Date: Sun, 15 Dec 85 18:55 EST To: irdis at vpi Subject: IRList Digest V1 #26 IRList Digest Sunday, 15 Dec 1985 Volume 1 : Issue 26 Today's Topics: Query - Videodisc jukebox - Expert Advice on Doc. Style for "Quarterly Journal of Speech" - Published examples of searches Announcement - New Group and new inf. science program at Chicago Cog-Sci Seminars - Inheritance / Data models / Data types Massively Parallel Networks that Learn Representations, English Dialogue with Computer Databases ---------------------------------------------------------------------- From: "Leonard N. Foner" Date: 7 Dec 1985 18:40 EST (Sat) Subject: Videodisc juke-box? I'm looking for a strange item, upon request from a friend. He wants to know who makes juke-box-like mechanisms for standard videodiscs. The intended application is a device which selects one of many such discs and plays an hour or so of video (some segment of the disc). The higher the capacity of the device, the better. He is also secondarily interested in what technologies exist to pack more video onto a disc (my ideas run to recording only the differences between frames, rather than the whole frame, etc). What ideas are being bandied about? Are there any that the industry might be doing or about to do? [Seems to me that the industry is not likely to make more expensive players in order to sell a $1K videodisc with 200 hours of programming on it (the expensive part being, of course, copyrights), but who knows...] His application calls for 144 hours of video to start with, with another 100-200 hours/year of add-on capacity, so he really wants to minimize discs if possible. How much can be put on a disc now? With what image quality? Is it possible to trade off quality for capacity with standard disc-making plants and players? Note that he's interested in video---moving images---not CD-ROMs and other digital media. Please send replies to me, unless you think the list(s) above are interested in this information as well. If you possibly can (barring internetworking difficulties), please CC a copy directly to me even if you do reply to the whole list. Thanx much, folks! ------------------------------ From: T3B%PSUVM.BITNET%wiscvm.wisc.edu@CSNET-RELAY Date: Wed, 11 Dec 85 23:22 EST Subject: Documentation style for QUARTERLY JOURNAL OF SPEECH Perhaps readers of IRLIST can give me some practical advice. I have been elected editor of THE QUARTERLY JOURNAL OF SPEECH (1987-1989). The journal, founded in 1915, publishes essays in speech,communication, and rhetoric. For many years it has used the footnote/endnote system of the Modern Language Association. But last year MLA abandoned the footnote, recommending instead a parenthetical reference in the text to a "List of Works Cited" arranged alphabetically at the end of the text. So, I am inheriting a journal without a status quo insofar as a documentation system is concerned. We seem to face three general solutions. (1) We could stay with a footnote/endnote system. To do this, we would either retain the style used for the past decades, or adopt the slight modifications that would be required to be consistent with what is now a not-recommended but still described part of the MLA Style Manual. Or, we could switch to the Chicago Manual of Style, which describes footnote usage in detail, and which is widely used by book publishers, and by journals in history. This solution has the advantage of continuity with our past and some alleged technical advantages for scholars who do classical and historical work. (2) We could switch to the new MLA parenthetical style. This would be consistent with many journals in the humanities, most notably PMLA, the major journal in literary criticism. Many printers and publishers think that parenthetical systems are more economical and easier to use than footnotes. (3) We could switch to the APA (American Psychological Association) style guide, which is also a parenthetical, author/date system. The main advantage of this solution is that it would make QJS consistent with three other journals published by the Speech Communication Association, all of which use APA. Further, it has been argued that this would have big advantages in information storage and retrieval, since if all four journals used one style sheet, the printer (Science Press, Ephrata PA) could store all the journals on a single tape, where they could be accessed as a data base, and from which they could be conveniently indexed with a key word system. I need expert advice, and concrete facts, about two sorts of issues: (1) What are the relative merits of the 3 major citation systems open to us as information retrieval devices in themselves? (2) Are the alleged advantages of APA usage real, in terms of where computer technology is, and is going in the near future? That is, is a single format for these journals a short-term or long-term constraint for information storage and retrieval? Sorry to burden the list with such a long note, but this is a complicated problem, and I want to get beyond merely polling the world for its opinion about footnotes vs. parenthetical references. Thanks for any help you can offer. Tom Benson Department of Speech Communication The Pennsylvania State University 227 Sparks Building University Park, PA 16802 814-238-5277 {akgua,allegra,ihnp4,cbosgd}!psuvax1!psuvm.bitnet!t3b (UUCP) t3b%psuvm.bitnet@wiscvm.arpa (ARPA) T3B@PSUVM (BITNET) ------------------------------ From: scott%truth.uucp.uchicago.csnet@CSNET-RELAY Date: Tue Dec 10 11:41 CST 1985 ... What would be really nice is if I could see a raw query, a processed query, and the results of a retrieval. Has anything like that ever been published? [Note: it is not hard to produce this - does anyone know of something already published? - Ed] ------------------------------ From: scott%truth.uucp.uchicago.csnet@CSNET-RELAY Date: Tue Dec 10 11:41 CST 1985 ... I've got two things for IRList. One is that the University of Chicago is initiating an information science program within the Graduate Library School, leading to an M.A with a concentration in information science. I'm pretty excited about it. The other thing is that the research group that started recently here, TIRA, is really gathering steam. I'll write a blurb about both sometime soon. ------------------------------ From: Peter de Jong Date: Tue, 3 Dec 1985 12:52 EST Subject: Cognitive Science Calendar From: nikhil at MIT-NEWTOWNE-VARIETY.MIT.EDU (Rishiyur S. Nikhil) Friday 6, December 2:15pm (refreshments 2:00pm) Room: NE43-512A Inheritance, Data Models and Data Types Peter Buneman University of Pennsylvania The notion of type inheritance (subsumption, ISA hierarchies) has long been recognised as central to the development of programming languages, databases and semantic networks. Recent work on the semantics of programming languages has shown that inheritance can be cleanly combined with functional programming and can itself serve as a model for computation. Using a definition of partial functions that are well behaved with respect to inheritance, I have been investigating a new characterization of the relational and functional data models. In particular, I want to show the connections of relational database theory with type inheritance and show how both the relational and functional data models may be better integrated with typed programming languages. Host: Prof. Rishiyur Nikhil ------------------------------ From: Peter de Jong Date: Fri, 6 Dec 1985 11:53 EST Subject: Cognitive Science Calendar [Extract - Ed] From: ELIZABETH%MIT-OZ at MIT-MC.ARPA Subject: 9.382 --- Seminar in Visual Information Processing Monday 9, December 4:00pm Room: NE43-8th floor playroom MASSIVELY PARALLEL NETWORKS THAT LEARN REPRESENTATIONS Geoffrey Hinton Carnegie-Mellon University I shall describe a new learning procedure for massively parallel networks of neuron-like processing elements. The procedure adjusts the connection strengths in a multi-layered network so as to make it give the correct output vector when given an input vector. The units in the intermediate layers come to represent important implicit features of the task domain that generates the pairs of input/output vectors. As a result, the network can generalise appropriately to new cases. I shall describe a pattern recognition example in which the network constructs a balanced ecology of feature detectors, and a higher level task in which the network learns a set of relationships. The second example illustrates the ability of the network to recognize isomorphisms and make use of them in encoding and generalizing knowledge. ------------------------------ Thursday 12, December 6:30pm Room: Intermetrics Atrium, 733 Concord Avenue Cambridge, Ma. "English Dialogue with Computer Databases" Prof. Carole Hafner Host: GBC/ACM For Info:(617) 444-5222 ------------------------------ END OF IRList Digest ********************