: The development of the Database System for Labour Accident Fact Exploration (abbreviated as Database "SAFE" in this series of studies) has been attempted, and the parts of it were already reported in recent publications of RIIS (The Research Institute of Industrial Safety).
    After the attempts, the Database "SAFE" has been in service for assisting the reserachers in planning and conducting research activities in RIIS. For example, it is used to extract desired labour accidents information related to research projects.
    During the course of utilization of the Database "SAFE", some demands for improvement of the "SAFE" arose. To satisfy the demands, enhancements of the SAFE were carried out as described in this series of studies. This newly developed information retrieval system is called "New Version of the SAFE", reffered to as "SAFE-Ⅱ". In this report, four main improvement matters are discussed.
    Firstly, two new databases on specific labour accidents are developed, which enable us to treat more detailed information-items in information retrieval for specific labour accident cases. For example, in the case of labour accidents associated with crane operations, information about "weight capacity of crane concerned" is indispensable. So, it is necessary to prepare new independent databases having specific information-items according to its nature. From this point of view, the Database of accident concerning crane, fire/chemical explosions were newly developed in the present studies.
    Secondly, databases on reference books and references concerning industrial safety and accidents are assembled to "SAFE-Ⅱ". In co-operation with the library of RIIS, more than 10,000 books or printed matters concerning industrial safety and accidents are collected and utilized for research activities in RIIS. These books are written in Japanese or Western languages (English, French, and so on). Development of a database for these reference books has been carried out since last few years, which started independently from the devleopment of the "SAFE" at the beginning. After the completion of information retrieval support system of the "SAFE", assembling of the book reference databases into the "SAFE-Ⅱ" is made to provide more efficient reference services.
    Thirdly, some utility functions for data interchanging are also developed in this series of studies. In connection with data handling in the "SAFE", it was recognized that development of easier methods for data handling by use of Japanese word-processors is vital in order to reduce the difficulty of task for data entry.
    Finally, new information retrieval support functions are also developed in this series. In the "SAFE-Ⅱ", plural databases on some specific accidents exist simultaneously according to information-items of each database. So, in some information retrieval procedure, it is necessary to mutually synchronize information retrieval results extracted from different plural database. By means of newly developed information retrieval support functions, it becomes possible to search a particular accident case from different databases at the same time and to synchronize their information retrieval results relationally or reciprocally.
    Then, this newly developed information retrieval system which is called again "SAFE-Ⅱ", is brought into practice in this series of studies. It consists of five databases, newly added infor mation retrieval support functions, and newly added utility functions for data interchanging in data handling. In this report, the ontents of this series of studies are divided into 5 chapters and the main substances of each chapter are as follows:

1. In Chapter 1, the historical process about dvelopment of the "SAFE-Ⅱ" and an outline of the "SAFE-Ⅱ" are introduced.
2. In Chapter 2, newly developed utility functions for data interchanging are explained. By means of these utilities, easy data entry by means of a Japanese word-processor can be realized.
3. In Chapter 3, newly developed information retrieval support functions in the "SAFE-Ⅱ" is shown. Especially, support function of relational information retrieval, which can synchronize information retrieval results extracted from different plural database, is discussed. In tihs chapter, some observaional results concerning automatically produced key-words from information-item of Japanese sentences are also presented.
4. Two newly developed database on specific labour accidents are explained in detail in Chapter 4. These new databases are the Database of labour accidents concerning Crane and fire/chemical explosion, and they consist of important part of the "SAFE-Ⅱ".
5. In Chapter 5, development of databases on reference books concerning industrial safety and accidents is explained in detail. These book databases consist also one of the important parts of the "SAFE-Ⅱ".

Keywords; Labour accident, Database, Information retrieval system.

: Five pieces of database are included in the information retrieval system of the Research Institute of industrial Safety, "SAFE-Ⅱ". They use FAIRS-I as a common D.B.M.S. (database management system). However, they are fundamentally independent of each other. So, the data entry system for each database has been developed in the different way. In this report the detail of such data entry system is discussed.
    There are six kinds of possible data entry methods to the database managed by that D.B.M.S. in the Research Institute of Industrial Safety (RIIS).
    The first is to use FAIRS-Ⅰ's command directly at TSS terminal. This method is suitable for handling a small amount of data.
    The second is to use a general editor prepared by PFD (Programing Facility for Display Users) to make an intermediate file. The contents of this file are supplied to input data of the database by batch job process. This method is useful only for skilled operators to PFD.
    The third is to use a special entry program of the host computer to access an intermediate file. By this method, it is possible to show operators what kind of datum should be entered by the preparation of menu screen, and the data can be checked as soon as the operator manipulates the keyboard. This method is adopted by the data entry system of the book database and that of labor accident report database.
    The fourth is to use a Japanese word-processor program for personal computers to create a document file. After compiling such file, the content of it is converted to a file of general type, then transmitted from the personal computer used as a terminal to the host computer by means of emulator program, and finally it is converted to a certain form of the intermediate file. This method is adopted to the data entry system of the database on accidents concerning explosion and fire and crane. The SAFE-Ⅱ system can treat three types of file made by following Japanese word-processor programs, Ichitaro, FM-OASYS, and EPOWORD-J. The created file is converted to the intermediate file in the host computer.
    The fifth is to use the editor of a database program for the personal computer. By this method it is possible to check the datum, and it is easier to make the checking program than by the third or sixth method.
    The last is to use a special program of personal computer written by BASIC to make a file in the personal computer and transmit it to the intermediate file in the host computer. This method has the flexibility, but it needs great efforts to make the program.
    In case to use a long Japanese sentence, a shift code has to be repeated twice to justify the character length before the sign of continuation. All the programs developed by RIIS support this function.

Keywords; Data entry, Database, Emulator, Personal computer, File transmission, Information retrieval system

: In the RIIS (Research Institute of Industrial Safety), the Information Retrieval System concerning labour accidents has been developed and was released in service for assisting research activities. This system has been referred to as "SAFE" which stands for "the Database System for Labour Accident Fact Exploration".
    In this series of studies, some improvement of the "SAFE" is carried out, and after this study, new information retrieval system is renamed as "SAFE-II". This paper presents newly added support functions concerning information retrieval in the "SAFE-Ⅱ" and describes the results of investigations about automatically produced key-words from information-item of Japanese sentences of the "SAFE". Outline of this paper are as follows:
    Firstly, the original data of the "SAFE" and the brief outline of the "SAFE" are explained. When serious labour accidents involving fatal victims or more than three injured workers oc curred investigations about causes of these labour accidents are examined by staff of the Labour Inspection Offices. After these investigations, mission reports on the accidents are submitted to Labour Standard Offices. Information contained in these reports are very important not only for working out preventive counterplans against similar labour accidents, but also for bringing into practice future research activities of RIIS.
    In this connection, development of database system about these pieces of information was carried out, and this experimentally developed system is called "SAFE" in this series of stud ies. Details of this system are already explained in the recent publications of RIIS, and a few important points of their contents are cited again hi this paper (Fig.3.1, Table 3.1).
    Secondly, the outline of improvements of "SAFE" to "SAFE-Ⅱ" is indicated. These improve ments are carried out, in this series of studies, in order to supplement a few additional func tions or utilities for information retrieval. As one of the main improvements, development of new databases on specific labour accidents is put in practice in this study. They are the Database of ac cidents concerning crane and fire/chemical explosion, and detailed contents of their development are explained in chapter 4 of this report. They are fundamentally independent from the Database of the Labour Accidents Report which has been already developed as the "SAFE". These newly developed databases treat more detailed information-items according to its characteristics, for example, in the Database of accidents concerning crane, information-item of "weight capacity of crane concerned" is treated, such particular/detailed information items are not contained in other databases in the "SAFE-Ⅱ". And so, some peculiarly labour accidents are registered in different plural databases each by each according to its particular/detailed information-items.
    In actual information retrieval procedure in the "SAFE-Ⅱ", results extracted from these different plural database, especially information from particular/detailed information-items of newly developed databases, can be used to complement the information obtained from the Database of the Labour Accidents Report. For assisting such new usage of the "SAFE-Ⅱ", it is necessary to develop some support functions for information retrieval which can synchronize relationally information retrieval results extracted from plural different databases. This new added support function is one of the important support functions for information retrieval in the "SAFE-Ⅱ", and their details are shown in this chapter of this paper (Fig.3.2 ˜ 3.18).
    Thirdly, in this chapter, observations concerning automatically produced key-words from information-item of Japanese sentences of the Database "SAFE" are also discussed. In this database, there are three types of information-items; numerical value items, Japanese index items, and Japanese sentence items. By means of a function of D.B.M.S. (data base management system), Japanese key-words are automatically produced from information-item of Japanese sentences, and they are registered at the inverted file of Japanese key-word.
    In actual information retrieval procedure, we can employ these Japanese key-words as free terms for information retrieval. And so, it is important to maintain these Japanese key-words, so-called, definition of synonyms, registrations to dictionary of computer. In order to acquire information for this maintenance, some utilities are developed in this study. By these utilities, we can monitor and store actual production situation of Japanese key-words and free terms used in actual information retrieval procedure (Fig.3.19).
    According to the observation about actual situation of separation of Japanese key-words, the average number of kind of key-words per one labour accident report is 9.24 words, and the average total number of key-words used per one labour accident report is 70.91 words. About automatically produced Japanese key-word, key-words of high frequency are indicated in Table 3.2.

Keywords; Labour accident, Data base, Information retrieval system.

: As is shown in chapter 3, the Database "SAFE" treats the informations described on the first pages of Investigation Reports of labour accidents by the Ministry of Labour.
    In some cases of special types of labour accidents, detailed information is necessary to inves tigate and analyze causes of accidents.
    In the Research Institute of Industrial Safety (RIIS), the staff of the Chemical Safety Research Division has been continuing to collect information described in the investigation reports of labour accidents for more than 30 years, and we could make full use of the information for the purpose stated above.
    The database on accidents concerning to crane operation is also developed for the same reason. The data source for this is mainly the Crane Yearbooks, and some data are added by researchers of the RIIS.
    In this chapter, development of the two Databases on labour accidents, one concerning ex plosion and fire and the other crane operation, is explained in detail.

Keywords; Labour accident, Database, Explosion, Fire, Crane

: The main library in the Research Institute for Industrial Safety was founded to provide centralized processing services for academic publications in the Tamachi head office in 1971. Since then, the collection of books and periodicals related to industrial safety have increased steadly. By the end of 1989, the registration of book collection exceeds more than 10,000 volumes including more than 2,000 foreign books.
    In accordance with the increase of the number of registered volumes, it becomes necessary to provide a more efficient system to facilitate book acquisition, registration, and cataloging procedure work and book information retrieval service system in the library.
    For this purpose, development of a new computer-based system to assist book registration, cataloging procedure and for book reference information retrieval system has been carried out for the last three years. This paper describes the outline of the developed system for book registration, cataloging and information retrieval system for book database system in the library of the Institute.

Keywords; Library, Computer-based book registration system, Database, Book reference and information retrieval system