When I added them in I was expecting a vote to be triggered. I had forgotten that they would AE due to them being blank.
Put it to the vote. Delete the two comments and post the links back here. Seems the democratic way.
I generally agree with the other thoughts in this thread. It is bogus, but harmless. A little bit of clarity has never harmed.
Like @Jim_DeLaHunt pointed out, there are times when trying to find an artist via the GUI to link to a recording and disambiguations are really handy then as you have no lookup ability at that point. I’ve been adding some old obscure punk bands lately and some of the artists barely have names on the cover art. Disambiguations help save a lot of time during selection.
Seems to me a lot of this discussion would be avoided by adding the following to search results:
area, if set
years of birth/death, if set
any targets of a member-of AR (if more than 3, perhaps those with the most releases?)
if a certain threshold of oerformance ARs us reached: the most common instrument(s) (counting vocals as an instrument) (requiring at least N such ARs, and again as a top 3) (*)
if a certain threshold of technical ARs is reached: most common of those (mixer, recording engineer)
That would likely remove the need for many disambigs (like “UK flautist” or “bassist for Some Band”).
In addition, for all of these, localised versions can be produced. This could probably all be done at indexing time too.
(*) can get tricky. Would be even better if this could aggregate things, like showing woodwind for someone playing multiple woodwind instruments.
Replacing disambiguation comments with auto-generated ones feels like the wrong path to go down, the strategy chosen by e.g. Wikidata to disambiguate everything (or at least as much as possible) feels much better due to the long-term impact IMO.
Localizable disambiguations would be a nice addition, and IMO the best technical solution to this whole problem. Wikidata is a good example for how nice this can work as well:
Why? Localising all or even a significant chunk of disambiguation comments is going to be a huge task. If these are automatically generated from relationships, all translations will already be in place.
But relationships can change, and since you can’t think about all the artists when entering relationships, a simple edit could create confusing auto-generated “disambiguations”. I am certainly not against showing more information from the entity data in search boxes, but I am against claiming they are equivalent or even superior to manually entered comments.
I am not sure it’s actually necessary to localize all the disambiguation comments all the time into all the languages, for a large portion of items Wikidata also only has English plus maybe German or French descriptions. But where it’s desired, necessary, or just possible because you know the language, you can localize them as you come across or create items.
I keep seeing large search pages being shown in this discussion. Where I need disambiguation is when I am editing relationships. If I am adding an artist, then I am dealing with that tiny list view control and the search results. Disambiguation is ideal for this
What was a simple question is getting out of hand here. Who on earth would do all that translation work?
If I need FULL details on the selected artist, then I can click on their name and open the page. You can’t put everything into a disambiguation.
The main benefit of Wikidata-style multilingual disambiguations is that we could have, say, French disambiguations for French artists (which are much more useful for French editors!) while still having an English one to allow non-French speaking editors to figure out who this artist is and whether they should or shouldn’t use it. We do want to eventually move towards a more international website after all
That said, this is quite a lot of work and probably won’t happen anytime soon - it works great on Wikidata because the entire system is built around it. But I at least would love to see it happen
Note that I was not suggesting /replacing/ disambiguations. I was suggesting tweaks to search results that would (potentially significantly) reduce the need for them (and result in localised text for no extra effort).