How to write and maintain natural language instructions?

Writing instructions in natural language that operate your system under test is a bit like programming without technical skills. All you need is knowledge about the system's domain and common sense. Both prerequisites will guide you to build a Domain Specific Language (DSL) that enables you to "program" what the application under test is supposed to do and to formulate expectations that prove whether the business rules have being applied correctly or not.

Unlike programming, before the first natural language instruction exists, you have no commands and syntax rules to follow. These will come into existence time by time when the library of your Domain Specific Language is growing. However, in the very beginning the degree of freedom is so high, that you may feel trouble how to start. To avoid this typical initial problem SysNat provides some guidance:

1. As in programming, different styles exist how instructions may be formulated (see How to formulate natural language instructions?). Choose the style that makes you and your team feel comfortable. Best Practice is to use declarative style on high level system descriptions which call natural language scripts. Those scripts are ideally written in imperative style for your low level system description.

2. Follow the given Best Practices (see Scalability in natural language testing).

3. By default, SysNat has example applications under test. Have a look into their language template libraries.

---

Back to Home