by Walt Warnick and Sol Lederman
OSTI has embraced a new paradigm for sharing scientific and technical information (STI). Historically, OSTI has fulfilled its mission  of providing STI to scientists, researchers, and the public by hosting, or collecting, documents and/or metadata. OSTI's new paradigm is to make content searchable that is often hosted by others; today, OSTI connects those seeking the content with the organizations that host it.
Beginning in the late 1940's, with OSTI's production of the Nuclear Science Abstracts - which was to go on for nearly 30 years, OSTI entered into the business of collecting information. Beginning in the 1990's, OSTI began creating web application to make the collected content openly accessible and conveniently searchable. ETDE Web, DOE Information Bridge, the Energy Citations Database, and DOE R&D Accomplishments are some of the successful applications.
In the last several years, OSTI's approach to disseminating STI has evolved. Recent applications such as the Eprint Network, Science.gov, DOE Science Accelerator, and WorldWideScience.org connect users with the highest quality science information without collecting or hosting it.
How does OSTI move beyond collecting to connecting and what does connecting mean? OSTI's new applications search content that is housed in document repositories owned by a number of government agencies and government-sanctioned organizations. OSTI applications search a number of these repositories on the fly and they aggregate the content from the sources they search and present the most relevant of the search results to the user. This simultaneous and real-time search of multiple repositories is called federated search. OSTI's federated search applications serve as portals to specific subjects. In being subject-specific, they connect users to the highest quality STI in their fields of interest.
Why is OSTI embracing the connection model? Quite simply, OSTI can far better achieve its mission by making great quantities of content openly accessible and conveniently searchable, but it is impossible to collect and keep current such quantities of content from multiple content sources. "Connecting" to content is doable, while "collecting" is not.
We believe that by connecting users to content, we provide a more comprehensive and authoritative search. In doing so, we accelerate the advancement of science.