Semantic Modeling Edge Case List

From GetSemantic

Jump to: navigation, search

Following a discussion with earle on #swig, the following is a list of edge cases that one should take into account when producing metadata formats, RDF vocabularies and ontologies, tools and database schemas.

[edit] People

  • Not all names follow the Western cultural practice of "Firstname Lastname".
  • Gender is not binary: transsexuality/transgenderedness exists.
  • Stage names, pen names and pseudonyms exist.

[edit] Books

  • Some books are ghost-written.
  • There should be some way of linking translations of books together
  • A translator is not an author, but has an author-like role.
  • Sometimes books are published posthumously.
  • Different chapters can be written by different people, and different sections of a book can be translated by different people.
  • Different chapters can be written at different times.
  • Some book sections denote where they were written.
  • Marginalia exists.
  • Sometimes books contain pronounciation marks.
    • eg. "Self-pronouncing" bibles
  • Some books contain the same text in multiple languages (eg. Bibles, ancient philosophical texts)
  • There are books which have different editions, but that is due to a translator updating a translation rather than an author changing the text.

[edit] Music

There are plenty of problems with music metadata formats, especially the ID3 format.

  • Musical recordings are created by more than one person, or more than one defined group of people. This includes collaborations, mashups, "features", rap battles, remixes, performances and many, many more.
  • There can be more than one genre for musical recordings.
Personal tools