On Fri, Jan 24, 2014 at 12:47 AM, Ulrich Klauer <ulr...@chirlu.de> wrote:
> Nicolás Tamargo de Eguren: > > > Now that the orchestra attribute is going away (thankfully) we > > should store this data where it belongs: on the artists. I would > > like to add a few new > > artist types to more clearly store this kind of info. > > > As a minimum, I'd like the top level here of orchestra, choir and > > chamber ensemble (ideally, of course, I'd want them all :) ): > > > Orchestra > > -Chamber orchestra > > -Symphony orchestra > > Choir > > -Male choir > > -Female choir > > -Mixed choir > > -Children choir > > Chamber ensemble > > -String quartet > > -Piano trio > > 1. All of those are actually sub-types of "Group", not top-level types > in their own right (like "Person" and "Character"). However, the > server software does not support the notion of subtypes at the moment; > this means that whenever a subtype (or subsubtype) is added, a code > change is required in order to display the start/end dates as > "Founded"/"Dissolved". Possibly in more places if, e.g., MBS-2604 were > to be implemented. Also, external software has no chance to find out > that new artist type 37 (say) is to be treated as a group. > Well, that sounds like something that would be reasonably easy to implement in the next schema change if we feel this is necessary. I'm fairly sure we can live with "Begin/End" dates until then, if nobody wants to write the code for this at the moment - all the places where this distinction is made inside MB at the moment are basically decorative except for deselecting Gender, which should hopefully be a trivial change anyway (so should the others, for that matter). 2. I don't actually think this is where the data belongs, at the very > least not for the subsubtypes like "String quartet". Instead, we > already have "is member of group playing <instrument>"; so, a group > consisting of two violinists, a violist and a cellist can be > considered a string quartet. > I'd argue it's more useful the other way around though - marking something as a string quartet automatically offering those four roles as a predetermined place to add the relationships from. I'm willing not to add subtypes anyway though, as I already said when originally sending the RFC. > 3. Introducing those very specific subsubtypes would lead to an > enormous proliferation. Just have a look at the "common ensembles" > listed in http://en.wikipedia.org/wiki/Chamber_music#Ensembles - other > responses already hinted at that, too. Only if you add every single type of ensemble, which would be pointless (and which is why I was disagreeing with turning it into a full-level ontology with all kind of extras). Again though, I'd be OK with just the main types. > All of these are fairly acceptably defined > > 3. Are they? Are The King's Singers a male choir? Is Rockapella? Since both seem to be considered vocal ensembles, the answer is no. The definition can be somewhat circular, but so is the Album/EP divide (and even the Group one, for one-official-member bands like Nine Inch Nails) . > Is a choir that sings TTBB, but admits female tenors (typically for lack of > male candidates), a male choir or a mixed choir? How many people are a > chamber ensemble, how many do you need for an orchestra? > Again, an orchestra becomes one when it considers itself one, mostly - certainly a string trio / quartet isn't one, even though some people keep using the orchestra rel for them (which was the main reason to introduce a separate "ensemble" type really, to try and get people not to pick "orchestra" for those - it's possible that it could be enforced with a guideline anyway I imagine though. For the choir, I would still call that a male choir probably, but for the few doubtful cases, there's nothing wrong with picking just "choir" :) > > An example of how this could be useful (apart than from just storing > > the data and knowing) is for having "Group" offer adding members > > from the edit artist page, while choir or orchestra artists would > > offer to add conductors too. > > A general type as outline above could achieve that, too. Agreed, that's why that's the minimum I'd want :) I'd like the orchestra / choir subtypes because the choir ones are *generally* (with a few exceptions like the one you mentioned I guess, which I don't doubt happens but I'm yet to see) clear, and the orchestra ones we had already on the orchestra relationship so I imagine people were reasonably able to make the distinction and thought it was interesting to a point. But I don't think they're basic for this to work, and could certainly be added later once 1) is dealt with if we want them. -- Nicolás Tamargo de Eguren
_______________________________________________ MusicBrainz-style mailing list MusicBrainz-style@lists.musicbrainz.org http://lists.musicbrainz.org/mailman/listinfo/musicbrainz-style