Our goals for Jag

We, of course, did not formulate our goals when we began work on Jag. At the time we were merely curious about Java and tired of C++. However, here are the goals that we should have had:

Jag should enable a researcher to

• prototype complex inversion problems,
• scale prototyped solutions to various data input sizes,
• reuse Jag code for different projects,
• collaborate with other researchers, possibly at remote sites or in other disciplines.
• publish results in various forms, particularly on the World Wide Web.

Did we succeed? We think we did, with the notable exception that solutions may not scale well to extremely large problem sizes, because of Java's inferior performance. We will revisit these criteria in our conclusion section.