Episode 119: DSLs in Practice with JP Tolvanen

Generator Optimization

Optimizing code generators involves addressing complexity through modularization and refactoring. Juha-Pekka Tolvanen highlights that structuring generators by domain concepts and refactoring them for commonalities can simplify the process, even though it may initially seem challenging 1. He suggests starting with easy cases to demonstrate immediate benefits, which can help convince management and colleagues of the value of code generation 2.

Once you convince management, when you convince your colleagues who don't believe that you can generate code, or you can convince your managers that they see, wow, this gave a good return of investment. Let's use this more so easy case first.

--- Juha-Pekka Tolvanen

This approach not only enhances productivity but also encourages the adoption of domain-specific languages.

   

Model Transformations

Model transformations play a crucial role in simplifying code generation by raising the level of abstraction. Markus Völter explains that adding rules to the modeling language can reduce the size of the code generator, making the process more efficient 3. While Juha-Pekka Tolvanen acknowledges the usefulness of model transformations, he emphasizes integration at the modeling time to ease refactoring and model management 4.

In our situation, if you see that something is changing in the data model, you can see maybe already that this is shared by somewhere else.

--- Juha-Pekka Tolvanen

This method allows for better handling of changes and ensures that models remain consistent and manageable.