Extensible Modeling Languages

Transcription

1 Extensible ing Languages Utilizing Libraries for Creation, Use, and Domain-Specific Extensions 5th MODPROD Workshop on -Based Product Development February 8, 2011 Department of Computer and Information Science Linköping University, Sweden Equation-Based Object-Oriented (EOO) Languages 2 Domain-Specific Language (DSL) Primarily domain: ing of physical systems Multiple physical domains: e.g., mechanical, electrical, hydraulic Equation-Based Object-Oriented (EOO) s and Objects Object in e.g., Java, C++: object = data + methods Objects in EOO languages: object = data + equations ica VHDL-AMS gproms Acausality At the equation-level u = R * i At the object connection level I

2 Agenda 3 I I 4 ease and possibility of expressing complex models or tasks Language versions: A, v1.0 A, v1.1 A, v2.0 A, v2.2 Standard library versions: L, v1.0 L, v1.1 L, v2.0 L, v2.2 I

3 5 mechanisms to add new language features Uses Simulation Optimization Code generation for real-time export Grey-box system identification etc. C, v1.0 gives many dialects and different languages B, v1.0 A, v1.1 A, v1.0 A, v1.1 A, v2.0 A, v2.2 gives larger and more complex languages I 6 precise semantics meaning of the language Language Specifications of state-of-the-art are informally defined hard to interpret unambiguously when developing compilers hard to reason about when extending the language hard to formalize e.g. ica due to size and complexity I

4 What is MKL? 7 Purpose: Research language explore new concepts ing Kernel Language (MKL) Bottom-up approach Small extensible language Precise formal semantics Base it on a proven foundation the lambda calculus Statically typed functional language Platform for experimental equation-based DSLs (Continuous-time, hybrid, structural dynamic, and acusal models) I 8 I

5 - HOAM 9 Higher-Order Acusal s (HOAM) Higher-Order Functions I.e. first class citizens, can be passed around as any value + Acausal s s in EOO languages, composing DAEs and other interconnected models. = Higher-Order Acausal s I.e., first class acausal models. I HOAM Example 10 Example of a mechatronic system with a DC motor and a flexible shaft One shaft element is created by standard components. I

6 HOAM Example 11 Example of a mechatronic system with a DC motor and a flexible shaft Higher-order function that can compose any mechanical component in series I 12 I

7 ica Environment 13 ica ica Tool Result (e.g., simulation) Language Specification - Type checking - Collapsing the instance hierarchy - Connection Semantics I MKL Environment 14 Benefits Tool vendors no need to update tool after lib ext. developer - less dependent on tool vendors A model behaves the same way in different tools MKL MKL Tool Result (e.g., simulation) for using models - Connection for Semantics using models for using models - Simulation - Connection (Runtime) Semantics - Connection Semantics Language Specification - Type checking - Collapsing the instance hierarchy - Connection Semantics I

8 Intensional Analysis and Lifting 15 Static Semantics Dynamic Semantics Lifted Lifted Lifting Collapsed using evaluation Type Checking MKL Equation System Analysis, models treated as data. Connection Semantics, Simulation, etc. Result I 16 I I

9 of Semantics 17 MKL MKL Tool Result (e.g., simulation) of for using models - Connection for Connection Semantics using models Semantics for using models - Simulation - Connection (Runtime) Semantics - Connection Semantics Executable Specifcation Static Type System Language Specification - Type checking - Collapsing the instance hierarch Small-step Operational Semantics I How do we verify our solution? 18 Prototype Implementation Type Safety Proof Two main lemmas Progress Preservation I

10 Conclusions 19 ( Approach) Expressivness (HOAM) (Operational Semantics) ing Kernel Language (MKL) Thanks for listening! I