Published on 05 January 2013 by @mathiasverraes
I used to care a great deal about having a coding standard. I often took the initiative to write a document for the team that described the do’s and don’ts, and I was often the one setting up automated formatting sniffers on every commit.
I’ve come to see that we have been ignoring an opportunity to use formatting to our advantage. By arguing and voting over standards, and building automated tools to do the formatting for us, and refusing commits that failed the formatting validators, we lost track of why we needed a standard in the first place. It’s not to fulfil the desire of the pedants in our teams, or for the esthetic of neatly aligned spaces.
The goal should be to write easily readable and maintainable code. What makes code more readable depends on context.
Can you spot the difference between these to statements?
$countries is written very compact, no whitespace, in a single line that laughs in the face of 80 character margins.
$options, despite having only three elements, uses ample whitespace and newlines. But why?
$countries is not very readable, because it doesn’t have to be. Chances are you’ll never need to change that list, let alone
read it entirely. Just glancing at the beginning of the line tells you what you can expect the rest of the contents to be like.
And if you do need to update the countries, you’ll probably copy/paste a list from somewhere else.
By writing the line very compact, you give a visual suggestion to the reader: “The contents of this line is not important to you, don’t bother reading it.”
The contents of the
$options array on the other hand are important. If one day you need to investigate why redirects aren’t
being followed, you will read this code and change it. You’ll be thankful you didn’t have to squint your eyes to find the relevant
element in a huge unreadable compressed line of code.
That’s just a simple example of making our formatting dependend on what works for a specific situation. Make the boring implementation details, look like boring code. Compact the stuff the never changes, the generic boilerplate stuff, so that readers understand intuitively that it doesn’t matter. But make the important code stand out. Highlight it using whitespace, well-placed comments, and attractive formatting. Be creative and figure out where formatting can help you be more expressive and intentional.
Nice post, thanks. Another advantage of the $option approach is that it is easier to read VCS diffs. For example, if you would change alwaysUsePost to true, this is very obvious in a diff. If you make a change to $countries, this is less obvious. I personally also format method calls this way, if the call is bigger than one line for example, and only if it improves readability. Like:
Adding an option to the array will then also just show a clear 1 line add in VCS diffs.
Follow @mathiasverraes on Twitter.
|Modelling Heuristics||Workshop @ Domain-Driven Design Europe||workshop||Amsterdam, NL||Feb 1|
|Conference||Domain-Driven Design Europe 2017||organiser||Amsterdam, NL||Jan 31 - Feb 3|
|Domain-Driven Design||Neos Conference||workshop||Hamburg, DE||Mar 30|
|Keynote (TBD)||Neos Conference||keynote||Hamburg, DE||Mar 31|
This work by Mathias Verraes is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 License.