Although some scientists, such as many physicists, may prefer a command line approach to submitting computational jobs, a majority of scientists want to be shielded from the innards of a computer. A popular approach is to build portals; user community web sites that allow job submissions from the convenience of a web browser. The development of such portals typically takes many months as the developer of such a portal has to learn how to integrate the various components: a portal system such as GridSphere (http://www.gridsphere.org/), a job submission architecture such as GridSAM (gridsam.sourceforge.net) , a security system such as the Java CoG Kit (www.globus.org/cog/java/), a system for staging input and output files, and so on.
In this project, you will _not_ use the conventional approach of programming a portal from scratch, but instead use a new system developed at the National e-Science Centre. This system, called Rapid (http://research.nesc.ac.uk/research/), has all the components mentioned above already in place. The philosophy of Rapid is to generate portals with the minimum amount of Java programming and with the aim to deliver a custom portal that hides the gory details of job submissions. You will wrap a number of the command line programs using the Rapid system in such a way that it helps brain imaging experts perform particular tasks using imaging tools developed by the SFC Brain Imaging Research Centre (http://www.sbirc.ed.ac.uk/).
It is up to you to use Rapid to design, build and deliver the portal, as well as to assist in getting the brain imaging tools to work on the hardware platform, the Edinburgh Compute Data Facility (http://www.ecdf.ed.ac.uk/). You will work in collaboration with both the development team of Rapid and the experts developing imaging tools at the SFC Brain Imaging Research Centre. This project provides the perfect opportunity to apply e-Science in practice.
If you have sufficient time in the project, you may also integrate a 3D viewer into the portal, such as JIV (http://www.bic.mni.mcgill.ca/~crisco/jiv/). An example of how this would look can be seen from a previous project in Computational Chemistry (https://research.nesc.ac.uk/node/296).