Science Archives in the 21st Century
There has been a persistent dream of a global data environment in which all of the Earth and space science data could be organized in a common way with .one stop shopping. for any data product. In this talk we will consider some practical principles for guiding data system integration. The highest priority in any science data development must be providing the scientific community with access to high quality and well documented data. That can only be accomplished with the active participation of scientists involved with the analysis of the data. This in turn requires that the data system be organized by scientific discipline. In the past decade the sources of scientific data have increased greatly as more spacecraft with more sophisticated instruments have been flown both in the United States and abroad. The advent of the internet has made widely distributed data sources both practical and desirable. Earth and space science data are now distributed around the world. The highest priority in organizing the world.s scientific data holdings must now be to integrate these distributed data holdings within each discipline. We will use two examples to demonstrate how this can be accomplished- the International Planetary Data Alliance (IPDA) and the SPASE (Space Physics Archive Search and Extract) consortium. In both cases the key to data system integration is the adoption of common metadata standards. In planetary science the existing metadata standards developed by the Planetary Data System (PDS) are being adapted by IPDA members to serve the global community while the international SPASE consortium is developing world wide standards for space physics data. In planetary science working standards existed before IPDA so they are being extended to serve the emerging international data system. In space physics there are many different data systems already in existence so the metadata standard is being designed as an intermediate language (an .interlingua.) to connect the ones already in existence. Further integration across disciplines must involve a careful balance between scientific need and cost.