CS has custom code for serializing JSON and CSV
Where?
What's the situation?
- JSON serialization uses a custom library.
- CSV is often written by string concatenation.
What does it make more difficult?
- Delimiter / encoding bugs could easily crop up.
- It's difficult to extend the code and be confident about it.
- If there are security holes in the custom library it's difficult to reliably patch.
- It's more difficult to switch wholesale to a new data format across our services.