There’s one problem with prose stored in version control systems: line breaks. Longer lines mean bigger diffs, which can make it hard to review changes. One approach, and the one I take, is to put each sentence on a line. This works pretty well, but there’s a better approach: semantic line breaks (SemBr). In SemBr, line breaks separate sentences into logical pieces.
The problem with SemBr is that I struggle to make my brain do it. It took me a while to figure out why I struggled with SemBr despite understanding the benefits. A while back I realized it’s a simple answer: I overthink it. When I write, I think about the meaning of words. When you add in thinking about the line breaks, the cognitive load goes way up.
Sentences (and character counts) are unambiguous. Everyone who uses the same style will end up with the same line breaks. Semantic chunks can be more ambiguous, so if you’re working with others, there’s a (self-imposed, no doubt) pressure to get the breaks Rightâ„¢.
My friend Matthew offered this succinct summary:
put things that are likely to change on their own lines.
Like
URLs,
dates,
and
list items.
So maybe I’ll give it a try again. And if not, at least I can explain why not.