In that thread I was picking up on how an undocumented random selection of characters need to be dropped from the search strings. Yet there is nothing that tells you this. If characters like dashes in a cat no are ignored then they should be removed as part of the search. i.e. you must search for “FRED123” if you want to find “FRED-123”. Very unnatural.
The noob as to know a bit too much about the way the database is structured here. If they want to do a quick search, the drop down box is not very nice for them. Noobs don’t know what works or release groups are.
Look at Discogs. You can throw ANYTHING into that search box and it chases for a match. Track title, album title, artist, barcode, catno - it doesn’t care. THAT is noob friendly.
If the search gives too many answers, you can then swap to their advanced search. Much clearer boxes put up onto the screen to fill in.
IMHO the search here at MB is just too geeky and too tied to database knowledge. It would be nice to lay a more noob friendly search over the top of this. MB is not really an accessible place.
Though as IvanDobsky says the standard “throw anything” discogs seach box is very noob friendly.
If there is any match then the new user at least gets a search result that appears to them reasonable -
Something that works great in discogs is searching for album name together with artist name, without using AND syntax and it works, it finds the album by this artist, usually in first results.
There are hours of editor time being wasted with the creation of duplicate Releases.
A friendly unisearch box would reduce this waste. And increase the proportion of time in which editors who are currently duplicate-adding feel satisfied by their contribution to MB.
(And no, this time it isn’t me who has added yet another duplicate Release.)
It could be easy, with a friendly unisearch box, to quickly check that a potential addition is not very similar to something already in the db.