Clarification on interface boundary warning and symbolic transformation in MethodOfLines.jl

[itsKgs]

Hi! I'm solving a PDE using MethodOfLines.jl, and when I try to discretize my system, I get this warning:

Warning: The system contains interface boundaries, which are not compatible with system transformation. The system will not be transformed. Please post an issue if you need this feature.
@ MethodOfLines .../pde_system_transformation.jl:43**

I'd like to understand more about:

• What exactly are interface boundaries in this context?
• Why do they prevent system transformation?
• Even though I see this warning, the numerical solution appears to be correct — can this safely be ignored?
• Is there a symbolic workaround if I want to use StructuralTransformations or Optimization.jl later (e.g., via remake, replace, DiffCache)?
• Should I avoid interface-style boundaries for symbolic work?

Thanks for all the great work on the SciML tools!

[xtalax]

Interface boundaries are any bcs which connect disparate domains together, and are how periodic bcs are implemented.
System transformation often involves adding new eqs and bcs to make certain types of terms compatible with the codegen. It wasn't clear to me how to implement this for certain boundary types, hence the warning. If your discretize doesn't fail with this warning, you don't need to worry, system transformation can also be done manually so if you need a periodic or interface, also don't worry.