This paper presents a word-based technique for storing natural language text in compact form. The compressed text consists of a dictionary and a text that is a combination of actual running text and pointers to the dictionary. The word-based technique nearly halves the storage required with no loss in information. Furthermore, the process requires only a moderate amount of time overhead to compress and store the text and can retrieve and decode the encoded text faster than the original text can be retrieved.
A systematic approach to the design of fixed-length, derived, truncated search keys for corporate author records is described. Certain distributional and informational characteristics of the elements of the entries are shown to be useful in search key design while some others are not. A statistical method is described for predicting the performance of search keys for files of arbitrary size. This method is employed for predicting the performance of a key known to perform well in a small sample file.
One of the most important impacts in cataloging in recent years is the use of a cooperative cataloging data base. In order to make the most advantageous use of such a system, it must contain uniform and accurate cataloging. This article reports the results of a study of cataloging records input into the OCLC data base by participating libraries (non-MARC records). The records were analyzed to determine the various fields of the cataloging record where most errors occur as well as the types of errors that are being made on the cataloging input into the data base.
An automated system has been developed for cataloging the monographs, periodicals, music, maps, films, recordings, and realia received by the Copyright Office for registration. Some of the problems in disseminating this unique data base through national networks are considered. The system features certain special techniques automatic inversion of personal names and the use of one entry as a model for the next.
A mathematical model of a regional interlibrary loan system was developed as a tool for the design and evaluation of a strategy to equalize the load of ILL on libraries in the system. The model simulates an automated ILL system that keeps track of each library's ILL activity and identifies those libraries that are overburdened with requests. The load-leveling capability of several algorithms for representing each library's ILL activity was tested. It was found that the algorithms were nearly equal in effectiveness but presented differing implementation problems. Because of recognized limitations in the environment represented to the model, further evaluation of the load-leveling concept in a test version of an automated ILL system is suggested.
Important deliberations are now going on in the House Communications Subcommittee regarding the proposed revision of the Communications Act of 1934. Subcommittee Chairman Lionel Van Deerlin (D-Calif.) introduced a revision bill, HR 13015, on June 8, 1978. A new law, once enacted, is likely to affect many aspects of telecommunications use in the United States for the remainder of this century, and well into the next. The following statement establishes the vital interest libraries have in this matter and presents many of the issues confronting the library community with respect to the formulation of an enlightened communications policy. This statement was prepared by the Washington Office of the American Library Association with assistance and contributions from several members of the LITA division (then ISAD) who are acknowledged in the concluding paragraph.