@andresmmera thank you for the update but I have some observations. The update you have made seems to be very similar to the existing mechanism of handling subcircuits in qucs. The principal drawback is that it mixes the device parameters spice and osdi libraries. The difference we would like to have is to import PDK data of the devices (which many times are modeled as subcircuits). Once imported we would like to provide mechanism to populate the parameters from the XML data and elaborate netlist entries according the specified format (SPICE, CDL). The separation of the geometry is also perfect and aligns with our ideas. The other feature is to (re)calculate the values of some parameters based on parsable equations. So your initial idea was much more close to what we expect.