86. Lightning Talks

Design Choices

Database design choices significantly impact performance and maintainability. Joe Zack and Alan Underwood discuss the challenges of treating stored procedures like functions in other programming languages, which can lead to inefficiencies. Alan highlights the importance of thinking in sets rather than objects when working with databases, as this approach optimizes performance 1. Joe questions whether using stored procedures that call other stored procedures is a code smell, as it complicates working with SQL Server results 2.

Good practices in your OO code make for horrible SQL, and good SQL practices make for horrible OO code.

--- Joe Zack

Balancing maintainability and performance often requires duplicating code, which contradicts traditional clean code principles 2.

   

Anti-Patterns

SQL anti-patterns often arise from treating SQL like any other programming language. Joe and Alan discuss the pitfalls of this approach, emphasizing the need to recognize when SQL is being misused as an application language 3. Joe points out that while duplicating code is generally discouraged, it can be necessary for performance in SQL, as true duplication differs from accidental duplication 4.

You can only scale SQL Server so much before you can't buy any more hardware.

--- Joe Zack

Alan notes that many applications won't face scaling issues, but when data grows unexpectedly, revisiting modularity for performance becomes crucial 4.

   

Logic Placement

The placement of business logic in databases versus application layers is a contentious topic. Alan and Joe debate whether logic should reside in the database or be handled by the application, with Joe noting that the decision depends on the type of logic involved 5. Alan argues that placing logic in the database can make it the central nervous system of an application, which may not be ideal for complex ecosystems with multiple data storage mechanisms.

It's not just the heart at that point. It's your heart and your brain.

--- Alan Underwood

This approach can complicate scaling and integration with other systems, suggesting that a balanced strategy is necessary for optimal performance and flexibility.