This post was originally posted on SDN on 07-10-2020. I decided to keep some of my posts here too.

Take a look at this: And this one: This one: vs this one: The flow: and Now with code.

Don’t focus on code semantics, but the visual representation. It is just dummy example, it’s all about formatting.

The first examples, following Clean ABAP preferred rules: and with helper variables: Now applying refactoring – Alt+Shift+R in Eclipse, some names changed: the second with helper variables after name changes: Now different style examples with line break and incremental indentation: with helper variables: Applying refactoring, names changed: and the second version with helper variables refactored: The last two images shows why I personally prefer using incremental indentation – no messed up code after renaming. Both styles are readable for me but this the second one without so strict alignment is refactoring-friendly, which is a very important thing for constant improvement of code and naming things correctly. I don’t want to manually correct main code + unit test code + dependent code after each name change (and be sure that everyone in a team remember to “fix” code after his or her refactoring changes). This heavily right-aligned code is also hard to read on smaller screens in editors or webpages like version control systems.

Speaking of version control – here are diffs for the Clean ABAP preferred rule, after refactoring I have aligned all things again.

ABAP Git diff (where are my name changes?): Bitbucket diff: Now incrementally indented code – changes are visible clearly.

ABAP Git diff: Bitbucket diff: An interesting video about this topic by Kevlin Henney (starting around 10-11 minute).

The key point from this presentation to me is that keeping the codebase stable after rather trivial refactoring operation is not conforming to a style, but rather to an invariant, a property.

There is also a discussion on this topic in Clean ABAP repo: