Alphabetical CSS, but let the machine do the sorting
Eric Bailey's post on organizing CSS declarations alphabetically got a shout-out, and the reaction starts from a blunt observation: the most common declaration ordering technique in the wild is no technique at all.
The most common CSS declaration organization technique I come across is none whatsoever.
The counterpoint is that a writer tends to group declarations by whatever comes to mind while drafting, which usually lands on semi-logical clusters — box model properties together, color properties together. In practice that hasn't mattered much, and that attitude is shaped by working mostly on small teams or solo.
The team argument
Eric's case for alphabetical ordering is about consistency across people:
[…] it imposes a baseline sense of structure across a team. This ask is usually enough, especially if it means cleaning up what's come before.
His larger point is that the resulting structure helps legitimize CSS in a field where CSS skills are undervalued. That part is hard to dispute.
Don't do it by hand
What is disputable: hand-alphabetizing an existing codebase is probably not a good use of time, and done blindly it could break things. That risk is exactly why Prettier punted on sorting, as documented in an issue comment.
If a team decides the convention is worth adopting, the advice is to wire it into the editor as an on-save action and back it with a pre-commit hook. Normalizing is work for computers, and the output can be checked while authoring.



