The AI-Enabled SDLC

Chapter 7. Automated Creation of Documentation

Documentation plays an essential role in the SDLC. Documentation within and about the codebase makes it understandable and reviewable, and it is critical for maintainability in the long run.

However, putting in the time and effort to create good documentation can seem tedious and often takes a lower priority to other development-related work. Given how well SDAs are able to work with text data and natural language processing out of the gate, they are a natural fit to ensure good-quality documentation for your code.

Documenting Content

In the most common use case, SDAs can produce header documentation and relevant comments in the body of their code. Many have built-in functions to do this, but with limited comments. For example, Tabnine has a built-in Document function available through VS Code’s Codelens functionality above blocks of code (Figure 7-1).

After selecting the Document function, Tabnine generates suggested header documentation that can be accepted or discarded, as shown in Figure 7-2.

Figure 7-1. Tabnine’s Codelens documentation function

Figure 7-2. Generated header documentation suggestions

To get meaningful comments throughout the code, you will often need to engage directly with the AI through the chat interface and ...