Medici is designed to support any data format and multiple research domains and contains three major extension points: preprocessing, processing and previewing. When new data is added to the system, whether it is via the web application, the desktop client or through the RESTful services, preprocessing is off-loaded to extraction services in charge of extracting appropriate data and metadata. The extraction services attempt to extract information and run preprocessing steps based on the type of the data. For example, in the case of images, a preprocessing step takes care of creating the previews of the image, but also of extracting EXIF and GPS metadata from the image. This raw metadata is presented to the user in both the Medici web and desktop clients. If GPS information is available, the web client embeds a Google to show the location of the dataset, while the desktop client shows the latitude and longitude values as textual fields in the additional information view. By making the clients and preprocessing steps independent and using RDF as a common domain-neutral representation, the system can grow and adapt to different user communities and research domains.
The Medici desktop client demonstrates how data processing steps can be supported when focusing on data management. The primary concern with processing and data management is to keep track of the data provenance and describe what happened to the data. This history of the data, the data provenance, is fundamental for experimental and process reproducibility. The desktop client enables users to launch external tools and capture the inputs and outputs of the processes and store this information as part of the metadata. The web client displays this information on the page so that researchers can easily identify derived data and determine its origin.
For more information check out a two page Medici whitepaper that was presented at the Microsoft e-Science in 2010.
