational entity. As we shall recognize, informational decomposition concerns an entity, formula, or formula system as an entity (informational unity) in question. Aserial(a) AparaUel(a) Aserial-parallel(0:) Aparallel-serial(o) A^eriäl-parallel(^) parallel— ^.etaphysical^^j ^metaphysical / \ serial—parallel^ / A metaphysical f \ ^parallel-serialt^J informational decomposition A of entity a\ serial decomposition-, parallel decomposition-, serial-parallel decomposition; parallel-serial decomposition; circular serial decomposition; circular parallel decomposition; circular serial-parallel decomposition; circular serial-parallel decomposition; metaphysical serial decomposition; metaphysical parallel decomposition; metaphysical serial-parallel decomposition; metaphysical parallel-serial decomposition 7.3.2 Serial Decomposition Serial decomposition of an entity a has to proceed in consequent steps from one decomposed state to another. The serial has the meaning of the consecutive. In this way, there exists an ordered sequence of possibilities in which something can be decomposed in a serial way. By serial decomposition of an informational entity, the process of informing enters into informational details, that is, subentities concerning the entity in question. By decomposition, informational components of something come to the surface and the entity is being analyzed to the possible constitutional details. Thus, a decomposition procedure always concerns the chosen entity and its own components, which within the informational framework of the entity, arise to the informational existence. Consequence 11 (Serial decomposition of an entity) For a serial decomposition situation of entity a, marked by Aserial(a), where entities - ■ ■ ,u! are a's components, there is, Aserial(a) ^ \ Aparallel(a) ^ etc. Also other patterns (decompositional alternatives) of serial decomposition are possible.. □ 7.3.3 Parallel Decomposition A parallel decomposition of something anticipates a parallel set of autonomous and also information-ally (sequentially) connected formulas concerning the entity in question and its components. The needs for a parallel decomposition appear, for instance, in semantic and pragmatic analysis of an entity (e.g., headword, sentence, paragraph, text, etc.), where by parallel formulas an entity is additionally (in a detailed way) interpreted in various possible (also unforeseeable, counterinfor-mational) ways, according to the semantic and pragmatic linguistic conventions and innovations. Understanding of something represents a characteristic process of interpreting not only in parallel, but also in various serial, circular (tautological), and metaphysical ways. Parallelism is one of the semantically most powerful approaches of decomposition. Consequence 12 (Parallel decomposition of an entity) Let us have the following of one possible general structures fmeaningsj of the parallel decomposition of entity a: nßi \\Wi,£J2, ■ • - T'irla, / C a \ /a h A; ßi h P2-, Ol h <^2; Voti,,-! H vv UJI 1= ÜJ2-, h ^n^ } i) □ 7.3.4 Serial-parallel Decomposition Another logical situation is the so-caUed serial-parallel decomposition of something. In this case, within a serial decomposition some parallel blocks appear. One can express this possibility by the following consequence. Consequence 13 (Serial-parallel decomposition of an entity) For a serial-parallel decomposition situation of entity a, marked by Aserial-parallel(a), "^here entities ^par, 7par, ' " Xparj i>pxT> tJpar mark a's parallel structured components, that is. par 'ßA \ßij i 7par 72; X2; V'2; Xpar ^ : \XiJ LÜ, par A.P. Železnikar respectively, one of the serial-parallel decomposition schemes, marked by Aserial-paraUei(a)) "^«J/ have the following form; Aserial-paraUel(o;) /(/3par,7par,---,'^par C ö) fa 1= {ßp^ h (7par N (■ ■ • (^par N (((••• (ah /Spar) ■•■) h Xpar)N^i'par)l= Wpar // Another scheme, demonstrating the dominance of the parallel principle, the so-called parallel-serial decomposition, belongs to the category of the parallel type of decomposition. 7.3.5 Parallel-serial Decomposition As mentioned in the preceding subsection, another possible logical situation is the so-called parallel-serial decomposition of something. In this case, within a parallel decomposition various serial decompositions appear. One can express such a possibility by the following consequence. AparaUel-serial(ö) ^ ßltß2, - • ■ ißnß-lißnß, COL a [=(^1 h(•••(/?«,-! NM•'•)); a t= K 1= (■ • ■ h t^n«) ■ ■ •)); (a 1= ai) h N «nJ'-O; (a l=a;i) 1= ((■••(a 1= di)..-) h 0!na-l) N «Ti«; ((■■■(« N A)--ONjSn.-ONiSn,; V((- • • (a N Wi) ■ ■ ■) N 1= } This is only one of the possible cases of parallel-serial decomposition of entity a. □ 7.3.6 Circular Serial Decomposition Besides of the circular serial decomposition in this subsection, we shall discuss circular parallel, circular serial-parallel, and circular parallel-serial decompositions in the next subsections. Circular decomposition of something closes its decomposition circularly into something itself. From the view of the decomposed entity, circular decomposition is a kind of the self-reflexive analysis or also the top-down design, by which the informational components of the entity in question come to the formal surface, for instance, in the form of extended and new informational formulas and formula systems. Consequence 14 (Parallel-serial decomposition of an entity) For a parallel-serial decomposition situation of entity a, marked by Apara]lel-serial(™); there is, Consequence 15 (Serial circular decomposition of an entity) Serial circular decomposition of entity a follows the principle of serial decomposition with the condition that the system must be circularly (cyclically) closed in regard to a. According to serial decomposition in Consequence 11, the right-most operand in any serial formula, where the left-most operand is a, must be a too. Thus, Af^^ia) ^ etc. where marks a serial circular scheme ofa's decomposition. □ From the technical point of view, circular serial decomposition has the meaning of an internal insight into the successive functional structure of entity a. 7.3.7 Circular Parallel Decomposition Consequence 16 (Parallel circular decomposition of an entity) Parallel circular decomposition concerning entity a may perform indirectly (in an indirectly visible, 'subconscious', hidden, or unrevealed way). So, instead of a serial structured formula system in Consequence 12, there can be an arbitrarily mixed structure of parallel formulas such that a) ß^T, 0 1= cj; Vw 1= Q y .a natural picture of a thing's structure and organization. Pure serial and pure parallel situations and attitudes are a matter of an abstract or technological approaches within aitiiicial and theoretical systems. Circular serial-parallel decomposition of an entity is a decomposition or a set of decompositions closed circularly into the entity itself. In this way, circular serial-parallel decomposition can be obtained from Consequence 13 by setting the last operands in serial sequences by a. Consequence 17( Circular s er ial-parallel decomposition of an entity) For a circular serial-parallel decomposition situation of entity a, marked by A^^p^^^i(a), where entüies /?par,7par) ' ' ' »(^par mark parallel a's components in Consequence IS, there is, Aserial-parallel(^) / (/?par, Tpar, ' " " ,Wpar C a) \ ^(ah(^par|=(7parh(---(V'parN ^ Wpar) h «)•••); Wpar) h a)'"); w, par )l=a // where marks a circular parallel scheme of a's decomposition. □ Circular parallel decomposition of an entity has its evident paragons in different physical and artificial systems, and in social informational systems. 7.3.8 Circular Serial-parallel Decomposition A circular serial-parallel decompositional scheme (concept) of a phenomenon seems to be close to The circular serial-parallel decomposition can be characterized also as a parallel type of circular-serial decomposition when considering various informing possibilities of a. □ 7.3.9 Circular Parallel-serial Decomposition Another possible logical situation is the so-caJled circular parallel-serial decomposition of something. In this case, within a parallel decomposition circular serial decompositions appear. One can express this possibility by the following consequence. Consequence 18 (A Circular parallel—serial decomposition of an entity) For a circular parallel-serial decomposition situation of entity a, marked by A^-^iti-^eriail«). ^/^^re is, ^parallel—seriiilC'^) '' (((ai,a2,- ■ \ C a (a h'^ONO-'K^N ")•■•); VV((- • ■(« h Wi) • ■ •) h '^n^) 1= « This is only one of the possible cases of the circular parallel-serial decomposition of entity a. □ 7.3.10 Metaphysical Decomposition We can standardize a certain form of metaphysical decomposition of an informing entity a, saying that the informing process is a structure and organization of entity's informing, counterinforming, and informational embedding. This decomposition standard carries a logical background—which considers the entity's intentional spontaneous-circular informing—by the informing of which information arrives and arises spontaneously, is accepted (embedded, interpreted, connected) in a certain amount and way, and all this processing runs in one or another circular way regarding the entity in question. A standardized and pragmatical metaphysical scheme was already discussed in [Železnikar 93]. For entity a, its standardized (metaphysically universalized or generalized) components are informing la, counterinforming Ca, counterinfor-mation informational embedding Calpha, and embedding information £a. For these components, including a, metaphysical serial, parallel, serial-parallel, and parallel-serial decomposi- tion can be determined in various possible ways. These standardized components can be pragma-tized (metaphysically particularized) according to the entity inner structure and organization, and according to the outer informational impacts. In this way, metaphysical decomposition is meant to be the a's own way of inner informational arising, its autonomous informing in the realm of a~ influencing environment. Consequence 19 (Metaphysical decompositions by an entity's standardized components) Let us have to an entity belonging metaphysical entities marked in the following way: a for an informing entity^ Xa for informing of entity o, Co for counterinforming of entity a, 7q for a counterinforming entity within entity a, £a for informational embedding of entity a, Sa for an embeddingly informing entity of entity a Then some of possible metaphysical decomposition schemes are, systematically, the following; C o) (e. h «))))); h ")))); (((((a 1= I,) h C) h 7.) N^») h v 1= a // for the long-cycle metaphysical serial informing; fa \= T.; \\ la h C«; Co. N 7^; 7a N ^ai £a ^a'i \eoc h « // (Za, Co,, 7a) C o) for a metaphysical parallel informing of entity a; A metaphysical / \ _^ ^serial-paraUel*."^ ^ (er N (^r N "))))); {er h (^r h a))))- A metaphysical / \ ^parallel-seriaU^J \ V (((((a h xr ) h en h in N for the long-cycle metaphysical serial-parallel informing where, for example, xr^ ir rPar ta ! jspont^^ / ^spont\ ^■circu jinten jäter . /-»par ^ 1 •-'Or ' /'circu ^ inten palter y / / ^spont\ ^circu la finteli fa alter lot . rPar ^ 1 t-oi ■■ ^ circu ^inten ralter ) / c-circu ? pinten j-sklter '■a ({..■ia^lD^ll)-. VV^S' 1= « •ON 7»''; /or a metaphysical parallel-serial informing of entity a. □ 7.3.11 Marking Composition Formulas In principle, composition proceeds from given, also to the greatest detail particularized entities and builds np a topical entity, which represents a new operand F, coming into existence by a procedure of composition of sequentially informing components. Definition 13 (Markers of informational composition) For informational composition F of a set tp of informational entities cti, 02» an, that is r(Qi, ctj, ■ • ■, Q„), we introduce the following particularized compositional cases: In this system of parallel structured metaphysical components, spent, circu, inten, alter, etc, mark spontaneous, circular, intentional, alternative, etc. parts, respectively which can be shaped pragmatically according to the situation and attitude of entity a. The pragmatical means sufßcienÜy concretized, for instance, intelligent in a certain (intentional, understanding) way (informational-associa tive, counterinformational-creative, embedding-interpretative) (look at [Železnikar -93] where intelligence as an informational entity's metaphysicalism is discussed). Thus, lastly, rserial(^) rparallel(^i') rsenal-paraUd(^) rparallel— serial(^) rS-(^) I'seriä'^paraUelC^) ^pSÌid-serial(^) pmeUphysical^^^ pmetaphysical/ , -, •■parallel i^) informational composition r of entity setìp; serial composition] parallel composition-, serial-parallel composition] parallel-serial composition] circular serial composition] circular parallel composition] circular serial-parallel composition] circular serial-parallel composition] metaphysical serial composition] metaphysical parallel composition] 264 Informatica 17 (1993) 245-266 A.P. Železnikar rS-pSel(^) metaphysical serial-parallel composition; rnetaphysical parallel-serial composition 7.3.12 Serial Composition In a serial composition Fserial of sequentially acting entities (components) ai, a2, ■■■ , a«, the procedure of composition runs from the left to the right by the informing insertion of operands, operators, and parentheses or, in the reverse direction, when alternative cases are composed. A serial composition may consider both the interior and exterior components in respect to a given informational components ai, ö2) ' ■ ■ , c(„. If only some interior components are considered, the composition may follow a (partial) metaphysical concept, in which some components become cyclically informed. If exterior components are considered, the composed formula models a new exterior, that is, a non-metaphysical phenomenon in which component entities are involved, A serial composition has to proceed in consequent steps from one composed state to another. In this way, there exists an ordered sequence of possibilities in which something can be composed in a serial manner. Consequence 20 (Serial composition of an informational entity) For o serial composition, marked by rseriai(ai,ö2,■ • •,o;„), where entities Ol, 02) • • 'j öf^ components, there is, (Oi \= as) 1= (aa N (• • ■ ("n-i N O") ■ ■ •); 1= aa)---) 1= "n-2) N Q;„_I) ^ etc. Also other patterns (compositional alternatives) of serial composition are possible. Q 7.3.13 Parallel Composition We have to make a clear distinction between the processes of formula or formula system decomposition and composition. A decomposition is a top-down development (project, design) proceeding from a top idea (concept, basic structure) into more specific details in parallel, serial, and circular ways following some informational intention in such or another way. A composition means a combining of already existing entities (a kind of bottom-up development) into new entities, where, for instance, new relations, combinations, opera^ tions between entities are introduced. The introducing of new operand and operator entities is characteristic for both decomposition and composition, because it follows the so-called principle of informational spontaneity. Consequence 21 (Parallel composition of an informational entity) For a parallel composition, marked by rparaiiei(ai,o;2, • • • ,a:n), where entities ai, 02, • • -, a« components, there is, rparaUel(oi, «2, ' " " ) "n) ^ / (a. h oì; \ \i,je{l,2,---,n}J «2; In this consequence only some a i ^ a j may exist, so the composed system in not necessarily completely informationally connected. □ Other compositional (e.g. serial-parallel and parallel-serial) cases, analogous to the previously discussed decomposition cases, can easily be constructed by the reader according to the scheme of compositional markers in Definition 13. 7.3.14 Circular versus Metaphysical Composition While a common circular composition is a free case of circularly structured components, a metaphysical composition is actually a case of metaphysical decomposition. Certainly, the metaphysical composition can also consider the so-called internalization of exterior phenomena which impact the entity in question. But, this means nothing else than an internally occurring informational affair within the entity's metaphysicalism. Consequence 22 (Circular serial composition of an informational entity) For a circular serial composition, marked by «^i» • • ', ^n), where enti- ties Q,aij 02, • • -, ön are components, and entity a is the cycling operand, there is, etc. Also other patterns (compositional alternatives) of circular serial composition are possible. a Consequence 23 (Circular parallel composition of an informational entity) For a circular parallel composition, marked by where entities a,ai, «2? ■ ■ -, are components and entity a is the cycling operand, there is, faìp af, a{ 1= ay, aj \= a; Vi,j€ {1,2,' \ M J /a; W öi; «2; \an/ Totally new technologies are required to automatically acquire and store massive knowledge as efficiently as possible. (...) Knowledge is varied, diversified, and comes in many forms. —[Knowledge 92] 3-5 In this consequence only some sequences a |= a,-; ai |=. Oj; Oj 1= a may exist, so the composed system in not necessarily uniquely circularly closed. □ The reader can easily construct the consequences which pertain to other circular and metaphysical compositions, following the ideas of the adequate (discussed) cases of decomposition (symbol A). In this way, the study of possible decomposition and composition cases is coming to a reasonable end. 8 Conclusion A very large-scale information processing technology has two aspects: one is to enlarge the amount of knowledge and the other to enlarge the capacity to process knowledge. (...) Very large-scale knowledge does not simply mean collecting knowledge in substantial amounts. The problem of a new informational theory is how to organize its structure and symbolism in a way (e.g. counteiinformationally), that formal, discursive (mathematical, scientific), and semantic conventions remain preserved and that communication (informational embedding, interpreting) to existing theories is possible. The problem is how to keep the individual of the new theory and put it into the discourse of a new reality, for which it is believed that it will find a philosophical, scientific, and technological background (understanding). The question of a general informational theory is how to set the boundaries sufficiently wide, apart from conventional possibilities, enabling the informational fragmentation (decomposition) and, simultaneously, the synthe-sization (composition). In this view, a general informational theory is possibility-trivial, possessing an infinite set of consequences (theorems, lemmas, rules) which follow from basic axioms. For instance, we learned that symbol |= represents an extremely general (powerful) informational operator, which can replace any particular or universal operational concept and acts in an informational formula game (scenario) as an operator joker (trump-operator). Such a trivial situation of a general informational theory changes in case of a theory particular) zation where informational operands and operators become informationally closed (e.g. tau-tologically determined) entities. But, formal expressions (formulas, formula systems) carry their own stories, inform and counterinform surprisingly in regard to the traditional, that is, logical, mathematical, and linguistically abstract forms. There is much more informational background in that what is called informationally arising, coming to informational surface and much less of that what in traditional theories is reduced, neglected, unconscious, and simply eliminated as insignificant, The realm of the informational is meaningly broader, parallel, interconnected; in short, it is more complex, perplexed, interwoven, circular, and spontaneous; it corresponds better than the traditional philosophies do to that that approaches as evolution of man's abstract, circular, and spontaneous mind and what has to a reasonable extent been impacted by human informational tools where, for example, information processing of information processing became an evident and logically possible, manageable task. The concept of an informational machine implementation [Železnikar 92 c] remains one of the most challenging projects for the coming decades. It has to follow some essential changes in the design of both machine's hardware and software, where the informationally arising machine components are regularly supported by the machine system. It was shown, how understanding of text can become informationally arising in the framework of an informational environment [Železnikar 92d] (e.g., at the analysis and interpretation of the most lucid, comprehensive, and circularly perplexed philosophical texts and their translation from one natural language into another [Heidegger 86]). At the end, it seems evident that informational concepts in the form of machines and programs need an informational formalism which can guarantee sufficiently safe, error-free and efficient design. This experience is nothing new in the realm of computer design and use where mathematically rigorous techniques have been applied at the design of computer architecture, operating systems, structured programming and the like. In this view, mathematics seems to be only a specific mental technique which can assure the repeatability, efficiency, and safety (security) of computational processes (algorithms). So let us close with the following remarkable citation ([Husserl 80] I, p. 253) concerning also the phenomenon of construction of future informational machines and programs (written down at the very beginning of this century, in the year 1900): Hier ist zu beachten, daß der Mathematiker in Wahrheit nicht der reine Theoretiker ist, sondern nur der ingeniöse Techniker, gleichsam der Konstrukteur, welcher, in bloßem Hinblick auf die formalen Zusammenhänge, die Theorie urie ein technisches Kumtwerk aufbaut. So wie der praktische Mechaniker Maschinen konstruiert, ohne dazu letzte Einsicht in das Wesen der Natur überhaupt und in das Wesen ihrer sie bedingenden Begriffe und Gesetze besitzen zu müssen. Ahnlich verhält es sich ja bei allen „SpezialWissenschaften ". References [Heidegger 86] M. Heidegger: Sein und Zeit, Sechzehnte Auflage, Max Niemeyer Verlag, Tübingen, 1986. [Husserl 80] E. Husserl: Logische Untersuchungen, Erster Band (Prolegomena zur reiner Logik) (Sechste Auflage); Zweiter Band, I. Teil (Untersuchungen zur Phänomenologie und Theorie der Erkenntnis), (Sechste Auflage); Zweiter Band, IL Teil (Elemente einer phänomenologischen Aufklärung der Erkenntnis) (Fünfte Auflage); Max Niemeyer Verlag, Tübingen, 1980. [Penrose 90] R. Penrose: The Emperor's New MJnd, Vintage, London, 1990. [Železnikar 92b] A.P. Železnikar: Basic Informa-tionaJ Axioms, Informatica 16 (1992), No. 3, pp. 1-16. [Železnikar 92c] A.P. Železnikaj: An Introduction to in form azionai Machine, Informatica 16 (1992), No. 4, pp. 8-29. [Železnikar 92d] A.P. Železnikar: An Informa-tionai Approach of Being-there as Understanding (in three parts), Informatica 16 (1992), No.l, pp. 9-26; No.2, pp. 29-58; and No. 3, pp. 64-75. [Železnikar 93] A.P. Železnikar: Metaphysicaiism of Informing, Informatica 17 (1993), No. 1, pp. 65-80, [Knowledge 92] A Plan for the Knowledge Archives Project, The Economic Research Institute, Japan Society for the Promotion of Machine Industry, and Systems Research & Development Institute of Japan, Tokyo, March, 1992. THE INFORMATION TECHNOLOGY (IT) FOR ENABLING DISABLED PEOPLE - A STRATEGIC AGENDA - On-K wok Lai Lecturer in. Sociology, Department of Sociology The Chinese University of Hong Kong Shatin, N.T. Hong Kong Fax:(852) 858 7604, (852) 603 5213 Keywords: Social Welfare, Health Promotion, Disabled People, Informational Society, Family, Community Network, S.E.Asia Edited by: Matjaž Gams Received: June 10, 1993 Revised: July 7, 1993 Accepted: August 7, 1993 The agenda, and strategy building for enabling the disabled and needy people via Infor-niation Technology (IT) will be the niajor health and welfare issue in both developing and developed countries in the coming century. This paper, based upon some comparative findings and projected scenarios in East Asian countries, tries to examine the feasibility of and argue for the IT enianced enabling (disabled people) approach in health and welfare promotion. Part One of the paper examines, analytically and historically, the case of Asian experience in; (1 ) traditional forms of social response towards the disabled people, (2) modern organized form of social services for the disabled, and (3) the futuristic individually tailor-made services for the disabled via IT provision in different communities. In Part Two, findings and observations in the Asian context are outlined and discussed. Three major arena of possible integration between IT and disabled people are identified: within family system, in the functional (e.g. disabled) groups, and in the new form of community. Coupled with a discussion on the gender issues of the IT in health and welfare promotion, the respective specificity of the arena will be articulated. Also, the differential albeit inter-related rate and intensity of IT utilization in different arena will be examined. The related analyses, as argued in Fart Three, indicate the feasibility of the informational based disabled community networking, in and through which the disabled can be integrated into normal life. But, the effectiveness of this approach is, articulated in the later part, contingent upon the variables embedded in the technology -namely, risks- and social structure. The paper ends with a set of agenda and policy-related strategies for enabling the disabled in the 1990s and beyond within the global context. 1 Introduction of capital in general, IT in particular [21, 38]. But the rapid economic development in this region is , , . , , -n . ^ • necessary implying that socio- economic con- The Newly IndustnaLzed Economies (NIEs) m ^^^ ^^^^^^^ ^^^^^^ underprivileged and dis- East Asia have been taking up a more impor- ^^^^^ ^^^ j^jj^ supported by tant position m the global economy and hence, ^^^^ ^^^^^^^^ ^ economic development of these countries is highly j ^ i. i. t, j u • • r ^ , s ponds to health and welfare issues, as ageing of and structurally associated with the globalization population and increasing disabled people in public sphere, will become a global issue. The global development of IT ~ one of the three major futuristic technologies: bio-technology, and material-cum-energy technology - and its application in non-productive sector, namely in social arena in general, and welfare and health sector in particular wiU be the crucial factors in shaping future societal development in these NIEs. In other words, how to utilize IT in these arena will determine welfare benefits of the people. Hence, the agenda and strategy building for enabling the disabled and needy people via IT wiU be the major health and welfare issue in both developing and developed countries in the coming century. This paper, based upon some comparative findings and projected scenarios in East Asian countries, tries to examine the feasibility of and argue for the IT enhanced enabling (disabled people) approach in health and welfare promotion [40]. This paper starts with an introduction of the contextual specificity of the Asian (yet highly differentiated) experience in state's welfare and health provisions, then some major observations wiU be discussed. In later parts, strategies of the IT based disabled community networking wiU be discussed. The paper ends with some remarks on the IT based promotion strategy for welfare and health in the coming decade. 2 Coping with Health and Welfare in Asian Context Despite socio-cultural differences in the East Asian countries, family system and the spatially bound socio-cultural units -village or community-have historical contribution towards the (yet non-)promotioii of health and welfare for disabled people or those-in-need [13]. The functional units (family in general and women in particular) in these economies are in fact the only agency that supported the disabled in traditional societies. Those needy in these societies were treated socially (with or without welfare services) at the interface between the traditional and modern structures, and adjacent to the Eastern and Western cultural heritage of welfare ideas [9]. Most of them were regrettably treated as "spiritual deviance" or symbols of immoral acts of their ances- tors. In short, they were treated, in our present societal standards, less humane than it could. The actual effects were the marginalization of these disabled in community and they became the burden of the less weU functioned family system on the way to modernization. The modern and organized form of social services for the disabled was developed alongside (in many cases) with (Western) charitable and religious institutional welfare services set up. With the exception of some socialist states in the region (e.g. PR China), the degree of openness of respective countries indirectly reflects the level of organized welfare provision and to a certain extent determines the ways and forms of welfare every citizen entitled. In short, economic-political development within the global networking, in most cases, shapes how and when the state provided welfare in these economies [12, 41]. With structural change within the networking of the New International Divisions of Labour, the futuristic IT service for the disabled has opened up a new arena of health and welfare promotion. The new development in IT, particularly in medical and health sector, and its application in reha-bihtation wiU likely have a paradigmatic impact on societal views on those welfare-needy or disabled people. This is because of the fact that, despite its complex system, the very nature of (people's immediate contacts with) IT: its simplicity, flexibility, mobility and multi-modal accessibility and user-friendliness, could enable if not inflate the expectation of people on the possible performance of those needy groups and also fulfilling the notion of excellence of helping professionals. 3 The Differential Impact of IT on Asian Societies The impact of IT is global and substantial, yet unequal and differential [8]! In the coming sections, we shall outline the major arena of possible integration between IT and disabled people:-within family system, in functional (e.g. disabled) groups, and in the (new form of) community, coupled with a discussion on the gender issue of IT in health and welfare promotion. The differential development course in Asian developing economies within the globally organized divisions of labour has uneven impact on the form of state and society relationship [33]. This differential albeit strong development of IT in production sector is quite obvious, particularly in shaping the positive strategic position of state policy on IT [11, 23]. But the applications of technology have done so on a vefy narrow base [6], particularly, IT utilization in health and welfare related sector is quite limited if not total neglected. Yet, studies and advocacy on this issue are seldom attempted. One of the major observations in these economies is the fact that many (both software and hardware) piracies happened at the beginning phase of integrating computer technology and IT in their production systems. But this piracy issue is not the major one as IT industry matures, particularly when the industry realizes and provides certain samples of both hard- and software gratis to non- profit making public bodies [34]. The related industry's (counter-)Teactions on the piracy issue thus have a strong implication for and in shaping the filtering or diffusion rate of IT in family, group, and community, upon which no existing state policy has acted. Beyond most of the academic discourses on them, most of the Asian NIEs have a very strong state intervention in the economies and hence, a crucial role played by the developing states in high technology related production, vis-a-vis social welfare provision that is considered negatively as collective consumption burden of the states [1, 17]. Within the hegemonic paradigm for (solely economic) development, the underprivileged groups (women and disabled in particular) have to depend much upon the existing support from their families or (to a less extent) social groups. In other words, they are less visible to be seen nor playing any significant role on the political platform of the "economic miracle" of these NIEs. In short, these groups of people in different (in home and factory) settings have continuously and historically absorbed the burden of economic growth [26,32]. For them, IT is only the mechanism for production rather than consumption, economic growth rather than social development. 4 IT for the Promotion of Health and Welfare In this Part, IT and more important, the related information (necessarily, the value and moralladen discourse) on health and welfare are considered as crucial elements in determining the level of socio-cultural development in these economies [7], and the feasibility of the informational based disabled community networking, in and through which, the disabled can be integrated into normal life will be examined. The outlined scenario that the NIEs' selective development in economic and technology, juxtaposing their underdevelopment in technology-for-needy, prompts us to consider the development strategy in and through which a truly equitable welfare can be enhanced via the Pro-Active involvement of IT, This development should be viewed within a wider context' that more (even some poor) economies are developing their IT and the possible involvement of technology aids in some Asian countries [14]. The following delineation is the core of our thesis. First, the disabled and underprivileged are in most cases less mobile than their counterparts thus they merely have to anchor in the community they live or work. The community based nature (and constraint) of these groups also limits their access to information and contacts with outside world. At this juncture, IT can enable them to live and work in their limited place (domestic setting) yet can have similar if not equal and equitable life chance, as the opportunity structure available is dependent upon information one does have. Second, gifted by the TOUCH (-screen, -plate, or -tablet) and (remote) SENSING technological set up, IT can and will enable the maximum level of communication of the disabled on the FLOW of information - networking, yet this might produce their omnipotent impacts on others, society at large! In other words, IT and its products, if effectively used by and accessible to the disabled people will likely shift our World-View on their (strong and potential) performance and thus considered them as beneficial to the community at large! For instance, the Handicapped Digest, a moderated bulletin board for and with the disabled accessible internationally via: "l-hcapQndsuvml.Bitnet", Internet, Fidonet and Compuserv has been serving multi-functional and multi-dimensional activities, say, policy advocacy, informational exchanges, and consultation. Last yet the most important one, the very nature of IT is a form of public good because once its hard and software are developed, if they are accessible, there is no cost to "make" more copy or duplication [2]. The implication of this nature of IT is very obvious that people in general and those-in-needs in particular can (and should be encouraged to) have equal access to this omnipotent means in enhancing theii individual as well as collective welfare. At this juncture, the state has an important role to play in Uiis process. Given these conditional observations, we outline the consideration for developing a futuristic brief for the IT based strategic plan for the promotion of health and welfare, particularly, for the disabled in the coming decade in Asian context. Considering the fact that state policies on computer technology might not be the most important variable in facilitating the diffusion of IT and its utilization for the benefits of society at large [23]: "such policies are more effective in promoting production than use of computers", we should consider the development of IT within the context of social development, namely how best IT can promote welfare of the people. Hence, the major consideration is that the diffusion of IT and its beneficial effects on society at large should be measured at qualitative terms. For instance, state intervention wiU provide the likely involvement of IT for those in normal schooling and perhaps, this provision is "necessarily" allocated and is extended to those underprivileged vocational training and rehabilitation programs. However, no single policy can cope with the diversified and complex society we have today. Therefore, it should be pointed out that, for both the disabled and underprivileged groups, the internal differentiation (i.e. individual differences) among themselves, and distinction between them and the normal population are very significant; and this particular condition might lessen the extent of a single, standardized, communicable and compatible network for them. In other words, rehabilitation or intervention programs should be tailor- made within a comprehensive state-wide framework and planning yet kept them compatible and communicable with each others [27]. The basic principle in developing IT in health and welfare promotion of these countries is soliciting the necessary socio-political supports, namely, any positive policy change in IT should be in accordance with Social Justice principle [36] through which the disabled or underprivileged groups can be directed towards a sustainable mechanism for further empowerment of themselves [28]. For further development of IT infrastructure and utilization in social sphere, efforts should be focused on developing IT beyond production and profit principle, particulary for family and social groups, in terms of technology and professional aids, and the building up of IT network in and through which the underprivileged groups can have equal and equitable access to new information and technology. More specific, the following arena upon which IT based health and welfare promotion strategies could be focused. Macro Level: Legal and political infrastructure should be established in such way that disabled people can participate and have empowerment in the process of rehabilitation back to society and enjoy similar (if not the same) level of rights to access and the utilization of IT. In short, the rights to have (or access to) IT should be incorporated into the social rights of citizens. Supportive aids foUowing justice and equity principle should be provided when and where appropriate. Here, it should be pointed out that some yet different forms of social security for the disabled have been provided in aU Asian NIEs, therefore, the IT rights should be considered as an extension of the existing provision in social rights. By this basic and necessary provision in legal and political infrastructure, the governmental intervention -in school system, medical and health services, rehabilitation and positive discriminatory programs- would be influential in empowering the disabled and underprivileged groups beyond welfare and health arena. In short, the IT based empowerment is for all citizens in general and those needy in particular. Only by enacting legislations that promote health and welfare of people via the IT rights, it is possible to develop a sustainable society. Meso Level: The basic principle for IT aids should go with citizens participation. For the first step, promotion of disabled people's health and welfare is to let them gain access to technology (and IT) with the support of facilitators, namely significant others in their daily milieu. The state's enabling, via cash payment, the provision of facilitators or IT hard and software, could be developed here. The involvement of IT in health and welfare arena should, firstly, take the gender-specificity into account as more and more women involving in the production and consumption processing of these services, this specificity is very much due to the biological (say, longer life expectancy of female) and soci- cultural (e.g. male tend to have a higher non-response toward weLfaie provision) structuration of WOMEN within the health and welfare market [5]. Hence, women are the main producers and consumers in the health and welfare market. Home and other social institutions, say, school or work unit, and their cultural context will shape the possible forms of promoting welfare for disabled people in future, and in fact, the related socio-cultural structure wiU enhance (or limit) the extent of technological diffusion in these arenas [17], More important, for strategic development, we should build up a supportive socio-cultural environment with IT for this change in future. For illustration, three major social arenas that are significant for the strategic promotion of health and welfare will be briefly discussed here:-(1) Family, (2) School, and (3) Work Unit. Family System: Diffusion of IT is, as other innovations do, at the least in home or family setting, therefore much work has to be done in future. As most of the Asian societies have a comparatively strong family network or at least the family system did, do and wiU play a significant role in socialization, it is logically to suggest that more IT (supported by the states) should be available for families. The basic condition of enabling families having access to IT regardless of class and race should be cultivated in social policy, bearing in mind that IT industry has experience in providing hard-and-software gratis to non-profit making bodies. The extension of IT beyond production sector (and interestingly back) to family and household domain seemingly is the coming major phenomenal trend juxtaposing the history of telecommunication. The strategy to increase the utilization of IT in household domain and at the community level via state intervention (by financing or other state policies), will enhance not just the living quality of people, but also strengthening IT consumption and (re)production in the Asia-Pacific Rim. This will eventually enhance the sustain-ability of the region. The global trend in socialization process of the next generation is quite clear that this process in future will be developed within a set of technological variables, mass media (as news and information provider) and the new consumption patterns via IT (e.g., music TV, video-text and games) within home setting. Different messages, life portraits and virtual realities via IT might challenge the very ascriptive nature of family system. Enabling families or households to have a positive and developmental access to and acquisition of IT within an appropriate socio-cultural framework therefore is becoming one of the rhajor tasks of the state for social development. One precaution must be noted here that, with the increasingly important role of IT in family system, more of the traditional assigned tasks and control of, the family on younger generation and the gender specificity between male and femzde, might become a dynamic and flowing rather than static condition. For instance, the logic of control via fiow, vis-a-vis hierarchial form - say, the quest for women power, of family governance might be the paradoxical phenomenon of, the increase of IT led knowledge base of family cycling process, on the one hand, and the loosening of parental (vs. childhood) control, on the other. 4.1 School System For schooling, "IT can make education information dissemination less costly and more extensive and efficient so that the opportunity for self-enlightenment is easily accessible to a much larger population than otherwise" [3]. More importantly, given the new development in IT in Asian context, particularly the radio and (satellite) television-led education can be less dependent on English- dominated cultural influence and, thus it is very crucial for the national and cultural integration of diifeient communities on the way to and for their development in the global system. One of the related development in school via IT is the vocational training that is positioned between the production and schooling. Vocational training in these developing economies will likely be the domain where more disabled and needy to be involved in the process of empowerment of themselves on the one hand, and the further advancement of quality of health and welfare, on the other. Paradoxically, the existing IT involved education has a strong bias towards the creation of knowledge infrastructure for the reproduction of profit or monetary value but not the welfare of people in general and those-in-need in particular [15, 35], To reverse this imbalance development, the state or supra-national organization, say, UNDP, UNICEF, has the obligation to counter-act this developmental trend for people's welfare (to have equitable distribution of collective resources) and the prevention of so cio-spatial segregation. One of the major challenge for school system in the process of enhancing the ability of students (including socially, physically and mentally deprived groups) is the dissemination of "expert" knowledge among laymen. That might increase the vulnerability of schooling in exposing itself towards a wider global community which engenders different sets of value and moral basis of knowledge, and the direct consequence would be loosening of schooling control on people. With this respect, a value-laden and morally charged guideline for utilizing IT in schooling is envisaged. 4.2 Work Unit: Since "Work" has a significant role to play in the empowerment process of the disabled, different albeit inter-related strategies in re-integrating them back to work environment with successful experience should be advocated [20]. This is particularly the case when IT-led labour empowerment can also reinforce and strengthen the development of an IT community. Hence, the increase productivity of the disabled within (a competitive) working environment via IT in general and the empowerment of workers' bargaining power in particular can be achieved quite readily [31]. Assistive Technology (defined as devise or know-how that directly assists an individual with a disability to have the highest performance as s/he can) for the increase of productivity of disabled people is the basic step for possible involvement or rehabilitation of them in daily living [19]. It should be note that not just the targeted IT but also the ACCESS (system) to IT should be provided in the process of integrating the disabled in normal daily working environment. For the integration of disabled and special needs people in work environment, the notion of individuals difference in the Sense of Control should be noted [22]. The related practical implication is that the extent of having the Sense of Control with similar assistive devices technology in work may vary among individuals. To cope with individual's (dis)Ability and strength is becoming one of the challenge for producers and coworkers. 4.3 Social Groups: Interaction in groups will be the major arena in which disabled people experience (both positive and negative) societal encounters. The end-products of IT have been useful in providing (perhaps, more than) the necessary interaction between technology-simulated environment and the disabled, say video games and the simulated virtual reality! Further stimulation for their integration into community, on the one hand, and the experiencing of other people's feeling, on the other, wiU enable the disabled to have further and more access to external environment to the fullest extent. Nevertheless, the actual link-up between the simulated environment and reality is yet to be worked out. 4.4 In the Community: It is half rightly pointed out that, among others, the informational support provided in social environment is very crucial in helping disabled people to make the right and appropriate decision when facing with problems [10]. Hence, the further networking of and for the disabled locally and globally enabled by the ever increasing new IT can be envisaged [16]. Networking via IT for subsequent direct personal contacts in community can enhance the sense of belonging and neighbourhood. One of the major networking effects will be the enlargement of personal knowledge basis and the diversification (vis-a-vis traditional form) of people's outlook, attitude, and lifestyle within the global cultiire(s) - the world becomes smaller! However, this might be paradoxical when further networking, in some instance, prevents normal face-to-face encounters in (traditional) community setting. To what extent should state and society assume a control or guidance (private vs. public ownership, centralized vs. decentrahzed control) on the development of media networks is and will be a major socio-political issue [37], 4.5 Micro Level: The advantage of IT in welfare and health promotion is its possibility of developing an "individually adaptive and tailor-made IT set up". Helping professional aid to provide a tailor-made, flexible and adaptive IT for the disabled and underprivileged should be the goal of rehabilitation in the coming decade. In other words, the rehabilitation contribution of IT should be developed along with the psycho-social aspects of the clienteles in terms of task and situation specificity. The development of IT in health and welfare sector should thus have to consider (seemingly contradictory requirement for) not just cognitive but also emotional aspects of the involving parties. Hence, a balance and comprehensive approach in developing IT in the related sector should have the domains in cognitive and behavioral development as weU as emotional involvement. In other words, the basic issue (or debate) in the IT based empowerment strategies for people's welfare is: how to develop People's (vis-a-vis machine) potential? The problematique is that the reinforcement on and its related positive effects on individuals via IT application are very crucial in determining long-term and sustainable (health and welfare) development of a community. 5 IT in a Risk Society! It shoiild be reminded that IT and its applications are not fault-proof! That is, social benefits derived from IT's effectiveness are contingent upon the variables embedded in the technology -namely, risks- and social structure. The diffusion of IT in production is more readily in Asia which is paralleling the development of the informal sector (in both economic and IT piracy sense). As a result of the hyper-mobility of technology, the form of informal economy and the development of underclass in the process of informatization of technology reproduction is obvious. For instance, the formation of 'Dual City' in which development juxtaposed under-development is possible [8]. Deep anxiety and perplexity with technology are becoming the fact of life as more technological disasters (accidents) happened in the last decade, Chernobyl and Bhopal are some of them. At this juncture, it is appropriate to describe our present form of civilization as technology-cum-risks society [4, 25]. The characteristics of this Technology-Risk-Environment Syndrome are the invisibility and global nature of risks, and the techno-risks multiply themselves at a geometrical rate and exponential scale:- the synergetic effects of technological development, associated risks and others, as calculated, reacted, mediated and felt by the shocked public via IT based mass media. Hence, the IT based media has been and wiU continue facilitating the bombarding process of the techno-related accidents in front of global audience. One implicit meaning of this process is;-the poverty of science regarding the failure and subjectivity of the assumed scientific professional risk assessment. The societal effect enabled by IT and mass media is the De-coupling of assumed scientific rationality and societal mass yet differentiated subjectivity on development issues, green and environmental protests are part of the manifested De-coupHng process. The further development of IT for promoting health and welfare, in the Risk Society, needs to be considered within the framework of the IT associated risks and ramification, as the target groups of having IT are normally quite vulnerable and fragile, in the accidental occurrences. How individuals, groups and organizations respond to and contribute towards the potential risks incurred in IT application should be re-examined. Furthermore, users (broadly defined) participation in IT development, particularly its application in health and welfare sector, is thus to be strengthened in future. The formation of Users -cum- Producers Groups for sustainable development should therefore be encouraged. 6 IT led-Welfare Development in Next Modernity? The syndrome of "IT illiteracy" (lack of know-how and personnel, coupled with the sceptical views on "foreign technology") is the major obstacle in applying IT for health and welfare promotion, particularly when people consider the nonprofit making nature of the development in this sector. This syndrome can only be dealt with when IT is being at least recognized by the state or supra-national development bodies (say, World Bank or United Nations) in national development strategy for the enlightenment of the entire population regardless of race and class. In other words, governmental Pro-Active policy of utilizing IT in aU sectors is the only possible way to develop its sustain ability. Hence, IT rights within the framework of soci-political rights of citizens should be recognized. But the endorsement of the rights on (accessing) IT would not be fault-proof in the development course, as the utilization of IT is only effective and functional if and when this praxis is sustained by a new conceptualization of natural and social events [30, 38] and more important, a set of critical thinking constructs (regarding the self-reflectivity on risks, technology, and environment) developed in public sphere. In short, the new conceptualization on IT-led development should be based upon socio-ecological sustainability. The IT mediated global exposure, say, the virtual reality, wiU be part of social life in future, and this conditional structure of (post-)modernity will shape - opinions, views and attitudes- the World-View of people regarding resources distribution on health and welfare provision, sustainability at large. In futuristic terms, how people articulate the health and welfare interests, vis-avis ecology, in public sphere and its results wiU form the basic Hving conditions of those needy or (under-)privileged in future. "Prevention rather than Cure " is the slogan that symbolizes the modernization process of health and welfare related technology. Confronting the coming (post-)modernity, issue on risks in particular, the em- powerment of everyone seemingly is the principle target or course of actions that we should take! Unless precautionary actions and re- thinking are considered at this critical juncture, social uncertainties, risks and IT related accidents might be the message in the media! On the way to another (post-)modernity, seemingly, the logic of (non)development take many yet fragmented forms that are in line with the accumulation process of the undesirable but hegemonic course of technological advancement, and this "underside" is not the logistical or the outcome of purely technical errors [29]. For the sustainable development of nature and people, some sort of "technological-ecological ethics" should thus be developed to guide the IT development and application in future! References [1] R.P. Appelbaum and J.W. Henderson, (Edited), States and Development in the Asian Pacific Rim. Sage, London (1992). [2] K.J. Arrow, The Economics of Information. In The Computer Age: A Twenty Year View, edited M.L.Dertouzos and J.Moses, MIT Press, Cambridge, MA (1979). 3] Asian Productivity Organization [APO], Information Technology APO, Tokyo (1990). Led-Deveiopment. [4] U.Beck, Äfsk Society. Sage, London (1992). [5] T. Bengtsson and M. Johansson, Population Structure, Community Development and Employment in Post-Industrial Sweden. Conference Paper presented at the International Symposium and the Comprehensive Development of Community, 2-6 December, Haikou, Hainan, PR China (1992). [6] M,R. Bhagavan, Technological Advance in tie Third World. Zed Books, London (1990). [7] V.D.L. Carstairs and M.A, Heasman, "Information Strategy as a Basis for Measuring Health Promotion and Protection" in Measurement in Health Promotion and Protection, edited T.Abelin, ZJ.Brenzinski, and V.D.L, Carstaris, WHO Europe Regional Office, Copenhagen (1987). [8] M. Castells, (1989) The Informational City. BlackweU, Oxford (1989). [9] N. Chow, Western and Chinese Ideas of Social Welfare, International Social Work 30: 31-41, (1987). [10] S. Cohen, and T.A. Wills, Stress, Soda] Support, and the Buffering Hypothesis Psychological BuUetin 98: 310-357, (1985). [11] A. Cotterell, (Edited) Advanced Information Technology in the New Industrial Society. Oxford University Press, Oxford (1988). [12] J.E. Dixon, Socia/ Welfare in Asia. Croom Helm, London (1985). [13] D. Elliot, (edited). The World of Social Welfare. Charles .C.Thomas, Springfield, m.(1990). [14] Far Eastern Economic Review, T.March: 4566, (1991); and T.July: 43-68, (1988). [15] Far Eastern Economic Review, IT. M arch: 68,(1988). [16] T.A. Fathy, Telecity. Praeger Press, London (1991). [IT] J.C. Ferraz, H.Rush and L Miles, Development, Technology and Flexibility. Routledge, London (1992). [18] T. Forester, Computers in the Human Context. MIT Press, Cambridge, MA (1989). [19] Futurist, Technology for Disabled Workers: Assistive Technologies are being Promoted by Several Countries. 26: 56-58, (1992). [20] R. Greenwood, Apply Technology in the Work Environment President's Committee on Employment of People with Disabilities f Arkansas Jiesearcii and Training Centre in Vocational Rehabilitation, Washington (1990). [21] J.W. Henderson, The Globalisation of High Technology Production Routledge, London (1989). [22] D.R. Kauppi and C.M. Dzubak, Pychosocial Factors in Assistive Technology and Employment. Technology and Disability, 1(4): 23-28, (1992). [23] K.K. Kraemer, V. Gurbaxani, and J.L. King, Economic Development, Government Policy, and the Diffusion of Computing in Asia-Pacific Countries. Public Administration Review 52(2): 146-156, (1992). [24] M. Kulessa, The Newly Industrializing Economies of Asia Springier Verlag, BerHn (1990). [25] O-K. Lai, "Agenda and Strategy for Community Development in the Risk Society" in Collective WorA'S, by the Institute of Population Studies, Chinese Academy of Social Sciences, Beijing, (1993). [26] V. Lin, Health and Weifare and the Labour Process Journal of Contemporary Asia 16(45): 456-T4, (1986). [2T] W.C. Man, Sate-Wide Planning for Access to Technology Application for Individuals with Disabilities, Journal of Rehabilitation 5T: IT-20, (1991). [28] M. Margalit, Effective Technology Integration for Disabled Children Springier Verlag, Berlin (1990). [29] J.W. Murphy, A. Mickunas, and J.J. Piletta (edited) The Underside of High-Tech. Greenwood Press, New York (1986), [30] J.W. Murphy, and J.T. Pardeck, "Technology, Computerization, and the Conceptualization of Service Delivery" in The Impact of Information Technology on Social Work Practice, edited R.A. Cnaan, and P.Parsloe, Harworth Press, New York (1989). [31] New Scientist, 26.May: p.50, (1988), [32] C.H. Ng, Identity, Power and Economic Change - 25 Years of Family Studies in Hong Kong, Paper presented on the Conference: 25 Years of Social and Economic Development in Hong Kong, Centre of Asiaji Studies, The University of Hong Kong, 16-19 (Dec.) 1992. [33] T.J. Pempel, Of Dragon and Development in Journal of Public Policy 12: T9-92, (1992). [34] M.A. Perelman, Information. Sodai Relations and the Economics of High Technology. Macmillan, London (1991). [35] P.W. Preer, The Emergence of Technopolis. Praeger Press, London (1992). [36] J. Rawls, A Theory of Justice. Oxford University Press, Oxford (1971). [37] J. Takahashi, Main Issues Involved in IT-led Development in Information Technology Led-Development, edited Asian Productivity Organization, APO, Tokyo (1990). [38] M. Tehranian, Technologies of Power. Ablex Publishing Co, Norwood, NJ (1990). [39] Time, Feb. 22: 16-25, 1993. [40] United Nations, Disafaiiity Statistics Compendium. U.N,, Geneva (1990). [41] United Nations, A Look Back: United Nation Decade of Disabled (1983-1992). The United Nations Chronicle 29, p.76 (1992). A COMBINED ALGORITHM FOR TESTING IMPLICATIONS OF FUNCTIONAL AND MULTIVALUED DEPENDENCIES Mirko Maleković University of Zagreb, Faculty of Organìzation and Informatics Pavlinska 2, 42000 Varaždin, Croatia Keywords: algorithms, correctness, dependency theory, functional and multivalued dependencies, implication problem, informativeness, time complexity Edited by; Rudi Murn Received: April 1, 1993 Revised: September 6, 1993 Accepted: October 11, 1993 In this paper a combiaed algorithm for testing itn plication probiem, F ^ f, where F is a set of functional or multivalued dependencies, and f is a functional or multivalued dependency, is presented. The algorithm combines two well known algorithms. The first algorithm solves the implication problem F j= /, where F is a set of functional dependencies and f is a functional dependency (the algorithm is based on the closure of a set of attributes [Maier 83]). The other algorithm solves the problem F ^ f, where F is a set of functionai or multivalued dependencies and f is a functional or multivalued dependency (the algorithm is based on the dependency basis [Beeri 80j). The time compiexfty of the new aigoritiiin is the same as one of the aigorithm in [Beeri 80], In addition, tie new aigorithm is more informative than the algorithms in {Maier 83] and [Beeri 80] in so far as the new algorithm includes the result explanations REI and RE2 that indicate how the imJication probiem F j= / is solved; the aigorithms in [Maier 83] and [Beeri 80] produce only the answer 'Yes' or the answer 'No'. Also, REI and RE2 contain the proof of the correctness of the new aigorithm. 1 Introduction 80]). This paper consists of 5 sections. Section 2 con-One of the important issues in the design of tains the basic concepts of the dependency theory, relational database schemas is the specification In Section 3, we describe two well known algo-of the constraints that the data must satisfy rithms for solving the implication problem F j= f. to model corretly the part of the world under In Section 4, we give the combined algorithm for consideration^. Of particular importance are the solviag the implication on problem F j= f, where constraints, called functional and multivalued de- F is a set of functional or multivalued dependen-pendencies ([Gallaire et al. 81], [Maier 83], [Uli- cies and / is a functional or multivalued depen-man 88], and [Vaxdi 88]). dency. First we present the basic propositions The specification of dependencies is based on and describe the general idea of our algorithm, implication problem solving ([Beeri and Vardi 81], Upon that we present the algorithm and consider [Beeri and Vai di 84], [Fagin 82], [Honey man 82], its correctness and time complexity. Conclusions [Maier et al. 81], and [Vardi 83]). In this work, are discussed in Section 5. we give the combined algorithm for testing the implications of functional and multivalued dependencies. 2 Basic Concepts The algorithm combines two well known algorithms (the algorithms in [Maier 83] and [Beeri attributes are symbols taken trom a given timte (nonempty) set Ä = {Ai,.., style: Mirko Varga, Faculty of Organization and The set R is caUed the relation scheme, In what Informatics Varaždin, Croatia, Fax; +385 42 55 594. foUows we use the letters A, B ,C., D, E, F,G, H 278 Informatica 17 (1993) 277-283 M. Maieković (possibly with subscripts) to indicate single attributes, and U,V,W,X,Y,Z (possibly with subscripts) to indicate sets of attributes. The union of X and Y is denoted by XF. We usually do not distinguish between the attribute A and, the set {A}. Thus, A-i,..Ak denotes the set We assume that with each attribute Ai there is associated a (nonempty) set, called its domain, denoted by dom (A,). Let R = {Ai,..,A„} be a relational scheme, and i) = (J dom(A,) be a domain of R. A tuple on scheme Ä is a mapping t : R ^ D, such that t(A,) € dom(Ai) for all Ai £ R. t[Z] is a projection of a tuple t on a set of attributes Z, where Z C R. A relation on relational scheme iž is an ordered pair (J?,r), where r is a finite set of tuples on R. We denote tuples by the letters t^u^v (possibly with subscripts). Also, (5| is the number of elements in a set S. A set of all multivalued dependencies on R is denoted by MVD{R), that is, MVD{R) = {X y I C Ä} Moreover, FMVD(R) = FD(R) U MVD{R) is a set of all functional or multivalued dependencies on R. 2.3 Logical Consequence Let il be a relation scheme, F C FMVD{R), and / € FMVD{R). We say that / is a logical consequence of F, or logically implies /, written as (V(Ä,.))[(ii,r) Ih /], where 2.1 Functional Dependencies Let X,Y C iž be subsets of a relational scheme R. A functional dependency is an expression of the form X -^Y. X ^Y holds in {R^r), denoted if {R,r) Ih-X^y, (vfi,i2 e t)[t,[x] = t2{x] => ii[y] = t2[Y]] Let FD(R) = {X ^ y I X,y C Ä} be a set of all functional dependencies on R. 2.2 Multivalued Dependencies Let XyY C i? be subsets of a relational scheme R. A multivalued dependency is an expression of the form X Y. X -^-^Y holds in (R,r), denoted if (R,r) IhX^^y, (3^3 € 0[Ì3[Xy] = h[XY]A Let F,G C FMVD{R). We say that G is a logical consequence of i^, denoted F [= G, if (Vf? € G){F 1= 5]. ^ is the complement of |=. 2.4 Closure of an Attribute Set Let F C FD{R),X C R. The closure of X with respect to i^, denoted X^, is defined by X^ {A€ R\F\= X ^ A} Then, the following equivalence holds [Ullman 88]: F \= X -^Y ^ Y C X^ 2.5 The Dependency Basis Let F C FMVD(R),X C R. Next, let M be a set of multivalued dependencies, defined as foUows: M consists of (a) all multivalued dependencies in F, and (b) for each functional dependency X —► y in i^, the set of multivalued dependencies Xi Al, where Yi = Ai..Ak. The dependency basis of X with respect to Af, denoted DB{X, M), is defined by DB{X,M) = {Z C R\X\Z ^^ A M]^X Z N [{M X ^^ V ^V <1 Z hJJ ^ = Z]} The foUcwing properties of the dependency basis hold: (1) DB{X,M) is a partition oÌR\X. (2) If y C iž and X n y = 0, then X ^^Y ^Y is a union of sets in DB{X, M). (3) If A e Ä and then F\=X ^ A^ [A G DB(X, M)] A F)[AnXi = (iiA A € i;]] For details, see [UUman 88] and [Beeii 80], 3 Algorithms for Implication Problem Solving In this section, we describe two weU known algorithms for implication problem solving. 3.1 Algorithm (A) The Algorithm (A) tests F \= f, where F C FD{R) ajid / € FD{R), and is based on the closure of an attribute set (the algorithm for the closure computing is in [Maier 83]; its time complexity is 0{|F| • |Ä|) Algorithm (A) Input: F C FĐ(R), Method: (1) Computex^. (2) F h X y if YCXp, if YgXp Obviously, the time complexity of the Algorithm (A) is also 0(|-F| • |Ä|) (note that item (1) makes up the body of the Algorithm (A)). 3.2 Algorithm (B) The algorithm (B) tests F ^ f, where F C FMVD(R) and / £ FMVD{R). The algorithm is based on the algorithm for the dependency basis computing, [Beeri 80]. Algorithm (B) Input: F C FMVD(R), f :X ^Y or / : X y, where Jt,ycÄandXny = 0. Output: Yes if F\= f. No if t^ /. Method: (1) Compute the dependency basic DB{X,M). (2) If / : X ^ y, then apply the module FUN-DEP. (3) If / : X ' MULDER FUNDEP y, then apply the module f-.X^Ye FD{R). Output: Yes if ^ /. No if ^ /. Let y = Because F ^ X Bi-.B^ ^ F 1= X —> for all (the uuion rule and the decomposition rule for functional dependencies, [UUman 88]), we have to test F \= X ^ Bi for all Bi- Method: FUNDEP: F\^X^Bii{ and only if 1) BiE DB{X,M), and 2) (3X,- y,- 6 F)[{B,} n X, = 0 and Bi £ Y,]. MULDEP: F ^ X y if and only if 1) £ DB{X,M))[Y U i?,]. (1) and (2). 4.3 Proposition 3 Let F C FMVD{R) and / FMVD{R). Then X A e A (riglit- Proof Uii. The proofs of (1) and (2) are direct consequences The time complexity for the algorithm (B) is of the definition of |=, and (3) easily follows from 0(|M| • jÄp) [Beeri 80]. 4 The Combined Algorithm for Implication Problem Solving In this section, we present the combined algorithm for implication problem solving. The algorithm tests f H /, where F C FMVD{R) and / € FMVD{R), by a combination of the algorithms (A) and (B). It is based on the following facts. 4.1 Proposition 1 Let X,Y,ZC R, and F C FMVD{R). Then (i)F\=x^Y=>F\=x^z if zcy (fd-decomposition) ; (2) F\^X^YAF^X (fd-union); simplifi cation); (4) F \= X Y F X (mvd-simplification); and (5) X^Y\^X - Z ^ F\= X ^YZ {Y\X) (fd^ Y (translation). Proof The proof is straightforward. 4.2 Proposition 2 The relation |= between sets of dependencies has the following properties: (1) C FMVD{R))[F C C? G |= f] (reflexivity); (2) {VF,G,H C FMVDiR))[F G AG \= H ^ F \= H] (transitivity); and (3) (Vi^, G,HC FMVD{R))[F CGAF\=H G ^ H] (monotony). (VX,- ^ y,- e side). Proof Because (VX.- Yi e ^ >;-]], we see that the condition 2) in the module FUNDEP is not satisfied. Hence we have F X A. Now we describe the general idea of our algorithm. In solving the implication problem F \= f, where F C FMVD{R) and / : X ^ y, we first build the set G = {Xi Yi\Xi ^ Yi e F}. Upon that, we test G \=' f hy the Algorithm (A). If we obtain that G \= f holds, then we have, by the Proposition 2 (monotony), /. In the opposite case, that is, if G f, we examine the condition (C) {m e {Y\X){3Xi ^ y- € G)[Bi € y] If (C) does not hold, then, by the Proposition 3 (right-side), we obtain F X ^ Bj, for some Bj in (y\X). Hence, by the Proposition 1 (fd-decomposition), we have F X (y\X). Therefore, by the Proposition 1 (fd-simplification), F X —>-Y. If (C) holds, we apply the Algorithm (B), without the item 2) in the modul FUNDEP as the item 2) is already tested in (C), in solving |= X Bj, for all Bj in (y\X). For solving the implication problem F |= /, where F C FMVDiR) and / : X > y, we make use of the next strategy. In the first place, we test G |= X y (by the Algorithm (A)), where G = {Xi ^ y,-|X,- ^ ^ G i^}. If G h X ^ y, then, by the Proposition 1 (translation) and the Proposition 2 (transitivity), we obtain G |= X —> Y. Therefore, by the Proposition 2 (monotony), we have F |= /. In the case G It^ X y, we proceed with testing F \= X Y by Algorithm (B). Well, the modus operandi of our algorithm is simple: We always make an attempt at solving 1= / by the Algorithm (A). The Algorithm (B) will be used only if the Algorithm (A) fails. In addition, the new algorithm will include the result explanations REI and RE2 that have threefold role: (Rl) REI and RE2 contain a proof of the correctness of the algorithm. We suppose merely that the algorithms (A) and (B) are correct. (R2) REI and RE2 go to show how the implication problem F f is solved. From this we can see not only if f |= / or F / holds but also if G [= / or G / holds, where G is the set of all functional dependencies in F, (R3) REI and RE2 malce possible the implementation of the algorithm in such a manner that the users of the algorithm can learn about the properties of the relation |=. This is a consequence of (Rl) and (R2), For the new algorithm will have the properties (Rl) and (R2), and because the Algorithm (B) produces only the answer Yes or the answer No, we shall say that the new algorithm is more informative than the Algorithm (B). Now we present the algorithm that we sketched a little while ago. Algorithm (C) Input: F C FMVD{R), f € FMVD{R) Output: F f with a result explanation REI, or F y: f with a result explanation RE2. Method: (1) If / : X ^ y, then apply module FUNI, (2) If / : X y, then apply module MULI. FUNI The module FUNI tests F\=X ^Y. 1) Let G = {Xi ^ y^ G F} be a set of all functiond dependencies in F. 2) Test G 1= / (by the Algorithm (A)). 3) If G h /, then F\=f. REI: The functional dependency / : X —> y follows from the set of functional dependencies G, G C F, by the Algorithm (A). Therefore, we have, by the Proposition 2 (monotony), |= /• If G /, then [ìiG = F, then F /]. RE2: Because G / and G = f, we obtain If G / and G F, then the algorithm proceeds with testing. 4) Apply module FUN2. FUN2 The module FUN2 proceeds with testing F \= X ^Y. 1) Check the condition (C): i^Bi € {Y\X){3Xi ^ y, € G)[Bi e y], 2) If (C) does not hold, then f. RE2: Because the condition (C) does not hold, we have, by the Proposition 3 (right-side), F ^ X for some Bj in (y\Jf). Hence, by the Proposition 1 (fd-decomposition), we obtain F ^ X —> (y\X). Therefore, by the Proposition 1 (fd-simplification), we have F X ^ Y, If (C) holds, then apply the Algorithm (B) (the modul FUNDEP without the item 2)) in solving F ^ X for all Bj in (y\X). The Explanation: As the item 2) in the modul FUDEP is already tested in (C), we only have to test the Item 1) in the modul FUNDEP. 3) If F h ^ ^ -Bj. for all Bj in (y\X), then REI: Because F ^ X ^ Bj, for all Bj in (Y\X), we have, by the Proposition 1 (fd-union), F \= X ^ Hence, by the Proposition 1 (fd-simplification), we obtain F\=X -^Y. If F X ^ BkJox some Bk in {Y\X), then RE2: Because F X Bk, ioi some Bk in (y\X), we have, by the Proposition 1 (fd-decomposition), F X ^ There- fore, by the Proposition 1 (fd-simplifLcation), MULI The module MULI tests 1) Test F 1= X ^ y (by the module FUNI with- out the modul FUN2). lfF\=X^Y, then F\=X ^^Y, that is, F\=f. REI: Because F |= X ^ y, we have, by the Proposition 1 (translation) and the Proposition 2 (transitivity), F [= X Y, that is, F^f. KF\^X^YotG\^X-^Y> then apply the Algorithm (B) in solving F \= X (y\x). 2) lfF\=X^^ then F 1= X y. REI: Because F X (y\X), we have, by the Proposition 1 (mvd-simplification), F\=X ^^Y. liF\^X ^^ then F^X ^^Y, RE2: Because F X {Y\X), we have, by the Proposition 1 (mvd-simpMcation), F / 1= y. Proposition (correctness + complexity) The Algorithm (C) is correct and the time complexity for the Algorithm (C) is the same as the time complexity for the Algorithm (B), Proof The proof of the correctness of the Algorithm (C) is given in the result explanations REI and RE2 that are included in the Algorithm (C). Also, the proof of the correctness of the same algorithm is sketched in the description of the general idea of the algorithm. To see how the time complexity of the Algorithm (C) is the same as the time complexity of the Algorithm (B), observe that, during the execution of the Algorithm (C), we are always using the Algorithm (A), that is, the module FUNI, previous to the Algorithm (B). The Algorithm (B) win be used only if the Algorithm (A) (the modul FUNI) fails. For the time complexities of the Algorithm (A) and result explanations REI and RE2 are less than the time complexity of the Algorithm (B), we can conclude that the time complexity of the Algorithm (C) is the same as one of the Algorithm (B), that is, the time complexity is 0(jM|-|Äj3). 5 Conclusions We presented a combined algorithm for implication problem solving. The algorithm tests the implication problem, F f, where jF is a set of functional or multivalued dependencies and / is a functional or multivalued dependency. It combines two well known algorithms (A) and (B) that are based on the closure of the set of attributes and the dependency basis, respectively. The time complexity of the new algorithm is the same as the one of the Algorithm (B), (It is 0{\M\ ■ |Äp). The Algorithm (B) produces only the answer Yes or the answer No. The new algorithm is more informative than the Algorithm (B) in so far as the new algorithm contains the result explanations REI and RE2 which indicate how the implication problem F \= f is solved. Moreover, from REI and RE2 we can see not only if F |= / or f ^ / holds but also if G ^ / or G / holds, where G is the set of all functional dependencies in F. Acknowledgement I would like to thank the anonymous referees for their valuable comments and suggestions, References [Beeri 80] C. Beeri, On the membership problem for functional and multivalued dependencies, ACM Trans, on Database Systems 5:3 (1980), pp. 241-259. [Beeri and Vardi 81] C. Beeri and M.Y, Vardi, The implication problem for data dependencies, Automata, Languages and Programming (S. Even and 0. Kariv, eds), SpringerVerlag, New York (1981), pp. 73-85. [Beeii and Vaidi 84] C. Been and M.Y. Vardi, A proof procedure for data dependencies, J. ACM 31:4 (1984), pp. 718-741. [Fagin 82] R. Fagin, Horn clauses and database dependencies, J. ACM 29:4 (1982), pp. 952983. [Gallaire et al. 81] H. GaUaire, J. Minker, and J.M. Nicolas, Advances in Database Theory, Vol. I, Plenum Press, New York, (1981). [Honey man 82] P. Honey man, Testing satisfaction of functional dependencies, J. ACM 29:3 (1982), pp. 668-677. [Maier et al. 81] D. Maier, Y, Sagiv, and M. Yan-nakakis, On the complexity of testing implications of functional and join dependencies, J. ACM 28:4 (1981), pp. 680-695, [Maier 83] D. Maier, The theory of Relational Databases, Computer Science Press, Rockville, Md., (1988). [TJUman 88] J.D. UUman, Principles of Database and Knowledge Base Systems, Volume I, Computer Science Press, (1988). [Vardi 83] M.Y. Vardi, Inferring multivalued dependencies from functional and join dependencies, Acta Informatica 19:2 (1983), pp. 137-145. [Vardi 88] M.Y. Vardi, Fundamentals of dependency theory, in Trends in Theoretical Computer Science (E. Borger, ed.), Computer Science Press, Rockville, Md., (1988), pp. 171-224. ELECTRONIC NETWORKING FOR PHILOSOPHICAL DEVELOPMENT IN THE PRINCIPIA CYBERNETICA PROJECT Francis Heylighen..,. ' ■ PO, Free University of Brussels, Pleinlaan 2, B-1050 Brussels, Belgium fheyligh®vnet3.vub.ac.be AND Cliff Joslyn System Science, SUNY Binghampton, 327 Sring St. #2, Portland, ME 04102, U.S.A. cjoslyn0bingsim.s. cc .binhampton. edu Keywords: cybernetics and systems theory, evolutionary philosophy, computer networks, electronic mail, hypermedia, World-Wide Web Edited by: A. P. Železni kar Received: September 13, 1993 Revised: October 11, 1993 Accepted: October 15, 1993 The Principia Cyfaernetica Project (PCP) is a collaborative attempt to develop a compiete cybernetic and evolutionary philosophy. Such a philosophical system should arise from a transdiscipiinary unification and foundation of the domain of Systems Theory and Cybernetics. PCP is meta-cybernetical in that we intend to use cybernetic tools and methods to analyze and develop cybernetic theory, which include the computer-based tools of hypertext, electronic mail, electronic publishing, and knowledge structuring software. They are meant to support the process of collaborative theory-building by a variety of contributors, with different backgrounds and living in different parts of the world. 1 Principia Cybemetica in a meant to support the process of collaborative Nutshell theory-building by a variety of contributors, with different backgrounds and living in different parts m, n ■ • • /-1 t- i.' Ti • ^ ■ of the world. PCP will thus naturally develop in The Principia Cybernetica Project (PCP) is a , , . ,, , ^^ J. ^ J 1 li the cyberspace oi interhnlced electronic networks, collaborative attempt to develop a complete cy- . , , . i . i ttt i i -.tt. i , .. J T T.-1 -L c -L as implemented lor example in the World-Wide bernetic and evolutionary philosophy, ouch a ,, , , . , , . , ^ 1 .1 • r ^ Web distributed hypertext software, philosophical system should arise from a trans- disciplinary unification and foundation of the do- PCP is to be developed as a dynamic, multimain of Systems Theory and Cybernetics. Simi- dimensional conceptual network. The basic archi-lar to the metamathematical character of White- tecture consists of nodes, containing expositions head and Russell's Principia Mathematica, PCP and definitions of concepts, connected by links, is meta-cybernetical in that we intend to use cy- representing the associations that exist between bernetic tools and methods to analyze and de- the concepts. Both nodes and links can belong to velop cybernetic theory. different types, expressing different semantic and These include the computer-based tools of hy- Practical categories, pertext, electronic mail, electronic pubBshing, As its name implies, PCP will focus on the clar- and knowledge structuring software. They are ification of fundamental concepts and principles of the broadly defined domain of cybernetics and systems, which includes related disciplines such as the sciences of complexity, AI, ALife, Cognitive Science, Evolutionary Systems, etc. Concepts include: Complexity, Information, Entropy, System, Freedom, Control, S elf-organization, Emergence, etc. Principles are for example Natural Selection, the whole is more than the sum of its parts, and the Laws of Requisite Variety, of Requisite Hierarchy, and of Regulatory Models. The PCP philosophical system is to be seen as a clearly thought out and well-formulated, global world view, integrating the different domains of knowledge and experience. It should provide an answer to the basic questions: Who am I? Where do I come from? Where am I going to? The PCP philosophy is systemic and evolutionary, based on the spontaneous emergence of higher levels of orgeinization or control (metasystem transitions) through blind variation and natural selection. It includes: a) a metaphysics, based on processes or actions as ontologica! primitives, b) an epistemology, which understands knowl- edge as constructed by the subject or group, but undergoing selection by the environment; c) an ethics, with survival and the continuance of the process of evolution as supreme values. Philosophy and implementation of PCP are united by their common framework based on cybernetic and evolutionary principles: the computer-sup port system is intended to amplify the spontaneous development of knowledge which forms the main theme of the philosophy. PCP is managed by a board of editors (presently V. Turchin [CUNY, New York], C. Joslyn [NASA and SUNY Binghamton] and F. HeyBghen [Free University of Brussels]). Contributors are kept informed through the PRNCYB-L electronic mailing list, and the Principia Cy-bernetica Newsletter, distributed in print and by e-mail. Further activities of PCP are publications in journals or books, and the organization of meetings or symposia. More information about PCP is available by anonymous FTP at isl.vub.ac.ba, directory /pub/projects/Principia_Cybernetica, on World-Wide Web (http ;//pespmcl.vub.ac.be) or by an e-mail request to fheylighQvnetS.vub.ac.be. 2 Rationale PCP is about Philosophy. But what is philosophy? Philosophy intends to answer the questions: Who am I? Where do I come from? Where am I going? What is knowledge? What is truth? What is good? What is evil? What should my supreme goals in Bfe be, and what is the meaning of life? But there is a huge literature on philosophy. What is new here? Every time has its own approach to these eternal philosophical questions, deriving from its knowledge and technology. We hold that in our time, the age of information, it is systems science and cybernetics, as the general sciences of organization and communication, that can provide the basis for contemporary philosophy. Therefore, this philosophical system is derived from, and further develops, the basic principles of cybernetics. Moreover, we start from the thesis that systems at aU levels have been constructed by evolution, which we see as a continuing process of variation and natural selection of the fittest configuration. Evolution continuously adds complexity and makes systems more adaptive by giving them better control over their environments. Control in cybernetics implies the existence of choice from a variety of actions (freedom), and the competence or knowledge to choose the action most likely to bring the system closer to its goals (constraint). The complementary creation of freedom and constraint is realized when a new level of control emerges. We call this a metasystem transition, and consider it as the quantum of evolution. As cybernetic theory informs our philosophy, so cybernetic technology lets us do things that philosophers of other times could never do. Using computer technology, we develop a large philosophical text from many nodes which are linked together with different relationships. Readers can navigate among the many concepts, guided by their individual understanding and interests. Disparate material can be integrated together while being written and read by collaborators from aU around the world. We hold that PCP is more than an interesting experiment, and that there is an acute need fox an approach similar to PCP. The ongoing explosion and fragmentation of knowledge demands a renewed effort at integration. This has always been the dream of the systems theorists; all they lacked was the appropriate technology to attack the complexity of the task. The remainder of this text sketches some of the electronic tools we are presently using to support the project. The proposed philosophy, constituting the content of the project, and the conceived distributed hypermedia/e-mail implementation, constituting the form of the project, are in fact closely connected. Both are constructive, in the sense that they start from primitive systems from a variety of origins (nodes containing expositions written by diverse participants), which are brought into contact (e-mail conversations , and links to shared files), connected (semantic links), and selectively stabilized, so as to retain those combinations which define a new, more integrated system. When constructing a cybernetic philosophy the fundamental building blocks that we need are ideas: concepts and systems of concepts. Ideas, similarly to genes, undergo a variation-arid-selection type of evolution, characterized by mutations and recombinations of ideas, and by their spreading and selective reproduction or retention. The basic methodology for quickly developing a system as complex as a cybernetic philosophy would consist in supporting, directing and amplifying this natural development with the help of cybernetic technologies and methods. It will require, first, a large variety of concepts or ideas, provided by a variety of sources: different contributors to the project with different scientific and cultural backgrounds. These people need a fast, flexible communication tool: electronic mail. Second, we need a practical way for storing the thus gathered information, arvd for making it available all over the world: lileservers. Third, we need a system that allows the representation of different types of combinations or associations of concepts: hypermedia. Fourth, we need selection criteria, for creating combinations of concepts, that are partly internal to the system (e.g. consistency, simplicity ,,.), partly defined by the requirements of the people that are developing the system (consensus, ...). Finally, we will need procedures for reformulating the system of concepts (knowledge structuring, closure, ..,), building further on the newly selected recombinations, with the help of the concepts of emergence, and especially of metasystem transition. 3 The PRNCYB-L Electronic Mailing List We are using the PRNCYB-L mailing list on the BINGVMB computer system at SUNY-Binghamton in New York. PRNCYB-L is a LIST-SERV mailing list combining the functions of a message board and a file storage server. 3.1 Purpose PRNCYB-L provides an open forum for aU participants in the project, allowing interactive discussions about aJl issues related to PCP. It functions as the main medium for communication about the Project. At present, over 50 people, representing the 5 continents, subscribe to PRNCYB-L. Topics that have been discussed include: entropy increase and s elf-organization, causality as covariation, thermodynamics and evolution of mortality, memet-ics and the evolution of cooperation, formal expression, criteria for reality. A selection, of relevant information (e.g. congress announcements, publications; on hypertext, electronic publishing, evolution of the brain, ...) from other electronic forums is regularly cross-posted on PRNCYB-L. A digest, or summary, of the discussions and node development on PRNCYB-L will be produced, and published in the Newsletter. It is sent to aU people on our list of potential PCP contributors, including those who do not have access to electronic mail or who otherwise want only summary information about the project. The mailing list is not meant for idle chattering or uninformative technicalities. It is also not meant as a general forum for discussion about cybernetics and systems science. Such a forum already exists in the mailing list CYBSYS-L (contact Joslyn for more information). Whereas CYBSYS-L is open to anyone with an interest in cybernetics and systems, PRNCYB-L is restricted to active participants and those who wish to be informed about the specifics of PCP. 3.2 How does it work? The functioning of the list is very simple, and does not require any technical knowledge, If you are subscribed to PRNCYB-L, you will automatically and immediately receive aU messages sent to the list on your e-mail address. In this way the mailing list functions as a live Newsletter about the project, keeping you abreast of conceptual developments, as well as practical opportunities, such as workshops or publications. Once subscribed, if you wish to contribute yourself, you just send an electronic mail message to the list address, and it wiU be automatically broadcasted to all others who have subscribed. In that way you can mail out e.g. questions, proposals, or reactions. The third function of PRNCYB-L is the file server. Files containing larger pieces of information (e.g. PC? nodes, publications, bibliographies, logbooks of past discussions) are stored in a central computer. They can be requested by any subscriber by sending a simple e-mail command message to the file server address. 3.3 Contributing to PRNCYB-L Participants can best contribute to PCP by posting to PRNCYB-L. But for those who cannot use e-mail, we wiU accept contributions on a diskette file. MS-DOS diskettes (double density only) can be sent to Joslyn's address below; Macintosh, or MS-DOS (3.5", 720 Kb only) diskettes to Hey-Ughen, The formats we can accept are straight ASCII or text only (pieferied); TeX, LaTeX or Word Perfect (IBM only) [Joslyn]; Word, WordPerfect or other common word processors [Heylighen] . Those people who do not use e-mail as yet are encouraged to inquire about available facilities at their university or research center (these may not exist in Eastern Europe, or certain other regions). Although e-mail does require a (short) learning period, its use is in general quite simple once somebody has explained the conventions to you (these can be different for different computers and networks). It does not require any knowledge of computer science or programming. If you wish to join PRNCYB-L, please send an e-mail note to the address: cjoslynQbingsuns.cc.binghamton.adu containing the following information: 1) Name. 2) E-mail address. 3) Postal address. 4) Phone. 5) Affiliations. 6) How did you hear about PCP? 7) Do you wish to be a full subscriber to PRNCYB-L, or just receive the Digest? 8) Please take at least one page to describe your work and how it might relate to PCP. You wiU then be added to the mailing list, and receive initial instructions on how to operate the LISTSERV software. 4 The Principia Cybernetica FTP-server The Computing Center of the Free University of Brussels (VÜB) has put up an anonymous FTPfile server for PCP. FTP allows a much more interactive, universal and user-friendly way of retrieving files than the PRNCYB-L file server. The server is freely available for everybody on the Internet. In a further stage we hope to complement this FTP server by a Gopher server. To retrieve a PGP file by FTP from a Unix/Internet site, connect to the host by typing: ftp isl.vub.ac.be. When you are asked for your login, type: anonymous. Enter your e-mail address as password. Then change directories ('cd' command) by typing: cd /pub/projects/Principia,Cybernetica To show the available files and directories, type: la. You will get a list of subdirectories, including the following: /Misc.Info: contains diverse information (reports, bibliography, ...) on tools that may be useful for PCP, such as WWW. /News: contains the newsletters distributed about POP, and reports on PCP activities. /Nodes: contains preliminary concept definitions developed by the POP-editors, to be organized as a hypertext, semantic network. /Papers.Heylighen: contains papers on PCP themes by PCP-editor Francis Heylighen. /Papers.Joslyn: contains papers on PCP themes by PCP-editor Cliff Joslyn. /Papers _Turchin: contains papers on PCP themes by PCP-editor Valentin Turchin. /Papers,Others: contains papers on or related to PCP by other contributors. /PCP-Web: contains a mirror of our hypertextfiles for World-Wide Web (see next section). /PRNCYB-L: contains material (discussions, list of members) from the PRNCYB-L mailing list. /Software: contains the HyperVision application, a simple hypertext viewer and editor for MS-DOS, and browsers for WWW (see further), /Texts_General: contains introductory or overview papers by the editors collectively, and collected contributions of others, including the workbook of the PCP Workshop in Brussels, 1991. /WF-issue: contains draft papers contributed by the editors and others for a special issue of World Futures: the journal of general evolution devoted to the theory of Metasystem Transitions, which forms the core of the PCP philosophy. Most of the files available on this server are in ASCII format, and should be retrieved as text, The '.txt' file suffix denotes pure text files, that can be read as such. The '.tex' suffix denotes ASCII flies formatted in I^TgXwhich should ideally be processed by a EiTgXeditor in order to reconstruct formats, formulas and figures, but where the text is mostly readable as such. Again change to the directory you want and list the available files. You can then retrieve the file you want with 'get' followed by the filename, e.g.: get Workbook.txt. When you have the file(s) you want, type: quit. In case of doubt or difficulty, consult your system manager. Where the above procedures are not available (e.g. from Bitnet or other networks), there are two gateway fileservers: ftpmailQdecwrl.dec.com and bitftp®pucc.bitnet that will do the transfer for you. To one or the other of them, send the following one line message by e-mail: help for instructions (which wiU be similar to the above, but wiU be in the form of a series of lines in an e-mail message ftpmail or bitftp that will then execute for you). 5 Principia Cybernetica on World-Wide Web We have recently started up a new PCP server for distributed hypertext, in the World-Wide Web (WWW) format. WWW is an extension to, and integration of, other Internet services, such as telnet, newsgroups, FTP, gopher and WAIS, WWW combines extreme power (it does everything the other systems do and more), with maximal simplicity and ease of use. WWW allows you to fetch files ('documents'), containing hypertext links to other, related files, which may reside in different parts of the world. By selecting one of the links, you automatically fetch the linked files. In that way you can navigate through a world-wide network of interconnected documents, without having to type in any commands. WWW also offers multimedia support on the appropriate platforms: hypertexts may contain color images, sounds and even animations. WWW software is freely available for all major computer platforms, and only requires an Internet connection. More information about WWW can be found by anonymous FTP to info. cern. ch, (directory: /pub/www/doc for 'paper copies' of articles on WWW in PostScript and ASCII forms) or to the PCP FTP server (directory: . . ./Misc.Info). An even better introduction can be got by directly logging in to the Web, using telnet to one of the following hosts (in mainframe systems, the command is normally telnet followed by one of the following addresses or IP numbers): info.cern.ch (IP number 128,141.201.74): No password required. eies2.njit.edu (IP number 128.235.1.43): Log in as www. A full-screen browser. vms.huji.ac.il (IP number 128.139.4.3): Log in as www. kufacts.cc.ukans.edu: Login as www. ukanaix.cc.ukaiis.edu: Full screen browser, requires a VTIOO terminal. Log in as wffw. Free WWW-software ('browsers') can be found by anonymous FTP at the following places: ftp.ncsa.uiuc.edu, in directory /Mosaic or /Mac/Mosaic: Mosaic multimedia browser for X-Windows, Mac and Microsoft Windows. fatty, law, comell.edu, in direc- tory /pub/LII/cello: Browser for Microsoft Windows. info. cern. ch, in directory /pub/www/bin: Several browsers (Mac, NeXt, Dec...). Once you are connected to WWW, the PCP Web can be found on the geographical list of all WWW-servers under 'Europe: Belgium', or on the following WWW-address ('URL'): http ;//pespmcl.vub.ac.be/ In the near future, aU material collected by PCP will be made available on the Web in the form of linked nodes. The Web is to function as the main medium for practical and theoretical development of PCP, containing all definitions of concepts and principles, linked together by semantic relationships. Later, the Web will also provide the capacity for people reading nodes to make annotations with comments or criticisms, thus providing a computer support system for truly worldwide collaboration. 5.1 Retrieving WWW-files by e-mail People who are not directly connected to the Internet, yet can use e-mail (e.g. through Bitnet or CompuServe) can stiU get WWW-files by sending a message to the address: listservQinfo.cern.ch The message should consist of one or more lines, each containing the command SEND followed by the WWW-address (URL) of a desired document. E.g. for the PCP default home page a command line would read; SEMD http://pespmcl.vub.ac.be/ General info about WWW can be found at the following addresses: http ://www.vuw.ac.nz: SO/non-locsLl/gnat/www-f aq.html; http ;//info.cern.ch /hypert ext/WWW/Summary.html http ://pulua.hcc.hawaii.edu /guide/www.guide.html http : / / info. cern. ch /hypertext/WWW/TheProject.html The SEND command returns the hypertext document with the given W3 address, formatted to 72 character width (ASCII, text-only), with links numbered. A separate list at the end gives the document-addresses of the linked documents, which can then be requested by a subsequent message. In this way, you can navigate through the Web, albeit only at mail speed. As WWW links to FTP, this mail service can also be used to retrieve FTP-files but it is better to use the previously mentioned gateways for FTP, in order not to overburden the system. 6 PCP Publications Several texts on the PCP are now available, and more are being planned. 6.1 Workbook Heylighen F. (ed.) (1991): Workbook of the Principia Cybernetica Workshop (Principia Cy-bernetica, Brussels-New York). This booklet (70 pages) contains short articles and abstracts presented at the Workshop in Brussels. Presently it provides the most complete overview of PCP related work. It is available on the FTP-ser ver or in printed form from F. Heylighen for free. 6.2 Papers In addition to tke papers in the Workbook, the following papers directly address PGP, or its underlying philosophy. Most of these texts are available on the FTP-file server, where they can be automatically requested. Papers can also be requested from their respective authors. However, because of onr limited facilities and the virtually unlimited number of people interested in PCP, we cannot send aU papers to everybody. So we would appreciate if you would only ask hard copy reprints for papers that you really need and cannot get in any other way. Heylighen F. (1991)'. Cognitive Levels of Evolution: pre-rational to meta-rational, in: The Cybernetics of Complex Systems, F, Geyer (ed.), (Intersystems, Salinas, California), pp. 75-92. Heylighen F. (1992): Principles of Systems and Cybernetics, in: Cybernetics and Systems '92, R. Trappl (ed.), (World Science, Singapore), pp. 3-10. Heylighen F., Joslyn C. & Turchin V. (1991): A Short Introduction to the Principia Cyber-netica Project, Journal of Ideas 2, #1, pp. 26-29. Heylighen F. (1993): Selection Criteria for the Evolution of Knowledge, in; Proc. 13*'' Int. Congress on Cybernetics (Association Internat. de Cybernetique, Namur), p. 529. Joslyn C., Heylighen F. & Turchin V. (1993): Synopsys of the Principia Cybernet-ica Project, in: Proc. 13^'' Int. Congress on Cybernetics (Association Internationale de Cybernetique, Namur), p. 509. Joslyn C. (1991): Tools for the Development of Consensually-Based Philosophical Systems: a feasibility study for the Principia Cybernet-ica Project (Principia Cybernetica Technical Report). Joslyn, C. (1991): Control Theory and Cybernetic Ontology, (Principia Cybernetica Technical Report, long version of the paper on p. 24 of the Workbook). Lichtenstein B, (1991): A difference that makes a differance: cybernetic inquiry and postmodem philosophy, in: The Cybernetics of Complex Systems, F. Geyer (ed.), (Intersystems, Salinas, California), pp. 11-20. Moritz E. (1991): On the Road to Cybernetic Immortality: A Report on the First Principia Cybernetica Workshop, Journal of Ideas, 2, # 2/3. Turchin V, (1990): Cybernetics and Philosophy, in; The Cybernetics of Complex Systems, F. Geyer (ed.), (Intersystems, Salinas, California), pp. 61-74, Turchin V. (1993): On Cybernetic Epistemology, Systems Research 10:1, pp. 3-28. Turchin V. (1993): The Cybernetic Ontology of Actions, Kybernetes 22:2, pp. 10-30. Turchin V. and Joslyn C. (1990): The Cybernetic Manifesto, Kybernetes 19:2-3, pp. 63-65. In addition to those, the following papers, presented at a Symposium on the Principia Cybernetica Project, have been published in the Proceedings of the 13th Int. Congress of Cybernetics (Int. Association of Cybernetics, Namur, 1993). Introduction (Heylighen F.), p. 507. Jdanko A.V,: On Fundamental Problems of the Principia Cybernetica Project, p.514. Umerez J., Etxeberria A. & Moreno A.: Emergence and Functionality, p. 519. Glück R.: The Requirement of Identical Variety, p. 524. Elohim J.L.; Automation: a conscious human tool to rationally accomplish human aims in order to purposefully push ahead human evolution, p. 534. Carvallo M.E.: Some Alternatives to the Representational Mind, p. 539. Maddock J. W.: Modeling Human Relationships via Dialectical Ecology, p. 544. Conclusion (Heylighen F.), p.549. 6.3 Books A first volume, to be completed in 1993, will be an edited collection of papers by invited authors, including the PCP editors, William T. Powers, Elan Moritz, Robert Glück fe Andrei Klimov, Jon Timerez & Alvaro Moreno, Charles Francois, Len Troncale and Donald T. Campbell. This volume will appear as a special issue of World Futures: the Journal of General Evolution, published by Gordon and Breach, New York. Its theme is the Theory of Metasystem Transitions. A number of these ideas can already be found in a pre-PCP book; Turchin V. (1977): The Phenomenon of Science, Columbia University Press, New York. A second, more long-term project is to synthesize the different ideas that were developed separately, in the form of a real Principia Cybernetica monograph, authored by the PCP editorial board, and similar to the set of linked nodes existing on the PCP-Web. A provisional outline includes the topics below. It is similar to the structure of the hypertext network on our World-Wide Web server. Introduction PCP in context of intellectual history; PCP in systems and cybernetics. What are Cybernetics and Systems Science? Systems concepts; Cybernetic principles (including semantic and anticipatory control, and metasystem transitions). What is philosophy? Metaphysics; Epistemology; Ethics. Evolution History of evolution: life, multicellular organisms, movement, irritability, complex reflexes, associating, human intelligence; Future of evolution: imrriortality, superbe-ings, social integration. The Principia Cybernetica Philosophy Meta-foundationalism; PCP metaphysics; PCP epistemology; PCP ethics: ultimate human values, continuance of evolution, four types of immortality. System architecture Syntactic: hypermedia, markup standards; Semantic: link and node types. 7 Other PCP activities PCP regularly organizes conferences or meetings. Until now there have been: 1. A Symposium on Cybernetics and Human Values at the S"® World Congress of Systems and Cybernetics (New York, June 1990). 2. The Workshop of the Principia Cybernetica Project (Free University of Brussels, July 1991). 3. A Symposium on The Principia Cybernetica Project at the Int. Congress on Cybernetics (Namur, August 1992). 4. A Symposium on Cybernetic Principles of Knowledge Development (in collaboration with Stuart Umpleby) is planned for the 12"' European Meeting on Cybernetics and Systems Research (Vienna, April 1994). The meetings allow researchers potentially interested in contributing the Project to meet in a relaxed atmosphere. The emphasis is on discussion, rather than on formal presentation. Contributors are encouraged to read some of the available texts on the PCP in order to get acquainted with the main issues. PCP also publishes a Principia Cybernetica Newsletter, which is freely sent by postal or electronic mail to all people who ask to be on our mailing list. There have been two issues to date (0 and 1), The Newsletter appears irregularly, and summarizes the main developments (meetings, publications, theoretical developments, practical issues). It is edited by Francis Heylighen. 8 Collaborators Needed We are still looking for people to work with us in a variety of capacities, including contributors, reviewers, readers, and general source-people. We would also appreciate help with the administration: sending out mail, editing and printing newsletters and documents, connecting different communication channels (e.g. translating printed or faxed text to electronic texts). If you would dispose of secretarial or technical facilities, or have the time to help, please contact us. We are in particular looking for people with experience in hypermedia and computer-supported collaborative work environments, who might help us in choosing or developing the right tools. If you feel a strong resonance with Principia Cybernetica and the views we are expressing, we would also be very interested in talking about involvement at deeper levels. K you wish to contribute in any way, please contact one of the editors below. We would appreciate that in your note you would give a short overview of your current interests and how they relate to Principia Cybernetica, More specific proposals about how you might contribute would be helpful. Valentin Tur chin. Computer Science, City College of New York, New York NY 10031, USA. E-mail; turccQcunyvra.bitnet. 9 Editorial Board The Principia Cybernetica project is managed by a Board of Editors. The Board is responsible for the collection and development of the material, and for the implementation of the computer system. All inquiries or proposals about PCP should be directed to one of the editors below: Francis Heylighen, PO, Free University of Brussels, Pleinlaan 2, B-1050 Brussels, Belgium. Fax: +32-2-641 24 89. E-mail: fheylighSvnet3.vub.ac.be. CliflF Joslyn, Systems Science, SUNY Bingham-ton, 327 Spring St. # 2, Portland ME 04102, USA. Phone/Fax: 207/774-0029 (Fax after notice by phone). E-mail: cjoslynSbingsuns.cc.binghamton.edu or j oslyn®kong.gsf c.nasa.gov. TO THE INTERNAL REPRESENTATION DEBATE On Internal Representation (F. Heylighen) Paul Pangaro made the following commeiit on internal representation (CYBSYS-L electronic mailing list); Much of Maturarla and Varela's work (and by association, recent writings by Winograd, Mores) (most of which is supported by previous work by von Foerster) is about the differences of approach of modelling from the perspective of internal representation versus not, I know of nothing that prevents a system without internal representation of the environment from being a priori not intelligent, and rather like the implications that arise from not being dependent on such a thing to create intelligence. (My term for the relationship between organism and environment where this is the case is through looping; the mapping made by the organism is not external to internal but rather through; looping is required for recursion and stabilization. As much as I hate to once again open the debate on representation, I would like to attract people's attention to Conant and Ashby's famous result that every good regulator of a system must be a model of that system. They have proven that for the most general definition of control (including feedback as well as feedforward), control of a system requires that there must be a homo-morphism (one-to-one or many-to-one mapping) of the events in that system to the actions used to control it. As the most general meaning of model or representation is just that of a homomorphlc mapping, the only possible conclusion is that any agent, who has any form of control over its environment, must in some way have a representation of that environment. The debate remains open over whether that representation is internal or not. Pangaro's view of a mapping through rather than into would be compatible with Conant and Ashby's result, as they only require a mapping from environmental events to actions performed by the agent, not necessarily to cognitive structures inside the agent. But it seems difficult to imagine how such a mapping could be imple- mented in a stable way without some form of stable, intermediate structures (neural pathways, or whatever) connecting the agent's sensors to its effectors. And why wouldn't we want to caU such internal connections a representation, as they map specific classes of perceived events onto specific patterns of activation in the cognitive system? Of course, if representation is understood in the most restricted sense as a mapping of static entities (objects in the environment) to static symbols (cognitive elements) then the Conant and Ashby result does not in any way entail a representation, The only correspondence required is one between processes (events and actions), not entities. What Brooks, Maturana, Varela and others seem to have discovered is that you can have control without such static representations. However, they seem a little bit too eager in their fight with the old symbolic or representationalist school of AI, and tend to throw away the child with the bath-water, rejecting any use of the concept of representation even though their own models do require representation in the dynamic sense. Reference: Conant Roger C, and Ashby, W. Ross (1970): Every Good Reguhtor of a System Must Be Model ofthat System, Int. J. Systems Science 1:2, pp. 89-97. Address: Dr. Francis Heylighen, Systems Researcher, PO, Free University of Brussels, Pleinlaan 2, B -1050 Brussels, Belgium. Phone:+32-2-6412525; Fajt:+32-2-6412489; fheyligh®vnet3.vub.ac.be. On Informing between Entities (A.P. Železnikar) Informing between entities a (environment) and ß (observer) can also be understood as a problem of ;3's internal representation of a. In informational language, the phenomenon of transition, marked by a j= has to be informationaUy interpreted. An entity (das Seiende) is an information-ally arising phenomenon for which the so-called homomorphic mapping (ono-to-one or many-to- TO THE INTERNAL Informatica 17 (1993) 294-296 295 one, even in the most free manner) is in no way a sufficient interpretation. The question is: How is the image Lß of environment a made (informed, constructed, phenomenalized) by the environment observer ß? Let us present (interpret) the process (entity, transition, operand) a \= ß through several formal situations. The observing of a by /3 should be without any (informational or physical) impact on a. This is an idealistic (reductionistic) situation being quite convenient for our informational investigation. So, let In this expression, marks an informational function (lß), a formula system, which depends on (considers) a, and this function is being in (operator c) ß- Operator C denotes the so-called Informational Being-in (in-volvement) of something in something. Through function iß{a), entity a becomes involved in entity ß, however, as an entity belonging to ß. The question is what is reflected in (.^(o) in concern to a, since this reflection belongs to observer ß observing environment a. Thus, a ^ ß (in fact, ß a) in the upper formula seems not to be completely reliable supposition! But on the other side, if ß does not impact o, this supposition is righteous. Informational Being-in (inclusiveness, includ-edness) of operands (entities) ^ and rj, as an autonomous informational entity, can be deflned (axiomatized) recursively, for instance, as i^cv) -Def Jl? h f) C T?) hpossible; V) impossible j J(l7 1= 0 C 0 Hpoasibki ^(f h T?) C f) ^possible / conditions—outgrow to an intelligent entity. Let us demonstrate this phenomenon informationally to our initial example, According to the second formula, if we choose the first and the second possibility, that is, [ß ty3(ö:)) C ß and (4/3(0) \= ß) C ß, two serial loops come into existence, that is, iß\=Cßia))\=ß; ß h N ß) as a consequence of C ß) (ß h ^M«); \ {ß cßia)] C ß {iß{a)\=ß)cßj The particularized informational operator hpossible says that the formula informs to be possible, according to the choice of an agent (e.g., entity mairked by r), both of them, or an exterior setter of the formula system). Informational Being-in concerning an informational entity (informational operand) is a conditio sine qua non for the so-caUed circular se-rialism, which can—under certain metaphysical In the fi.rst circular formula, ß is the observer of the transition ß \= Lg(a)., while in the second circular formula the role of the observer of ß belongs to the transition |= ß. Similarly, one can say that in the first formula, transition ß [= ißiGt) is the informer of ß, while in the second circular formula ß is the main informer. In both cases, formulas are circular in respect to ß. Informational entities (agents) are dynamic phenomena (also systems, processes, events, actions) and their intelligence depends on their own and environmental spontaneity and circularity (e.g., on the spontaneous through—operator 1= between entities a and ß—instead of the spontaneous into together with informational looping). If function iß{a.) does include a kind of understanding as a property within ß, marked by understanding i//}(c(), then this understanding cyclically produces the so-called meaning /iw^(a), concerning environment (or, more precisely, an environmental phenomenon) a. We say that iß{a) within itself is decomposed (identified, interpreted) as an understanding entity in regard to a by (MW^(a) C Uß{a)) C fcj3(a) The whole story of spontaneous cy eli city concerning the observer ß and the observed environmental entity a now proceeds into the complex circular scheme, expressed by the formula a N ((K(«) C C ißia)) C ß) (1) The reader can imagine how this formula could be decomposed according to the definition of ^ C t? recursively and, for instance, pragmatically interpreted (additionally decomposed) into further informational detail. The difference between the object a and its internal representations in the form of spontaneous-cyclically connected entities tß(a), liß(a), and fin^ia) depends on the internal cyclic informing of ß or, precisely, on its meta-physicalism. That what informs within /? as a consequence of a's impact, and informationally concerns a, is certainly not the environmental entity a itself, but /3's imagery (physical, informational phenomenal-ity) of a in the form of, for example, ißia), ^^3(0:), and ßUß{Q.) within an intelligently structured entity ß. One of metaphysical shells of the observing entity ß consists of three standard parts called informing, counterinforming, and (informational) embedding. We have the following basic metaphysical (circular) serial scheme concerning an informational entity in general, and entity ß in particular: metaphysical informing of/3 as a whole N Iß) N £ß) informing counterinforming'---- embedding Other schemes can be constructed by all possible permutations of parenthesis pairs. In short, in this scheme, Ip, Cß, £ß, and Eß are called informing, counterinforming, counterinformation, embedding, and embedding information, respectively. This particular scheme of serial metaphysical shell is a consequence of a scenario of informational includedness being informed by an exterior entity a so, according to Formula (1), «h ^{Sßia) C C lß(a))cCß{a))cIß{a^Cß This metaphysical shell of ß recognizing a can now be filled intelligently, that is, the informing part of ß by specific components characterizing /0 in a narrower sense; the counterinforming part by understanding and by it produced concept of meaning; and the embedding part by further interpretation of meaning in informational connection with ß. This formal philosophy (theory) reveals the problem of internal representation in a complex, dynamic, and never definite way. Informational formulas are arising informational entities by themselves and in this manner essentially deviate from abstract mathematical notions (theories, systems, formulas). The other approach of internal representation of a by entity ß is metaphysically parallel and much more straightforward in respect to the traditional understanding of informational includedness. It proceeds from the basic parallel metaphysical shell of entity ß which is entity ß )/3's embedding ^^ ^ Jf^X ß^s counterinforming Iß C Cß } Decomposing this shell in respect to operator |=, the internal representation of entity a by (within) entity /3 is a consequence of the metaphysical system ah ^ß\=£ß{ay, ß\=Ißia)- \ Bßiitem{} First Author: {Ssl Title}, Magazine, Vol. 1, Io. 1. \end{thebiblj o graphy} \d8f\joarnal{Informatica {\bf 17} page xxx—yyy} \immediate\Hritel6{'Informatica' VI,2 by B,-'Z} \newif\iftitle \titlefalse \hoffset=limm \voffset;=amm \oddsìdemargìn=-21miin \evensìdemar6ìn=-14mm \tapmargln=-33mm \headheight=17inm \headsep=10mm \footheight=8.4min \footskip=52.Emm \teKtheight=242iiTO \textiiidth= 170mm \colttansep=5iiini \coluninseprule=0pt \tvocolumn \slappy Vflushbottom \parindent lem Meftmargini 2em MeftmarginMeftmargini Meftmarginv .Sem Meftmarginvi .Sem \def\labelitemi{\bf —} \def\labelitemii{—} \def\labelitemiil{~} Nsetcounter{3ecnunldepth}{3} \def\mBketitle{\tBocolumn[% \vbox{\hs ize=\t ex tu idth\Large\bf\raggedr ight \uppercase{\atitle}}\vss\blgskip\bigskip \vboi{ \hsize=\textBidth \0author}\bigskip\siiiallsltip \vbox{\hsiz6=\textBidth {\bf Keywords:} XOkeynords} \bigskip \hbox{{\bf Edited by;} \eedited} \smallskip \hbox{{\bf Received:} \hboi to lQein{ \areceived\hss} {\bf Revised:}\hbox to 10em{ \Srevised\hss} {\bf Accepted:}\hbDx to 10em{ \@accepted\hs$}} \bigskip \vbox{\hsize=\te*t«idth \l6ftBkip=3Gm \rightskip=3Biii \sl NSabstract} \b igsk ip\bigski p]\ti tletrue} \def\maketitleauthor{\tBocoluiiim[Ä \¥box{\hs ize=\t eitwidth \l.arge\bf\ raggedr ight \«title}\v3s \bigskipNbigskip \vbox{\h3Ìz«=\t extwidth \aauthor} \bisskip\bigskip] \gdef\9title{}\titletrue} \def\tnakeonlytitle{\tiiocolumn[y, \¥box{\hsize=\textijidth \Large\bf\raggedright \«tit1e}\vss\bigskip\bigskip] \gđef\Stitle{}\titletrue} \def\8title{} \def\eauthor{} \def\®titleH{> \def\aauthor«{} \def\®keyBard3{} \def\aedited{} \def\fflabstract{} \def\«received{} \def\«Tevised{> \d«f\eaccepted{} \def\authoreven#l{\gdef\ÖauthorH{»l}} \def\titleoddSl{\gdef\atitleH{\uppercase{#l}}} \d ef\keyword s»l{Ngdef\«kaywo rds{«l}} \def\edited#l{\gdef\8edited{#l}} \def\received#l{\gdef\8received{#l}} \d ef\rev isedt1{Vgdefrevised{il}} \def\accepted»l{\eđef\«aecepteđ{»l}} \long\def\abstract#l{\gdef\®abstract{#l}} \def\soction{\estartsection {section} {l}{\za}{-3,5ex plus -lex minus -.2ex} {2.3ex plus .2ex}{\Large\bfVraggedright}} \def\3ubsection{\®etartsection{subsection} {2}{\z«}{-3,25bi plus -lex minus -.2ex} {l.Sex plus .2e*}{\large\bf\r4ggedright}} \d«f\subsubsection{\(9startsection{subsubsection} {3}{\z9}{-3.2Sex plus -lex minus -.2ex} {l.Ea* plus .2ex}{\notmalsize\bf\raggedright}} \def\aevenhead{\hbox to 3eni{Sbf\thspage\hss} {\sinall\ j ournal \hf il \ if t it le\« 1 s e \8authorH \f1}VglobalVt it1efalg e} \def\8oddhead{{\small\iftitle\6lse \etitleH \fi \hfil\jottrnal}\hbox to 3eiii{\hss\bfVthepage} \global\titlefalse} \def\(t6venfoot{\hfil} \daf\®oddfoot{\hfil} \endinput \end{do cument} REVIEW REPORT Basic Instructions Informatica publishes scientific papers accepted by at least two referees outside the author's country. Each author should submit three copies of the manuscript with good copies of the figures and photographs to one of the editors from the Editorial Board or to the Contact Person. Editing and refereeing are distributed. Each editor can conduct the refereeing process by appointing two new referees or referees from the Board of Referees or Editorial Board. Referees should not be from the author's country. The names of the referees should not be revealed to the authors under any circumstances. The names of referees will appear in the Refereeing Board. Each paper bears the name of the editor who appointed the referees. It is highly recommended that each referee writes as many remarks as possible directly on the manuscript, ranging from typing errors to global philosophical disagreements. The chosen editor will send the author copies with remarks, and if accepted also to the Contact Person with the accompanying completed Review Reports. The Executive Board will inform the author that the paper is accepted, meaning that it will be published in less than one year after receiving original figures on separate sheets and the text on an IBM PC DOS floppy disk or through e-mail - both in ASCII and the Informatica LaTeX format. Style and examples of papers can be obtained through e-mail from the Contact Person. Opinions, news, calls for conferences, calls for papers, etc. should be sent directly to the Contact Person. Date Sent: Date to be Returned: Name and Country of Referee: Name of Editor: Title: Authors: Additional Remarks: All boxes should be filled with numbers 1-10 with 10 as the highest rated. The final mark (recommendation) consists of two orthogonal assessments: scientific quality and readability. The readability mark is based on the estimated perception of average reader with faculty education in computer science and informatics. It consists of four subfields, representing if the article is interesting for large audience (interesting), if its scope ajid approach is enough general (generality), and presentation and language. Therefore, very specific articles with high scientific quality should have approximately similar recommendation as general articles about scientific and educational viewpoints related to computer science and informatics. SCIENTIFIC QUALITY Originality Significance Relevance Soundness Presentation READABILITY Interesting Generality Presentation Language FINAL RECOMMENDATION Highly recommended Accept without changes Accept with minor changes Accept with major changes Author should prepare a major revision Reject INFORMATICA AN INTERNATIONAL JOURNAL OF COMPUTING AND INFORMATICS INVITATION, COOPERATION Since 1977, Informatica has been a major Slovenian scientific journal of computing and informatics, including telecommunications, automation and other related areas. In its 17th year, it is becoming truly international, although it still remains connected to Central Europe. The basic aim of Informatica is to impose intellectual values (science, engineering) in a distributed organisation. Informatica is a.journal primarily covering the European computer science and informatics community -scientific and educational as well as technical, commercial and industrial. Its basic aim is to enhance communications between different European structures on the basis of equal rights and international refereeing. It publishes scientific papers accepted by at least two referees outside the author's country. In addition, it contains information about conferences, opinions, critical examinations of existing publications and news. Finally, major practical achievements and innovations in the computer and information industry are presented through commercial publications as well as through independent evaluations. Editing and refereeing are distributed. Each editor can conduct the refereeing process by appointing two new referees or referees from the Board of Referees or Editorial Board. Referees should not be from the author's country. If new referees are appointed, their names will appear in the Refereeing Board. Informatica is free of charge for major scientific, educational and governmental institutions. Others should subscribe (institutional rate 50 DM, individual rate 25 DM, and for students 10 DM). Send a check for the appropriate amount to Slovene Society Informatica, Vožarski pot 12, Ljubljana, account no. 50101-67851841. Please, return this questionnaire. QUESTIONNAIRE Send Informatica free of charge Yes, we subscribe We intend to cooperate (describe): Proposals for improvements (describe): Informatica is published in one volume (4 issues) per year. If possible, send one copy of Informatica to the local library. Please, complete the order form and send it to Dr. Rudi Murn, Informatica, Institut Jožef Stefan, Jamova 39, 61111 Ljubljana, Slovenia. ORDER FORM ~ INFORMATICA Name: ............................................. Office Address and Telephone (optional): Title and Profession (optional): ... ;........................................................ E-mail Address (optional): ............. Home Address and Telephone (optional): ........... .................................................... Signature and Date: ................... EDITORIAL BOARDS, PUBLISHING COUNCIL Informatica is a journal primarily coveting the European computer science and informatics community; fscientific and educational as well as technical, commercial and industrial. Its basic aim is to enhance communications between different European structures on the basis of equal rights and international refereeing. It publishes scientific papers accepted by at least two referees outside the author's country. In addition, it contains information about conferences, opinions, critical examinations of existing publications and news. Finally, major practical achievements and innovations in the computer and information industry are presented through commercial publications as well as through independent evaluations. Editing and refereeing are distributed. Each editor froni the Editorial Board can conduct the refereeing process by appointing two new referees or referees from the Board of Referees or Editorial Board. Referees should not be from the author's country. If new referees are appointed, their names will'appear in the Refereeing Board. Each paper bears the name of the editor who appointed the referees. Each editor can propose new members for the Editorial Board or Board of Referees. Editors and referees inactive for a longer period can be automatically replaced. Changes in the Editorial Board and Board of Referees are confirmed by the Executive Editors. The coordination necessary is made through the Executive Editors who examine the reviews, sort the accepted articles and maintain appropriate international distribution. The Executive Board is appointed by the Society Informatika. Informatica is partially supported by the Slovenian Ministry of Science and Technology. Each author is guaranteed to receive the reviews of his article. When accepted, publication in Informatica is guaranteed in less than one year after the Executive Editors receive the corrected version of the article. Executive Editors Editor in Chief Anton P. Železnikar Volaričeva 8, Ljubljana, Slovenia E-mail: anton.p.zeleznikar@ijs.si Associate Editor (Contact Person) Matjaž Gams Jožef Stefan Institute . Jamova 39, 61000 Ljubljana, Slovenia Phone: +386 61 1259 199, Fax: +386 61 219 385 E-mail: matjaz.gams@ijs.si Associate Editor (Technical Editor) Rudi Murn Jožef Stefan Institute Board of Advisors Ivan Bratko, Marko Jagodic, Tomaž Pisanski, Stanko Strmčnik Publishing Council Tomaž Banovec, Ciril Baškovič, Andrej Jerman-Blaaič, Dagmar Šuster, Jernej Virant Editorial Board Suad Alagić (Bosnia and Herzegovina) Vladimir Batagelj (Slovenia) Andrej Bekeš (Japan) Francesco Bergadano (Italy) Leon Birnbaum (Romania) Marco Botta (Italy) Pavel Brazdil (Portugal) Andrej Brodnik (Canada) J anusz Brozyna (France) Ivan Bruha (Canada) Luca Console (Italy) . Hubert L. Dreyfus (USA) Jozo Duj movie (USA) Johann Eder (Austria) Vladimir A. Fomichov (Russia) Janez Grad (Slovenia) Noel Heather (UK) Francis Heylighen (Belgium) Bogomir Horvat (Slovenia) Sylva Kočkova (Czech Republic) Miroslav Kubat (Austria) Jean-Pierre Laurent (France) Jadran Lenarčič (Slovenia) Angelo Montanari (Italy) Peter Mowfòrth (UK) Igor Mozetič (Austria) Stephen Muggleton (ÜK) Pavol Navrat (Slovakia) Marcin Paprzycki (USA) Oliver Popov (Macedonia) Saso Prelern (Slovenia) Luc De Raedt (Belgium) Paranandi Rao (India) Giacomo Della Riccia (Italy) Wilhelm Rossak (USA) Claude Sammut (Australia) Johannes Schwinn (Germany) Jifi Šlechta (UK) Branko Souček (Italy) Harald Stadlbauer (Austria) Gheorghe Tecuci (USA) Robert Trappl (Austria) Terry Winograd (USA) Class Wohlin (Sweden) Stefan Wrobel (Germany) Xindong Wu (Australia) i I ■ 1 :• ! ; An International Journal of Computing and Infonnatics Contents: Editorial Program of Informatica . Profiles: H.L. Dreyfus 213 214 - - - t An Token Based Distributed Mutual Exclusion Algorithm . Kia Makki Niki Pissinou Yelena Yesha 221 Quality of Decision Rules: Empirical and ^ Statistical Approaches - Ivan Bruha Sylva Kočkova 233 Logos of the Informational Anton P. Zeleznikar 245 The Information Technology (IT) for Enabling Disabled People - A Strategic Agenda - On-Kwok Lai 267 A Combined Algorithm for Testing Implications of Functional , ■ ' Mirko Maleković 277 Electronic Networking for Philosophical Development in the Principia Cybernetica Project ' . ^ Francis Heylighen Cliff Joslyn 285 To the Internal Representation Debate. ' 294 Global Modernization Plan of Telecomrhunications Infrastructure and Services in Slovenia (MTIS) - Gorazd Kandus Iztok Tvrdy 297 Reports and Announcements 314 r'ii