I'm writing an application using a FirstSQL database. Thus far everything has worked pretty well out-of-the-box, even though FirstSQL isn't listed as a supported DB. I'm simply using the DefaultDialect.

I've progressed to a point now where I'm considering the potential advantages/difficulties of writing a custom FirstSqlDialect.

Could someone explain to me whether this is feasible and how it might be done, or point me to resources/documentation for writing a Dialect?

I would like to shift from my current bottom-up approach to a middle-out approach, using the mapping to maintain both generated Java skeletons and the schema DDL. From what I understand from reading (and rereading) Hibernate-In-Action, I'll need a proper Dialect for hbm2ddl to work properly.

Many thanks for any help,
Thom

Should probably mention as well, that I'd be happy to share any Dialect work with the Hibernate project, for other users, should anyone else choose to use Hibernate w/ FirstSQL in the future. FirstSQL has informed me that we are the only clients experimenting with Hibernate that they know of, to date, but they are very interested in the outcome.

If the DefaultDialect works - what is the problem then ?

The best source of how to write a custom dialect is to see on the ohter dialects how they have handled their "quirks".

If you want to find out which methods hbm2ddl uses go look at Configuration.generateSchemaCreationScript() and .generateSchemaDropScript().