So, our main search tool will get a relevance search, which means that the output is somewhat sorted by how important an entry is. Before that the entries were sorted by date and if two entries had the same date the order of the entries was not specified.
If a customer is going to search for some stuff with the relevance search the table is ordered by how many words are hit in several fields, also the date has some influence.
The order of the entries is calculated by the backend (or db, i don't know) and could be exposed in the user facing table as an own column, which could also be used to filter entries with low relevance aka low score.
I feel like this should not be done, because mainly I think it is wrong to burden the user with complexity used to give an elegant behaviour. But the rest of the team insists on doing it, to give the user the power to filter irrelevant entries.
So is there a general advice? Is it a good idea? Is someone doing it?