I would say that yes, the queries could be complex. I did forget to mention this. Ideally it would be possible to query on X + Y + Z.

Let me make an example. Let's ay my data will be a large index of recipe books and each book is a collection of recipes by different people. Recipes have names, ethnicities, date of origin, authors, ingredients, etc...

I may want to do a search for an Italian recipe that features a specific ingredient. Or any recipe that is in a book that has a title starting with Red and the recipe itself has the word Pie in its name and was written before 1990.

I have also been informed that it's quite the science knowing how to structure the database itself, including what to index, to make sure it is of a desired size (doesn't bloat), performs at a reasonable speed, etc...

Damn, I'm just the project coordinator and designer, not a database guru. I just know we need some type of one. We can make our own in house which would be suitable for a smaller number of indeces, but we could not develop something that is as robust as mySQL and that would stand up in performance to the same brute-force or convoluted searching.

Keep in mind I don't want to be saddled with having to run a daemon on the system at all times. Access should be able to be made dynamically by our application(s) whenever we need to. When we're not running there shouldn't be any overhead on the system other than the disk space our data takes up.

Thanks for the very fast responses. I knew this would be one of the best places to ask (when isn't it?)

Bruno