DLF 2011 - How to explain to librarians the utility of formal data modeling
What does formal data modeling mean to people? It's easy to rely on tools like RDF, UML, but does there need to be something more abstract behind that?
What lightweight tools can be used to do this?
Dangerous to do this outside the context of actual use cases.
Have them participate rather than explain.
Separate what's in a record or on a screen from what the structured data behind it is.
Show something, just don't talk about it. Show a real thing that can be done with more structured data, a more rigid data model, that can't be done now.
Big problem with LD now - have to imagine what's possible, it's hard to show today what the benefit would be.
Does the W3C LLD report (http://www.w3.org/2005/Incubator/lld/XGR-lld/) help get at this issue at all? Best thing about the document are the use cases articulated there.
Shift from name authority records to RDA Person Entities is an example of moving to more formal data modeling in libraries.
In Archives, they've also made a shift to describing a person rather than a name. This takes shape in EAC-CPF (http://eac.staatsbibliothek-berlin.de/). This is XML-based, not RDF-based.
Use case for more metadata about a person: allows reading a book, following a link to what that author was bookmarking in Del.icio.us the year they were reading the book. VIVO documents a person's research activity, which could be used for this purpose.
Formal data modeling can be hard to pitch, while we're pitching backing off on strictness of data. They're not alternatives to one another, though - they serve different purposes, so comparing the two doesn't really make sense. Strategy: tie it back to use cases. Focus on the structure, and let the fluff be the fluff. Ask to justify why what you think might be fluff is important, what use case it fulfills. It takes time to make time.
Is there some way we can make formal data modeling, or using stricter data standards, seem like less of an imposition?
- work incrementally