One of the main goals of the THREDDS project is to develop a framework that allows for the growth of a scientific data web. Several components will be created during the two year development period. First, a language for communicating the necessary metadata between data providers and data users, we refer to this as the PICats (Publishable Inventories and Catalogs). The PICats are currently implemented as XML documents over HTTP. We will be working with metadata standards groups on translation from the many existing and developing metadata standards. Tools will be developed for communicating existing metadata stores as PICats and for allowing the creation and addition of new metadata stores. Second, a set of tools and APIs for accessing and using the scientific data/metadata web from data visualization/analysis applications and from data search sites. Third, search and discovery capabilities. We are working with several groups, including DLESE (Digital Library for Earth System Education), to make sure these data collections can be referenced and searched using their digital library systems.
We have started from the data user point-of-view, concentrating initially on the metadata needed to use data ("use metadata") rather than that needed to find new data ("search metadata"). By developing the THREDDS framework hand-in-hand with several data server sites and several client applications, we have been able to quickly prototype the system and advance the design.
In this paper we will discuss the details of the THREDDS framework, where we currently stand in the development, and what we have learned in the development process.