Comparing protein structures is a long standing problem in protein
science. A general solution is still missing. The current solutions
are either optimized for a certain structural aspect or a specific
application. An ideal method should be able to capture the different
types of structure similarity, should be fast enough to keep pace with
the rapidly growing structure databases and should deliver biological
meaningful alignments.
We give an overview of the basic problems. We analyze several methods
for how near they come to an ideal method and where they fail. Then we
introduce a distance based method, which aims to be fast enough to
perform database searches on a standard PC hardware but also delivers
alignments.