When you say "people assume..." do you mean users? In that case, I think
they assume that don't have to worry about the underlying algorithms;
people want relevant results, which shouldn't have to depend on any deep
understanding of what happens behind the scenes.
>But does it need to be linear? How does a
>
> log (*that* word) / total number
Our density operator is not linear... and it gives pretty good results,
although density is rarely the only evidence involved in coming up with a
relevancy score.
Nick