October 20th, 2005, 8:18 am
on the other hand, for lots of numerical methods problems it's difficult to know what the answer is until you've got your hands dirty and written some code and examined the alternatives - in such cases a research proposal will be more tentative and exploratory rather than set in stone