myTea: Connecting the Web to Digital Science on the Desktop

Bioinformaticians regularly access the hundreds of databases and tools that are available to them on the Web. None of these tools communicate with each other, causing the scientist to copy results manually from a Web site into a spreadsheet or word processor. myGrids' Taverna has made it possible to create templates (workflows) that automatically run searches using these databases and tools, cutting down what previously took days of work into hours, and enabling the automated capture of experimental details. What is still missing in the capture process, however, is the details of work done on that material once it moves from the Web to the desktop: if a scientist runs a process on some data, there is nothing to record why that action was taken; it is likewise not easy to publish a record of this process back to the community on the Web. In this paper, we present a novel interaction framework, built on Semantic Web technologies, and grounded in usability design practice, in particular the Making Tea method. Through this work, we introduce a new model of practice designed specifically to (1) support the scientists' interactions with data from the Web to the desktop, (2) provide automatic annotation of process to capture what has previously been lost and (3) associate provenance services automatically with that data in order to enable meaningful interrogation of the process and controlled sharing of the results.