The Benefits of Diligence

Victor Arrial,Giulio Guerrieri,Delia Kesner

2024-04-19

Abstract:This paper studies the strength of embedding Call-by-Name ({\tt dCBN}) and Call-by-Value ({\tt dCBV}) into a unifying framework called the Bang Calculus ({\tt dBANG}). These embeddings enable establishing (static and dynamic) properties of {\tt dCBN} and {\tt dCBV} through their respective counterparts in {\tt dBANG}. While some specific static properties have been already successfully studied in the literature, the dynamic ones are more challenging and have been left unexplored. We accomplish that by using a standard embedding for the (easy) {\tt dCBN} case, while a novel one must be introduced for the (difficult) {\tt dCBV} case. Moreover, a key point of our approach is the identification of {\tt dBANG} diligent reduction sequences, which eases the preservation of dynamic properties from {\tt dBANG} to {\tt dCBN}/{\tt dCBV}. We illustrate our methodology through two concrete applications: confluence/factorization for both {\tt dCBN} and {\tt dCBV} are respectively derived from confluence/factorization for {\tt dBANG}.

Logic in Computer Science

What problem does this paper attempt to address?

### What problem does this paper attempt to solve?

This paper aims to solve the challenges faced when embedding two computational models, **Call - by - Name (CBN)** and **Call - by - Value (CBV)**, into a unified framework - **Bang Calculus (dBANG)**, especially the problem of maintaining dynamic properties. Specifically:

1. **Maintenance of static properties**:
- Some previous studies have successfully studied certain static properties of CBN and CBV (such as normal forms, quantitative types, compact types, inhabitability, etc.) by embedding them into dBANG. These properties can be derived from dBANG through appropriate encoding.

2. **Maintenance of dynamic properties**:
- Dynamic properties (such as the behavior of reduction sequences, concurrency, factorization, etc.) are more complex and have not been fully explored. In particular, the reverse simulation from dBANG to CBV has not been well - handled in the existing literature, resulting in the inability to directly derive the dynamic properties of CBV from dBANG.

3. **Introduction of a new embedding method**:
- The paper proposes a new dCBV embedding method that can maintain both static and dynamic properties and meet the requirements of reverse simulation. This method uses the modal operators `!` and `der` delicately and performs "diligent administration" during the application process to ensure that all administrative steps are completed as early as possible.

4. **The concept of diligent administration**:
- Diligent administration is a method to ensure that all redundant `d!` steps (i.e., administrative steps) are executed as early as possible. This helps to avoid unnecessary complexity and ensures that the reduction sequences from dBANG to CBN/CBV can be correctly mapped back.

5. **Specific applications**:
- Through two specific examples (concurrency and factorization), it shows how to easily derive the corresponding properties in CBN and CBV from the properties in dBANG. This unified framework not only simplifies the proof process but also provides a three - in - one solution: only one proof is required to cover the properties of CBN, CBV, and dBANG.

In summary, the main contribution of this paper lies in solving the problem of maintaining the dynamic properties of CBV, and by introducing a new embedding method and the concept of diligent administration, it makes the property derivation from dBANG to CBN and CBV more systematic and reliable.