CORPUS SEARCH SYNTAX
--------------------

The search query syntax is a work in progress, but mostly follows
established text/web search conventions.

Search is based on matching the terms in the query string with the
contents of text files in the corpus. Relevance is determined by
ranking the number of search term occurrences, or 'hits', in each
file. The file with the most hits will appear at the top of the
results list. When multiple search terms are present in the query
string, the ranking is based on the term with the least number of
hits.

If multiple terms are present in the query string, there is an
implicit logical AND between each term. A file will only match the
string if it has hits for all of the terms.

By default a search term will match any occurrence of a word that
starts with the term. For example, a search for cat will match the
words cat, cats, catlike, cateye, cattle, and so on. If only exact
matches are desired, that can be accomplished by putting quotes around
the term. If quotes are placed around multiple words -- a phrase --
instead of a single word, the algorithm will match the exact
phrase. However, the current implementation is very slow for this type
of search. The gopher client will disconnect before most searches of
this type will complete.