We developed a database system to store and analyze crop data which are obtained in breeding tests of various plants. In operating breeding tests, researchers pay considerable labor not only for cultivating but also analyzing obtained data. To improve it, we performed a data-modeling of crop data through investigating a number of reports and also through having interviews with several researchers of breeding. By managing crop data with this model, our system enables researchers to analyze the crop data efficiently, which results in reducing analysis labor of researchers. Our system is designed in consideration of usability: researchers can easily store crop data with their own spreadsheet files. They can also analyze the stored data through web browsers with simple operations such as selecting conditions and items to compare. Further, they can search data with conditions and the result can be downloaded. We evaluated our system by applying some breeding tests, and then we obtain good evaluation results with the comment that the system is useful and it reduces the labor of analyses.

View full abstract

There are increasing numbers of systems designed to share information resources among different organizations. However, as the main point of discussion of these systems is the method of mapping the metadata element set of each database onto the others', users' viewpoints are kept out of the discussion of constructing systems. This paper aims to analyze user demands by investigating the actual use of information retrieval systems to examine search methods in shared resource systems. People's retrieval processes were investigated using a cognitive psychological experiment, with protocol analysis and interviews. Through a detailed analysis of their information retrieval processes, the following results were found: 1) a query that allows searches on all search elements together is effective for users who do not understand the details of either queries or databases; 2) elements such as 'Title', the definition of which is clear and easily understood, are used effectively; 3) even if an element is effective for retrieval, such as 'Date', it is not used when the format of the values in the element is not shown. Based on the results of this investigation, we discuss a method for constructing search elements in resource sharing systems.

View full abstract

A web platform has been developed for experts to solve a given problem strategically by taking advantage of available information- data and/or knowledge. Information handled here are contexts (causal relation, consequence and time dependent properties), contents (textual information about specialized information and values in databases), and know-how (how to use computerized resources including simulation codes). It is the main concern of this prototype platform that a balance of sharing of information and appreciation of intellectual property right. This platform was used among materials experts in US and Japan through internet, and the results by such test users were reported.

View full abstract

Many universities and other research organizations have proceeded construction of academic resource repositories under the leadership of National Institute of Informatics (NII). Because it requires much labor to create metadata contents of academic resources, development of tools or services enabling to easily create metadata contents is expected. This paper proposes a system which converts BIBTEXmetadata of academic resources - widely used by researchers in science and engineering - into NII metadata format in order to support constructions of academic resource repositories. The proposed system enables to reduce almost four-fifth of workloads of creating metadata contents. In addition, utilizing compiler compiler to develop the proposed system reduced the develop cost.

View full abstract