I am all for normalization and referencial integrety, but most of the 'more professional' places i have worked for seem to think they are bad things (and joining tables by email addresses is a 'good' thing).

If the database will remain small, then choosing not to normalize for a perfomance gain smells like false optimization to me.

When i worked at a very large IT center for a nameless accounting firm, i noticed that NONE of their tables were normalized and none had any referencial integrety. They claimed it spead up searches - i claimed it slowed down code production (hey - they also claimed finding child nodes in their org chart was best accomplished by using the LIKE operator *shudder*).

And as for the the continents . . . well, what if you want a list of available continents? You really should put them in their own table (SELECT DISCTINCT(CONTINENT) FROM COUNTRY is just plain silly).

But i am a big fan of normalization, other opinions may differ.

Jeff

R-R-R--R-R-R--R-R-R--R-R-R--R-R-R--
L-L--L-L--L-L--L-L--L-L--L-L--L-L--