There are many opensource applications that support postgresql (eg , gforge , bricolage , dspam ..) but does not use schemas(namespaces) as a result of which you are forced to use/create a new database and loose the advantage of linking the application data with your existing database. as no-cross database queries can be done in PG.
my question is , is it a feasible idea to have some special kind of database in the postgresql cluster that mimics a schema of an existsing database.
say rt3 is a special database that links to existing maindb databases' rt3 namespace then any DDL / DML done to public schema of rt3 database is redirected to rt3 namespace of the existing database maindb's rt3 schema.