In this paper we report on a series of tests we have done over three   generations of search engines and comment on the completeness of their   coverage. The existing and new search systems, while becoming more   focused, tend nonetheless to generate misses and false hits; this is due  to the fact that they attempt to match the specified search terms without  context in the target information resource. Our tests, repeated over  three generation of search engines, over the last few years show that  while the recall for a simple search has increased, the precision has  actually ecreased. We give some reasons for this decrease and suggest  the need for increasing the intelligence of the search process by adding  semantics information to the indexing scheme.