A few days ago, while hunting for a bug in PL/SQL code, I stumbled upon the strangest way to write NULL. If the e-mail address parameter was empty, the genius who wrote this PL/SQL procedure set it to , then compared it with ten lines below, in order to log a problem. I really don't know what is it about NULL that scares people so much, but over the years I got used to occasional -1 and 0, or even 'EMPTY'. However, this is the first time I ran across Donald.

Accessing the database from the outside world basically comes down to two options - direct querying or executing stored procedures. Procedural access is often chosen for the wrong reasons - making maintenance significantly harder.

A common practice for handling errors in PL/SQL procedures is to catch all errors in the top-most database layer and convert them into error codes and human readable messages for client applications. This technique is a relict from the past and, in fact, a very bad practice from today's perspective, since it can lead to data corruption.