Statistical offices have always had the problem of how to distribute the data they generate. Despite advances in computer composition, printed publications remain expensive and time consuming to produce. Printed media also do not easily lend themselves to re-use in extracts, newly generated tables, or calculations. Data provided on computer tape not only require access to a mainframe computer, they also generally involve investment in specialized ad hoc programming services. With the spread of personal computers, some statistical offices are providing their data on diskettes that can be read by these machines. However such diskettes are often simple downloads from mainframe computer tapes, and most users do not have the programming support needed to utilize the data or to understand and analyze them in the absence of extensive accompanying documentation. As a result, sales have been small and the productive use of data much curtailed. The research carried out under Phase I of the present project successfully tested the feasibility of developing a form of electronic publishing that would integrate statistical data with their documentation so that users could read the output directly, as in a printed publication, and the computer could utilize it directly for processing and analytic purposes. Phase I also successfully developed prototype software that would allow users to list, retrieve, and browse through tabular data and documentation, make extracts from them, and print them out or translate them into forms suitable for processing by other software packages. The recommended award will support the final (Phase II) stage of a project funded under the Small Business Innovation Research Program. Phase II will be concerned with advancing the accomplishments of Phase I into a generalized reference library system capable of serving the varied dissemination practices of different statistical agencies, providing facilities for retrieving and merging data from different databases, including bases containing diverse types of data, and furnishing processing and analytic tools that meet the needs of the casual user at the same time as they satisfy the more complex requirements of advanced users, including data providers. Phase II will also be concerned with creating modifications to the basic software that adapt its operating characteristics and special features to the changing landscape of computer technology and electronic publishing. The products and services to be developed under Phase II research will be undertaken in cooperation with statistical offices of various sorts both at home and abroad. The resultant technology will thereafter be offered to such data sources via direct sale or leasing arrangements. Since large corporations face many of the same data handling problems as statistical offices in the public sector, extension of the technology to corporate users will be explored.
