The results show that by using J%, we should expect slightly increased LOC and number of files. This is the main problem of
J% in its current specification. For each SQL query a new external type should be declared, which practically means the
creation of a new source file. Consequently, the increased LOC can be explained by the declaration of boiler-plate code that
each external type requires. External LOC (Ext. LOC) counts the lines of code for the external types’ source files. Since Java does
not support external types, the count is always zero for the Java version of the projects. Note that the number of SQL Queries is
reduced in AddressBook, ExamJ and RUBiS. J% encourages query reusability and for the same queries, the same external type
can be reused. For example, the RUBiS project has originally 44 SQL queries and 16 of them could be reused.