Great Moments in Taxonomy

Horrible edge cases to consider when dealing with music.

Pairs nicely with Falsehoods Programmers Believe About Names, Time, Time, Time Zones, Geography, Addresses, Maps, Plain Text, Gender, Lists.

Previously, previously, previously, previously, previously, previously, previously, previously, previously, previously, previously.

Tags: , , ,

20 Responses:

  1. The one I remember driving me mad was the artist "Antistatic" released an album called "Workbench", because of course there's no such thing as an Antistatic workbench, is there?
    Oh wait...

  2. jer says:

    I like how the article is littered with typos, as if the subject matter didn't already complicate the task of its intended communication. (For example, it helps to know what "Mezbrow" refers to without needing to look it up.)

  3. Lloyd says:

    Band names: there's no mention of or love for the Kopyright Liberation Front, better known as the KLF, furthermore known as the JAMS, or Justified Agents of Mu-Mu, but also as the Timelords, 2K, and the One World Orchestra?

    Let me ask you a question, what time is love?

  4. Waider says:

    To some extent this feels a little contrived, since some of the "confusing" things are clearly not all that problematic from a perspective of software, and some of the more "creative" titles are clearly only possible because someone's laid hands on UTF-8 while in an altered state so they're de facto going to be ok in a properly-configured database (or cloud-scale wastepaper basket, or whatever you're storing them in). I prefer the "names" list because it seems a bit more to be coming from a place of "things you might plausibly (culturally) believe about names that are demonstrably not true, and which will fuck up your schema/application/day"

  5. thielges says:

    The title of which was stolen appropriated from an Abbie Hoffman book with the same name.

    I used to have a first edition copy of Hoffman's book on my shelf but has gone missing.

  6. robert_ says:

    Great shoutout for Navidrome, which I recently migrated to after Subsonic development seemed to grind to a halt.
    One thing worth noting is that Navidrome takes its album/track metadata from an MP3's ID3 tag so if there are problems with any of those edge cases, it's a question of whether to blame the ID3 limitations or Navidrome or both (though I believe ID3v2 handles most Unicode).

  7. Kyzer says:

    I have been on a MusicBrainz bender recently, so going through this list I was calling out things like "MB would handle that with an alias" or "that's what the sortname is for" or "MB guidelines say to normalize titles and artists unless there's strong evidence of artist intent" or "use the script and language on the release" (so if the tracklisting says "Tchaikovsky" you write Tchaikovsky, linked to Пётр Ильич Чайковский's artist ID). I was pleased to see the last line mention MB as an afterthought, and then as you go back through the article you see that loads of the examples are linked to MB.

    The Discogs people are even more anal, photographing the SID codes and CD matrix or the inner record label. Publishers enjoy fucking about and making multiple releases with the same catalogue number / barcode (as it mentions regarding Art of Noise). Sometimes the only disambiguation is "has a red label instead of blue" or "has a two column tracklisting on the back of the case instead of a single column", literally everything else is the same. Did they run out of blue labels at the factory that day? At least publishers can't pretend a physical release didn't happen, or it had different tracks, or a different artist, as they can and do with digital releases.

    What's nice about the MB database structure is that it used to be less complex, but it met and dealt with all the crazy edge cases as people wanted to put them in, and yet the result is still relatively clean, and Picard does a pretty good job translating the relational data to flat data in the form of ID3 tags (or MP4 atoms, or OGG/FLAC metadata)

    • thielges says:

      No kidding about Discogs. I started liquidating the remainder of my vinyl there and got lost in a maze of twisty passages, all alike. One old album has apparently been released in over a hundred editions. I was able to narrow my copy down to one of about a dozen points in their taxonomy. Since I can't tell whether it is worth $3 or $300, I'll avoid the wrath of persnickety collectors and dispose of my copy via an easier route, even if it doesn't capture top dollar.

  8. Scott says:

    Finding out more about the band !!! after hearing some of their music proved to be an exercise in frustration.

  9. Mark says:

    In my experience, all classical music recordings are edge cases.

    • tfb says:

      I always assume that's because no-one is putting any effort into arranging any kind of taxonomy for them, because the audience for classical (which, of course, is itself the wrong term) music doesn't overlap with the audience who uses these taxonomies much and they fit badly. There were (are, I think) vast and moderately comprehensive guides to recorded classical music: the people who made those must have had a taxonomy that works.

      • Mark says:

        Classical is the wrong term, but there is no right term and you knew what I meant.

        But yeah, the biggest issue is that classical recordings fit a completely different paradigm in terms of how you want to identify a given track/piece/album/artist than popular music (another wrong term for you), and iTunes etc. are written with the latter in mind. Classical recordings (mostly) won't be edge cases for taxonomies designed with them in mind, but such taxonomies aren't used by the systems most of us use to listen to music.

        • tfb says:

          Sorry, by 'the wrong term' I wasn't getting at you: it's the wrong term but I don't know what the right term is or if there is one in fact. Sometimes I hear 'art music' because, apparently, music made with electric instruments isn't art.

          • thielges says:

            Sometimes I hear 'art music' because, apparently, music made with electric instruments isn't art.

            Feel free to ignore with enhanced disdain anybody who paints with such a broad brush.

          • Mark says:

            Trying to name genres consistently, sensibly, and usefully is the deepest of taxonomical quagmires. As for the name "art music"...

          • Michael V. says:

            I've heard "compositional music" (because of its focus on the composer, rather than the performer,) and "orchestral music" (because it often uses instruments associated with orchestras.)

  10. jwz says:

    Those of you who keep seeing this post as a prompt to mention that you know of a song or a band that has a funny name -- one that in no way causes a taxonomy or even a collation problem -- should consider the value of silence.

  11. Dan Reish says:

    I can remember when, on Google Maps, it was impossible to get driving directions to addresses on Boulevard SE in Atlanta, a major commercial corridor.

    A few more fun pathologies about addresses in Atlanta beyond the tired "50 Peachtrees" jokes: there's an intersection of West Peachtree Street NW and West Peachtree Place NW, a street half of which is simply called "The Prado NE" (alphabetized under "the" in the city's old notecard-based street index), and better yet, the other half is called "South Prado Northeast", 30 Clifton St NE and 30 Clifton St SE are a couple of blocks apart and routinely get each other's mail, Main Street in Atlanta is an undistinguished suburban arterial miles from downtown, and the full name of a short stretch of West Peachtree is West Peachtree Street Northeast.

  • Previously