In a lot of places, I've hacked together a solution to take the "first" source in the list and return that as source_url so that legacy code works, but I don't feel like that's the most elegant thing to do.
Some UIs are still only designed for a single source URL.
Scraping doesn't autofill the source URL field anymore.
Multiple sources needs some extra features:
source_count/sources_countsearch field