cik@l.cc.purdue.edu (Herman Rubin) (11/18/89)
In article <1989Nov17.173402.21820@world.std.com> bzs@world.std.com (Barry Shein) writes: > >From: brnstnd@stealth.acf.nyu.edu (Dan Bernstein) <>> The Library of Congress probably has <>> hundreds of thousands of subject classifications and I haven't heard <>> anyone calling the nation's library systems useless. >> >>Because most books are put into the right sections. In other words, the >>official subject divisions are easy to understand and widely respected. >>USENET groups don't work that way. > >I bet you'd never get a librarian to agree to that statement. It's all >ultimately a fuzzy, humanist endeavor. Which is fine. > >Cataloguing is a very difficult specialty, and worse, where something >seems to fit changes over time (not that it gets changed, but it can >seem very out of place twenty years later, where would you go to find >early books on cognitive modelling? Psychology? Neurobiology? Math? >Physiology? Philosophy? Computer Science????) I would agree that most books are put into the right sections, but not boast about it. Also, I know of no library catalog system which has cross-classification. Posting to multiple newsgroups, or something like that, is necessary. Suppose that I want to look through our library to find the books to put on reserve for a graduate course in mathematical statistics. There are the standard classifications for probability and mathematical statistics. But many of the books are in the social science statistics section (and many of them have no social science whatever, and are straight mathematical statistics). There are also many engineering texts which are mainly mathematical statistics. There may be good ones under economic statistics or medical statistics. The ones in engineering and medicine may not even be classified under the statistics subsection of those fields. Suppose that a librarian gets a book entitled "Elementary Theory of Cylindric Algebras" to catalog. What are the chances it will be catalogued under mathematical logic? -- Herman Rubin, Dept. of Statistics, Purdue Univ., West Lafayette IN47907 Phone: (317)494-6054 hrubin@l.cc.purdue.edu (Internet, bitnet, UUCP)
sean@dranet.dra.com (11/19/89)
Distribution: usa Organization: Data Research Associates, St. Louis MO Lines: 61 In article <1722@l.cc.purdue.edu>, cik@l.cc.purdue.edu (Herman Rubin) writes: > I would agree that most books are put into the right sections, but not > boast about it. Also, I know of no library catalog system which has > cross-classification. Posting to multiple newsgroups, or something > like that, is necessary. Well, actually there are several systems that have "cross-classification." The UDC (Universal Decimal Classification), which is based somewhat on the Dewey system, allows a work dealing with two or more subjects to be classed by two or more UDC class notations (often linked by a colon sign). There are also several others, but the UDC is one most likely recognized by someone who doesn't work in a specialized library. For example, a book about the use of computers in the management of hospital personnel 362.1 : 658.3 : 681.31 Translation => Hospital : Personnel management : Computers likewise in rotation 658.3 : 681.31 : 362.1 and 681.31 : 363.1 : 658.3 so it could be found under any of those three areas. Also, a side note, the UDC appears to adapt itself very well to computerized information retrieval, and is already available in book or machine readable form. Although it is popular in several parts of the world, it is generally only found in a few technical libraries in the United States. > Suppose that I want to look through our library to find the books to put > on reserve for a graduate course in mathematical statistics. There are > the standard classifications for probability and mathematical statistics. > But many of the books are in the social science statistics section (and > many of them have no social science whatever, and are straight mathematical > statistics). There are also many engineering texts which are mainly > mathematical statistics. There may be good ones under economic statistics > or medical statistics. The ones in engineering and medicine may not even > be classified under the statistics subsection of those fields. The problem comes about when the class is used for shelf listing (the physical placement of the book in the library). The library could buy X number of books to place in each of X locations, or choose one of the classes for the call number on the book. Libraries, being always short of money, choose the later approach. They handle the "pointer-problem" through their catalog, either on-line or on cards (being much less expensive than books for the library, though the researcher may dispute this). Misclassification is always a problem. In USENET people post things to the wrong newsgroup (even cross-post it to multiple wrong newsgroups), and catalogers make mistakes in their subject analysis of a book. The problem with making USENET over to use some type of comprehensive classification scheme, is that someone is going to have to teach everyone how to use it. How to analyze their postings to make sure they post in the "correct" newsgroup, and generally a lot of red-tape. I can see it now you will need to get a PhD in USENETology before being allowed to post, then maybe that's not such a bad idea :-). -- Sean Donelan, Data Research Associates, Inc, St. Louis, MO Domain: sean@dranet.dra.com, Voice: (Work) +1 314-432-1100 Affiliation given for purposes of identification, not representation