4 Responses to “Metadata and data structures”

  1. Bruce D'Arcus says:

    I don’t mean to be pedantic, but you’re conclusions are a little sloppy. You say:

    1) “Dublin Core does not define how data should be encoded. It is simply a list of elements.”

    So you’re using a concept from XML (“element”) to describe a format that is independent of XML, and so in fact distorting its purpose in the process.

    I’d say the core DC terms (title, etc.) are data properties or attributes. But DC is of course much more than just those core terms thee days.

    2) “RDF is similar to METS”

    “Similar” how?? I see very little similarity, except at the most superficial level that both allow mixing of different metadata structures. RDF, however, is a data model; METS is not.

    3) “SKOS is an XML format for thesauri”

    SKOS is an RDF vocabulary that can (like any RDF) be serialized as XML. But it is NOT fundamentally an “XML format.”

    I don’t mean to pick on you, but if you’re making sweeping suggestions like this, you need to be more careful about the details.

  2. Thank you for the feedback, and you make a number of great points.

    Yes, the word “properties” would have been better than “elements” to describe Dublin Core items. Yes, exactly, RDF, just like METS, can be used to mix content from various vocabularies together into a single XML file. That is what I meant. Regarding SKOS, again, your distinction is more precise than my description.

    The devil is in the details.

  3. I’m pretty much a fan of MODS at this point, it’s much less fiddly than MARC for my purposes.

    However, I’m wondering if there’s any discussion of a “full-text” tag for journal articles, etc. I don’t mean articles that an author has posted, but something more general, like the PLoS or preprint servers. In my case, there’s no URI because the text doesn’t exist as separate from its metadata, the MODS record *is* the online version. Any guidance for me?

  4. Avi, thank you for the feedback.

    If I understand your question correction, then I know for certain MARCXML is not designed to contain full-text mark-up. I’m pretty sure MODS is the same way. Both are intended to contain bibliographic metadata. On the other hand, it would be entirely possible to included a link (think “call number”) in either a MARCXML or MODS file pointing to the full-text of a journal article.

    As for the mark-up of the journal article itself, I would advocate the use of TEI. Many might think this is overkill since TEI leans towards the very analytic and scholarly, but in reality, TEI is well-suited to general mark-up of text — prose or poetry.

    HTH. –ELM