3 thoughts on “Metadata is a Derivative Work”

  1. Let’s not start down the wrong path by talking about derivative works or licensing. Both assume that metadata are copyrightable. I’d start from the opposite assumption, that metadata are uncopyrightable facts.

  2. Ah! David, I rarely disagree with you, and I hesitate to do so without having my references ready, but this is too important. To classify metadata as “derivative work” would be a huge mistake. That would make writing metadata of someone else’s work copyright infringement.

    Metadata falls into the *fair use* category of “critical editions.” Unfortunately for those who back GNU-like openness, that places metadata into a BSD-license-type of space. You and I are free to create all the metadata we want, but so is company X, and they are free to charge for it.

    This is apparent in companies who make DVD players that automatically skip offensive content. Even though there is an ISO standard for marking up temporal media (MPEG-7, which we presented two years ago at your open learning conference), these companies prefer to use their own proprietary metadata format.

    Please, please, please, reconsider this stance. I fear some lawyer will stumble upon this post, knowing who you are, and use it a copyright dispute on this very topic.

    -jb

  3. Work that my colleagues and I have done on the Core Description Profile of MPEG-7 has been based on the assumption that information on films (a sort of metadata) constitutes a “critical edition” of sorts and is thus NOT a violation of copyright. This assumption has been guided by legal opinions from various sources but so far is untested in court. I strongly suggest that metadata falls into this category.

    For example, if we make available a list of comments on the scenes in a movie, then there is a great deal of precedent that suggests this is just fine. On the other hand, it seems clear that if we transcribe a movie and make that available as a sort of metadata (i.e. describing exactly what is said in the movie), then we would be violating copyright. If we make available an alphabetized list of the words with pointers to where those words appear, however, then we would not be violating copyright.

    Finally, I do not quite understand why anyone would want to put their stuff in the no-man’s land sort of limbo you seem to be describing. Essentially, why would someone create open courseware materials with accompanying metadata, making the materials openly available and then turn around and restrict “access to” and/or “reuse of” the metadata?

Comments are closed.