I don’t use them right now, but when I experimented, I prefer alphabetical. If I wanted to cluster properties in proximity, I would do something like
---
author: Dumbledore
dataAnalysis: discourse analysis
dataType: interview
---
Alphabetical is for easy scanning, matches my linear brain, and also means less decision fatigue regarding trying to figure out how I want to sort.
The only exception is tags, which I always put at the bottom for easy clicking and aesthetic reasons, however, if this feature request were implemented then I wouldn’t have a tags field referred to as such, but would use tags in other properties, i.e. the below vs. the above:
---
author: Dumbledore
dataAnalysis: "#discourse-analysis"
dataType: "#interview"
---