Calvin: Thanks for your thoughts. I'm a software engineer also, so I appreciate the complexities you are pointing out.
Calvin Walton-2 wrote: > > ...Speaking as a person who is currently attempting to write a music > player > that uses Musicbrainz metadata... Inheritence of ARs makes a lot of > things in the database model a /lot/ harder to deal with, particularly > when doing things like searching for all works composed by a particular > artist.... > I agree, there is complexity here. But is it really the Inheritance proposal which is causing the complexity? I'd argue that the complexity is inherent in the real-life domain we are trying to model. Composers really do create compositions which have multiple levels of structure. Releases really do map Tracks to recordings of fragments of compositions (movements, scenes), rather than to complete compositions (symphonies, operas). Some ARs really do indicate structural relationships between (parts of) compositions, while others really do indicate attributes describing compositions. I'm hearing a couple of people in this thread allude to a preference for having Relationships describing compositions be applied to every parent and child Work entity in the Parts Relationship tree of Work entities. This is a legitimate choice for the data model. It makes application development easier, at the cost of making data entry harder, accidental inconsistency of data more likely, etc. But we could decide we prefer that trade-off. Calvin Walton-2 wrote: > > ...Speaking as an editor, I actually don't really mind doing adding these > duplicated ARs manually. Although I admit that I don't do much with > opera or musicals, etc. :) ... > Ah, then come do an opera sometime. I typically see 30-50 partial compositions under a single Opera entity, which accounts for the number of acts and scenes in the opera, plus the number of different choices Release producers make about where to insert track boundaries. Regardless of how the Relationship Inheritance proposal works out, the data entry software will have to improve before I would expect any sane editor to enter Work entities for a fully fleshed out Opera entity. Calvin Walton-2 wrote: > > ...It would be /really/ helpful to have some sort of support for this by > the musicbrainz server and webservice; preferably /before/ the guideline > starts to be applied to data on the site.... > I had intended this proposal to guide addition of new data into MusicBrainz, not to initiate a campaign to get rid of ARs on partial Work entities which inheritance renders redundant. But I get the impression that you and others are concerned about delaying the purge on such ARs until the software is up to snuff. i'm OK with that. I'm even OK with an interim policy which says, put ARs everywhere for now, while we get the software able to handle inheritance, and then do a much bigger purge of redundant ARs. Would that answer your objection? Calvin Walton-2 wrote: > > None of the existing portions of the website (e.g. the recording page) > or tools (e.g. bitmap's release releations script) support going > multiple levels deep to pull out a composer AR; so in these cases you > simply wouldn't see one even if it's there. Other users of the > webservice would have similar issues. > The way I look at it, I'm not distressed that editing and search and tagging software won't support inheritance the moment this proposal would get approved. From my point of view, the software is all broken *already*, because the Composer relationships are already mostly inaccessible. I think it's common for editors NOT to put ARs on the child-level partial Work entities. And those are the entities which search and taggers use. Approving this proposal doesn't make the software any more broken, it just provides a clearer statement about how the software needs to change to fix its present brokenness. Making a decision that, as an interim step, we will encourage editors to put ARs on all levels of the Work Parts tree -- that is a workaround which would make our present software appear less broken. I do agree with that. I don't agree it's the smartest long-term design, but we could collectively come to a different decision. -- View this message in context: http://musicbrainz-mailing-lists.2986109.n2.nabble.com/RFC-339-Partial-Works-Relationship-Inheritance-tp6939661p6942736.html Sent from the Style discussions mailing list archive at Nabble.com. _______________________________________________ MusicBrainz-style mailing list MusicBrainz-style@lists.musicbrainz.org http://lists.musicbrainz.org/mailman/listinfo/musicbrainz-style
