Sorting the fields alphabetically is not straightforward. Currently, the human-friendly labels are accessed dynamically (“on the go”), rather than all at once. What we have at hand initially are only the raw field identifiers, e.g., "mex:institution" or "mex:name".
Additionally, proper sorting depends on the current language, which further complicates things.
Implication
To sort fields alphabetically in a user-friendly way, we would first need to fetch or resolve all the labels for the current language, perform the sorting based on those labels, and then render the fields. This adds extra complexity compared to sorting by the raw field keys.