Megan began by describing the differences in the scope between the FASST (Future of Archive Services at Space Telescope) and SHARE study groups. SHARE is primarily looking at MAST archive services such as reprocessing Hubble data. The scope of FASST is wider and encompasses the archive's role within the context of multiple missions, other archive services, and the NVO.
Niall gave a presentation on what goes into an archive in the 21st century. It should be science driven, should include interoperability, and will require new technologies. We will need to define a plan. Some of our concepts will doubtless parallel those of the "Virtual Observatory". He described the complexities of how science is done, and indicated the aspects that an archive contributes to. What is in an archive? What can it do? What could it do for you? What do we need it to do?
Daryl suggested a graphical way of data mining. Stefano suggested that one might want to know all the observations performed in a particular spot in the sky. Chris indicated that in the future people will do more catalog-based research, intercomparing information from multiple sources. Cathy noted the need to provide data as a reliable, understandable product, where possible, so that the user does not need to expend an inordinate effort to understand it. She suggested an interface where one can ask a question such as "how large is the aperture?" for an observation and get an answer easily, without having to dig through documentation.
Stefano noted that one should keep separate in conversation the ideas of relating data to other data versus the process of extracting new information. Niall suggested that an archive can help provide expertise through its services, such as in the extraction of sources from an image to create a catalog. Rick suggested that the system should be a fully integrated one, including getting information back from the researcher into the archive. This would include value-added products, for instance fields for which someone has provided the astrometry.
Stefano noted that a uniformity of interfaces is very helpful, but can be hard to do while also providing needed flexibility. Niall noted that the standardization of metadata is an important part of the NVO effort. An example in progress is the current effort for Starview to handle the various MAST data sets.
Mark suggested we do some case studies of specific HST data products such as the HDF. Which ones have proven to be useful to the astronomical community? What is currently available is not accessible for searches and further manipulation. The question of including ground-based ancillary data is being debated among the various efforts.
Daryl noted the the user-provided data requires documentation on how it was generated. Rick noted that the usual way is through refereeing and publication. This raised the issues of how one can insure the quality and reliability of user-provided data, and how far one should go to require documentation. Niall noted that the researcher's name is associated with the data product. Mark noted the AR Legacy program will specifically solicit such data products. Niall noted that one method could be to referee reduced data products which are submitted to the archive. People already publish information on the web. Another way is to collect feedback from users of the data. Megan noted that some sites already collect contributed software.
For next time, we will put forth several specific science projects.
These will be used as case studies to explore what capabilities are most
needed by researchers.