命令式vs 声明式 models

Imperative and declarative models

Pesic’s PhD thesis 里总结到，

Imperative models take an “inside-out” approach; that is, every possible execution sequence must be modeled explicitly. As a consequence, imperative modeling may lead to over-specification and lack of flexibility, making it difficult to defer decisions at runtime and to change existing process models.
To overcome these shortcomings, declarative modeling approaches have been proposed. In contrast to imperative approaches, declarative models take an “outside-in” approach. Instead of describing how the process has to work exactly, only the essential characteristics are described. To this end, constraints are specified that restrict the possible execution of activities.

命令式模型采用“由内而外”的方法;也就是说，每个可能的执行序列都必须显式建模。因此，命令式建模可能导致过度规范和缺乏灵活性，使得在运行时延迟决策和更改现有流程模型变得困难。 (mutual exclusion constraint 互斥关系很难表达)
为了克服这些缺点，提出了声明式建模方法。与命令式方法相反，声明式模型采用“由外而内”的方法。而不是描述过程必须如何准确地工作，只有基本特征被描述。为此，指定了限制活动可能执行的约束。

Declatative language

A declarative language only describes what the essential characteristics of a process are while it is insensitive to how the process works.
For instance, a possible key quality of a process can be that a specific action is “just being executed”. Formalizing this quality as a predicate ranging over a set of actions, one can use the temporal logic LTL to model how executions of actions relate to each other over time.

声明性语言只描述流程的基本特征，而对流程的工作方式不敏感。

声明式(Declarative)编程和命令式imperative)编程

链接: link.