The SRW Editorial Board proposes to add the following index to the CQL
context set (see:
http://www.loc.gov/z3950/agency/zing/cql/context-sets/cql.html#indexes )

index name: allRecords --
A special index which matches every record available.  The relation and term
are to be ignored, but the recommended syntax is:

                      cql.allRecords = 1

For example, you could:

1.  harvest an SRW server's records with:

              cql.allRecords = 1

2. match all records apart from those with cat in the title, with:

       cql.allRecords = 1 NOT dc.title any cat