The reason for this post is that I just remembered this edit some time ago.
The performer is the group Berliner Saxophon Quartett consisting of 4 members. I had added the group and the 4 persons as recording artist to the recording. In the mentioned edit, the group was removed with the argument that it is already credited in the recording artist field. As this argument still feels wrong to me (is it covered by some official guideline?), I would like to have a discussion about that.
I think it would be better allowing to add relationships to the recording without any restrictions coming from the entry of the recording artist field.
Simply because the relationships offer much better possibilities for giving structured data, like the actual role of the artist. In fact, itâs not too clear to me why we need an recording artist at all. Relationships can everything the recording artist field can, plus much more.
Listing the four performers is listing the quartet - only with better info. After all, it was the people who really performed there, the quartet is just a way to group them.
I would definitely like the ability to indicate âX performed on Y as part of group Zâ somehow, but otherwise I donât think it makes any sense to have them both, suggesting that this was performed by these four saxophonists plus a quartet.
I think if we added a relationship to a Beatles song with all members credited that just said âperformer: The Beatlesâ, it would feel absolutely redundant and be removed quickly. I donât really buy that this is any different in classical.
I completely agree with you that âX performed on Y as part of group Zâ would probably be the best solution.
I also agree with you that this has nothing to do with classical. But why should it be bad to add a relationship for the Beatles in your example? There are already redundancies in the relationship: The ârecorded atâ relationship gives the date 1962-11-26. Giving this again at ârecording ofâ is redundant.
Trying to put my thoughts in a nutshell: Relationships are the superior way of representing metadata. The recording artist field has much fewer possibilities (and given the rich framework of relationships, I donât see a clear reason why we need it at all). Thus, the entry in the weaker recording artist field shouldnât restrict the data which is allowed to be added the the relationships.
I just came across of another example. In the performer relations of that recording, the string quartet âAmadeus quartetâ is listed, and additionally the 4 string players.
reosarevok, do I get you right that âAmadeus quartetâ should be removed here? I donât like this, as we are throwing away information on a recording without good cause. Classical recordings often appear on dozens of releases (in this case, already 5 releases are linked) without an evident connection to each other. Commonly, this makes it quite hard to dig up the right recording. We should leave anything in place which might be of help for finding the recording.
You get me right that I would not add it. Thatâs not a guideline, though, just my personal view on the issue. The recording will already appear on the recordings page for the quartet, for finding needs.
That said, since itâs not a guideline, thereâs no need for you to remove it (itâs not âwrongâ, officially). I can live with it being there like that until we can do multi-point relationships, although I am afraid thatâll take basically forever
Yes, but this is separated from the performer-relationships where it wonât show up. The higher the chance to find the recording, the better.
Or a slightly different point of view: Relationships provide the most fine-grained possibilities for searching (at least in theory), so preferably the full set of information should be contained in the relationships.
Thank you for clarifying!
If I changed everything wrong that Iâm spotting, there wouldnât be time for sleep.
Mostly 1, in my view. Groups just, well, group performers, but they donât perform as such, do they? I still think itâd be very weird to have a âperformer: Black Sabbathâ relationship in all tracks in Paranoid when we already have the members, for example. But maybe thereâs a point, I dunno.
Oops, really? I always had assumed that 2. is your point. So the performer-type orchestra (among others) should never be used?
This is a bit like saying: All my CDs have only a single track. Itâd be weired to have track numbers.
What I mean is: To get a clear picture, one should better look at the more intricate examples, not only the simplest ones. To stay with a rock music, letâs look at the Deep Purple album Concerto for Group and Orchestra. It is a live performance together with The Royal Philharmonic Orchestra. The recordings have the Royal Philharmonic Orchestra as the performing orchestra relationship. So you are saying that in your opinion, this relationship should better be removed (no matter whether the members of the orchestra are listed separately or not)? This feels really really wrong to me.
No, it should definitely not be removed if the members arenât listed separately, because linking to a grouping of performers is better than not linking to anyone at all But if all the orchestra members are listed, then having that relationship too sounds at best redundant to me, and at worst confusing (âall this people play, and also the orchestra? This sure must be a work for huge forces!â).
I think we would both agree though that the ideal option would be to have â(list of performers) performed (recording) as part of (grouping)â though - which is also how itâs usually printed on albums! Itâs just not possible right now, and might not be for years
In the meantime, Iâm fine with not removing the group relationships when present, and I think I might stop doing that - Iâm unlikely to start adding them myself though
Right now what is possible, though, is to link the members of a band (or orchestraâŚ) to the group via the member of relationship.
Unfortunately, that doesnât let you specify who played on this recording, or that while Bob normally plays electric guitar, on this recording he played acoustic guitar.
But if you do bothâput the people who played and the group on the recording and in addition make sure the member-of relationship is in the database, then that actually is clear: Bob and Tom play drums on this recording. So does the group The Hypothetical Band. Bob and Tom are members of the Hypothetical Band, so theyâre not additional people playing.
(If you just have Bob and Tom, without the band, you donât know which of two bands theyâre in, or if maybe this is something that they did without the other band members).
[But please donât do this with orchestras. I doubt the page would even load once youâd added 60+ people to each recording].
Let me comment that in my opinion, groups clearly are performers. However, I see the redundancy problem of listing people twice, which apparently we cannot completely avoid at the moment.
And let me once again add my second point: The recording artist is a weak to dubious piece of information for me. We should better not rely on it, all the information should preferably also be given in the form of advanced relationships (such that the recording artist can be created automatically from that information).
Just out of interest: Why? What is so hard about implementing those multipoint relationships?
Thank you! I already seized the opportunity and re-added the âBerliner Saxophon Quartettâ in the initial case.
Our entire relationship structure (and database tables) are designed as âA relationship Bâ. To change that from 2 points to n points, weâd probably need to change a looot of stuff under the hood.
Resurrecting this interesting thread as it seems relevant to a question Iâve hesitated over a few times recently. There is the marvellous script âReplace recording artists from a Release pageâ which transfers all performer relationships to artist credits. But I wonder if it is always OK to do this.
Should we care about the order of the artist credits for a recording? This came up in this edit, for example. I think the answer is probably not.
The recording artist field is supposed to be a âsummaryâ according to the CSG. But the script will put everything there, as in this edit of mine. Does this matter?
If the answer to either question is âyesâ, it would be lovely to have an enhanced script that gave more control over what was transferred, but I donât know quite how that would work.
