On Jan 3, 2008 5:33 PM, Bram van Dijk <[EMAIL PROTECTED]> wrote: > > * I do presume that _because_ AR's are being propagated, anything that > > is specifically AR'ed has had a higher standard of fact checking > > applied by the editor who put it there, and they did it to the highest > > possible level of precision, so release level really meant 'applies to > > every track' or the release as a whole. > > > Right, this is the main point I think. In this discussion there are some > IMHO valid reason given to stop propagating AR's.
Indeed. And if everyone hasn't (and they clearly haven't) been using this principle, it's a good reason to not autopropagate the ones we have. > > * I do think if data is ambiguous or incomplete, it's better to > > document clearly what is known, and where it's known from, in an > > annotation, than to add fuzzy data to the AR's. > > > I disagree that a release level AR (take the example artist A performed > drums on release B) is ambiguous of incomplete or something. On the > track level this AR would not mean that A played at 1:15, there might be > no drums in the first 2 minutes or something, it only means that > somewhere in the track a drum was played by A. It is ambiguous however, on a per-track basis. So why _must_ it be an AR, and not noted in the annotation? > At the release level we have the same, a drum, somewhere on the release, > was played by A, not necessarily all the time, and not even necessarily > on each track. Liner notes are more often than not given on the release > level, and there is nothing fuzzy about them, they are just less > detailed than we might wish. ... they're not fuzzy but they're not detailed? You can't have it both ways. They _are_ fuzzy, if they're not detailed, c'est la vie. You'd be amazed how often further research turns up more detail though. Interviews, videos, press releases, artist websites, often contain detailed information. One Swedish musician I track has been adding details of his old albums to his website for the last couple of years. He's been publishing along with them the notes he wrote at the time, occasional personal diary excerpts, sometimes studio schedules, and all sorts of detail that just wasn't available on the original 1973 or whatever liners. I've noticed several other people using this information to flesh out the AR's track by track on some of those old albums. > > * I think it's better to add only actually known information to the DB > > relationships, because it's always easier to add than to clean data. > Like a said, a release-level credit that does not apply to each track is > known information, not ill-researched, ambiguous or fuzzy, but verified > known information. But it is all of unknown, ambiguous, and fuzzy from the point of view of any particular track. That doesn't mean it's bad or non-useful, but if we don't know, we just plain don't know. Would it be so terrible if this kind of information lived in the annotations, unless and until more information turns up? -- Lauri Watts _______________________________________________ Musicbrainz-style mailing list Musicbrainz-style@lists.musicbrainz.org http://lists.musicbrainz.org/mailman/listinfo/musicbrainz-style
