Which is preferable, a data sublanguage based on the relational algebra, or one based on predicate logic? Why? Of two sublanguages based on predicate logic, which is preferable, one that uses tuple...

Which is preferable, a data sublanguage based on the relational algebra, or one based on predicate logic? Why? Of two sublanguages based on predicate logic, which is preferable, one that uses tuple variables; or one that uses domain variables? Why?