Questions That Should Be Asked More Frequently

[Previous] [Next]

Q. So what's the best way for me to update my database? Should I write all the code myself, or should I use updatable Recordset objects throughout my application? Should I believe the marketing hype, or should I believe the gurus who say that if you want a job done right you must do it yourself?

A. The simple answer is that there's no simple answer. You can go to either extreme, or you can look for some middle ground. Keep an open mind, and try a variety of techniques.

One recommendation: try new techniques, but give yourself time to experiment with them. If you plan ahead and design your application well, you might be able to move from one method to another. For example, you could build an initial version of your application that passes Recordset objects across process boundaries in a three-tiered application and uses the ADO Cursor Engine to modify data in your database. In later releases, you could use the Recordset objects to interact with the user and store changes he makes. Then you could examine those changes and call your own stored procedures to update the database. You could later modify your components to use two-dimensional arrays, Variants, or UDTs to store the results of queries and maintain data. Almost anything is possible.

The more you experiment with the different ways to update your database, the better equipped you'll be to decide what's right for you. Think about what you need in functionality and performance. Then figure out a way to determine whether a particular approach will give you what you need.

In Part II of this book, we'll discuss features of the ADO Cursor Engine. As you read about these features, you'll discover what the ADO Cursor Engine has to offer—and you might decide to use it in your applications. Or you might develop a clearer understanding of how to implement similar features in your own libraries.