[Koha-devel] OO-RDB mapper

[Rick Welykochy]

Have the Koha developers given any thought to recoding the raw DBI
calls in Koha with an OO-RDB mapper?

There are a number of them available on CPAN.

I've written these things, at varying levels of complexity and completeness,
in the past and the advantages are numerous:

(*) separate the Perl code from the SQL

(*) rows in tables are represented by objects so no more errors
     like this: $row->{naem} -- instead: $row->name

(*) the abstraction layer of an OO-RDB mapper can provide 99%
     database driver independence; Koha could suddenly be supported
     by MYSQL, Postgresql, oracle, sqlite, etc.etc.

(*) the abstraction layer allows application-level hooks to be
     installed *anywhere* in the RDB transcation process and total
     global control of RBD operations is suddenly available (instead
     of scanning/editing 1000's of lines of source code to make
     a basic functional RDB change)

(*) audit logging can be automated; all insert, updates and deletes
     can be caught in the abstraction layer -- do what you want with them


(*) updating a row in a table becomes foolproof:
     $row = OORDB->fetch(criteria);
     $row->name("new name');
     $row->debt(undef);
     $row->commit;

The list of advantages does go on and on.


cheers
rickw



-- 
_________________________________
Rick Welykochy || Praxis Services

Q: What's the difference between a software sales person and a car sales person?
A: With a car sales person there is a slight chance that he actually knows how to drive.