Up: CS Research Paradigms and
One builds a system, and measures its performance, either in
actual use or in simulation. Ideally, one is able to compare the
results against previous work.
Accepted standards of quality include:
- The technique should be sound, including the design of the
experiment, the implementation, the measurement techniqe, and
any statistical methods that might be applied.
- It should have some practical relevance.
- It should address a well-defined problem, of clear practical importance.
- It should provide new information, not already available from
theory or previous experimental work.
- It should take into account any existing relevant theory.
It may test the existing theory, explore ideas for new theory, or attempt
to empirically answer questions that have proven to be intractable in theory.