lib-s-mocs-kmc364-20141005044842 182 TECHNICAL COMMUNICATIONS REPORTS-LIBRARY PROJECTS AND ACTIVITIES Ohio State University Health Sciences Library Uses Autamated Bookstack System The new Health Sciences Library at Ohio State University began serving stu- dents in May 1973 with some of the most advanced features in any library in the country. It contains an automated book- stack system to locate and file books, and is the fourth library in the country to have the system ( Randtriever, manufactured by Remington Rand Corp.), says Jo Ann Johnson, director of the Health Sciences Library. "The bookstack system will find and de- liver a book via a conveyor belt in about a minute," Miss Johnson said. The chief advantages of the system are that it saves space and is speedy and accurate, she pointed out. "The book stacks in the new library take up about 15 percent of the total space while in most libraries the stacks take 40 to 60 percent of the space," Miss Johnson said. Aisles in the stacks are narrow, about 15 inches, and the shelves rise through two stories of the library-twenty-two feet in all, she said. The library has a ca- pacity of 175,000 volumes. The accuracy of the system will reduce the problem of misfiling. Also, book theft should dwindle because the stacks will be closed to users, she said. The library is connected with the com- puterized circulation system of the univer- sity library, made up of a main library and twenty-three branch libraries. This circu- lation system is the first of its type in the country and permits library users to place telephone calls to learn titles and authors and to charge out books. Other features of the modern library will include a computer-assisted instruc- tion area to be completed later, and con- nections to MEDLINE, the international computerized information system of med- ical journals. Miss Johnson explained that the auto- mated books tack system works like this: A library staff member sends instructions via a terminal to an electronic device in an aisle. The device travels on vertical and horizontal columns in the aisles. It picks out a small bin of books containing the requested one, then travels to the end of the aisle and places it on a conveyor belt. At the terminal, the staff member se- lects the requested book from the bin, usually containing about eight volumes, and sends the bin back for refiling. A glass window permits observation of the system. University of California, Berkeley Serials Key Word Index The University of California, Berkeley, General Library has published a Serials Ke y Word Index to titles of 45,741 seri- als. The computer-produced index is the largest of a fairly new variety of key word indexes, covering titles of serials rather than articles. The 360/ Assembler pro- grams written by the Library Systems Of- fice include a number of innovations. Berkeley serial records are stored in MARC format, upper-lower case, capital- ized by citation rather than catalog stan- dards. The key word extract program ig- nores prepositions and conjunctions, etc. (which are not capitalized); treats certain multiword terms (La Paz, United Na- tions) as single words; prepares a library- standard sorting key (with U.S. filing as UNITED STATES, & filing as AND, and distinction made between two types of hy- phenation); and does no stop-list search- ing or other searching for excluded words. Key lines are sorted by key word; all other processing is based on an alphabetic file of key words attached to main entries. Thus, vocabulary control (forced inter- filing of abbreviations, synonyms, cog- nates, etc.-not heavily used in this edi- tion) is a fast, simple runtime operation, changing certain key words (on a single alphabetic pass) and generating "see" ref- erences. Exclusion of low-content words is also a fast, simple runtime operation, done in the printing program, allowing ex- cluded-word entries to print if the word occurs first in either title or author, and generating an explanatory note under each excluded key word. Listings are main-entry, alphabetic un- der key word groups, with brief holdings, campus location, and call number where available. The key word appears in all capital letters within each entry, and re- dundant entries are collapsed-that is, if a wqrd appears more than once in an en- try, each occurrence is capitalized, but the entry is only listed once under the key word. The first edition limits entries to 98 characters and holdings to 13 characters; the programs have since been revised to allow up to 193 character entries and up to 45 characters of holdings. Both versions T echnical Communications 183 of the programs retain as much runtime flexibility as possible, while maintaining extremely low running time. The first edition, including mostly non- document currently-received titles, is photocomposed in a 6-point slab-serif type and published in three paperbound vol- umes. Copies are available for $60 a set from Systems Office, Main Library, Uni- versity of California, Berkeley, CA 94720. - Walt Crawford, University of California, Berkeley PROGRAMMING AND COMPUTERS PLEA, a PLI 1 Efficiency Analyzer PL/ 1 users find that the language of- fers infinite ways to invoke inefficient code. Partial defense is provided by care- ful manual reading. Another, and very PLEA UT~~CTIO'! ANALYSIS PACE STAT£NENT T~A~ COUNTS FOil MAIN P~OCEOORf T!SliMEo OFFSET 000)68 IN LOAD NOOULEo 'l 141 :' 69 ·r- 10 1 3 11 "\J ~ u •• ~~ 2 II ,, u 4 »- ' I Z1 l6J ---lt 25 9 35 14 u 5 ' ' f: Hi ' .. "s .. , u ~- 34 10 !i 3 .-... !. · . Ul .... _,... 6 ' E0Tt1) ; J = (TOF + BOT) I 2; IF B > .A(J) THEN BOT = J; ELSE TOP • J; END; IF TOP> 101 B = A(TOP) THEN; /•NOT FOUND•/ ELSE FLAGl • 'l'B; 'f, Total ~ 20.4 10.1 13.4 s.z 0.7 19.4 1.1 1.5.6 88.9 9.4 Totals 98.)* ( 0 )- Truncation error. No.Traps 151 197 121 1Z 28J 17 232 1312 146 14.58 Fig. 2. Each test repeated 2000 times with Argument B in Array and 2000 times with Argument B not in Array. Sampling interrupt interval was .00 seconds. to get a reasonable sampling of the re- mammg blocks. A comparative run showed that the optimization overhead was charged to the proper statement groups, but pragmatists will note that the problem setup was biased against the bi- nary search solution. However, using the trap totals from Tests 2, 3, and 4, the 50 percent Proba- bility Test indicates that the probability of no significant difference between meth- ods 2 and 4 is more than 5 percent; the probability of no significant time differ- ence between methods 2 and 3 is more than 1 percent. PLl:A is available at a program distri- bution fee of $25 from the SHARE Pro- gram Library Agency, Triangle Universi- ty Computer Center, P.O. Box 12076, Research Triangle Park, NC 27709. Thanks are due Dr. David Gomberg, University of California, San Francisco Computer Center, for most of the runtime and several of the statements used in the test.-]ust·ine Roberts, Systems Librarian, UC-San Francisco INPUT To the Editor: I am writing to you concerning the ar- ticle which occurred in the September 1972 issue of Journal of Library Automa- tion entitled "The Shared Cataloging Sys- tem of the Ohio College Library Center." I also note that this issue of your jour- nal, even though dated Sept.-mber 1972 was not received until July of 1973 by this library, and, indeed, it was a timely arrival for at the present time the North- west Association of Private Colleges and Universities is investigating the feasibil- ity of seeking service from the OCLC for some of its library requirements. However, in talking with Mr. Kilgour and his asso- ciates at ALA this summer it was exceed- ingly difficult to get a complete cost pic- ture of participation in the OCLC and to this date we have not been able to get a complete cost breakdown obligation. In this regard, this article was extreme- Technical Communications 185 ly interesting and I requested one of my staff members to do a careful analysis of the cost aspects of the OCLC services. I am attaching this analysis for your inter- est and perhaps it will be of suitable per- tinence for the readership of your journal. Certainly I, and other of my colleagues at this university and in NAPCU would more than be interested in response by Mr. Kilgour and his associates. SUMMARY: Desmond Taylor Library Director Collins Memorial Library University of Puget Sourul Tacoma, Washington AN ANALYSIS "Average cost per card for 529,893 cat- alog cards in finished form and alphabe- tized for filing was 6.57¢ each ... the system is easy to use, efficient, reliable, and cost beneficial. An off-line catalog card production system based on a file of MARC II records was activated a year be- fore the on-line system." Requests were hatched weekly. Library of Congress card numbers were keypunched onto cards for searching. Seventy percent were found the first search. "Members could specify a re- cycling period of from one to thirty-six weeks ... before unfulillled requests were returned." Lowest price in lots of one-half million Permalife cards was $8.01 per thousand. CPA's checked the system and found that all direct costs were included in 6.57¢ cost. No mention is made of the preexisting cataloging systems-