Introduction
This article serves as an editorial preface to a special issue of Acta Informatica dedicated to the field of formal synthesis. Drawing from the 7th Workshop on Synthesis (SYNT 2018) held in Oxford during the Federated Logic Conference (FLOC), this collection highlights a significant paradigm shift in how we approach software and system design.
In the traditional programming model, engineers dedicate the majority of their effort to the "how"—the specific sequence of instructions required to meet a requirement. Formal synthesis proposes a different path: programmers focus on specifying what a program should do, while automated algorithms handle the implementation details. In this framework, formal specifications effectively become high-level source code.
The Problem: The Burden of Manual Programming
Traditional software development is inherently difficult and error-prone. The gap between human intent and machine execution is often where bugs and security vulnerabilities reside. By focusing on formal specifications, researchers aim to reduce this gap, allowing synthesis algorithms to generate "correct-by-construction" code or control logic. This move shifts human creativity further up the abstraction stack, focusing on requirements rather than low-level instruction management.
Synthesis as High-Level Compilation
A compelling mental model for this field is viewing synthesis as a form of high-level compilation. Just as a C++ compiler translates human-readable code into assembly, a synthesis tool translates formal logic specifications into executable machine code. This perspective positions synthesis not as a replacement for programming, but as its next logical evolution toward higher levels of abstraction.
The Three Pillars of Modern Synthesis
The research presented in this special issue—comprising ten papers selected from seventeen submissions through a rigorous peer-review process—defines the state-of-the-art across three main pillars:
- Applications to Programming: Automating the creation of software snippets or entire modules to streamline development.
- Control of Physical Systems: Generating logic for hardware or robotics that must interact with the real world. This highlights a growing intersection between formal methods in computer science and control theory in engineering.
- Algorithmic Advances: Theoretical and practical improvements that make synthesis faster and more scalable.
The Synergy of Theory and Practice
A key takeaway from the SYNT 2018 workshop is the maturing nature of the technology. The field is no longer purely theoretical; practical performance improvements are now being published alongside deep theoretical advances. This synergy suggests that synthesis is moving toward broader applicability in complex, real-world systems.
Conclusion and Future Outlook
While synthesis is a rapidly evolving field, challenges remain in scaling these techniques to larger systems and broader classes of specifications. However, the transition from manual instruction-writing to automated synthesis represents a fundamental change in the engineering of reliable systems. This special issue serves as a window into that evolution, showcasing the transition from theoretical foundations to practical, scalable implementations.
Summary updated