Lamenting Grid access

Thomas makes some good points about his experiences and the still greater need for GRID computing. I am all for people writing interfaces on the GRID, but there doesn’t seem to be a very easy system. Presumably things like myExperiment, MOBY, and other approaches will make this an easier prospect, but I have yet to see a system that has the flexibility in using custom code, large datasets that need to be local to the compute, can be used by non-computer scientists (but computational saavy biology-types).